petroff

defence

by A. Raetsky
& M. Chetverik

EVERYMAN CHESS
Gloucester Publishers pic www.everymanchess.com
First published in 2005 by Gloucester Publishers pic (formerly Everyman Publishers
plc), Northburgh House. 10 Northburgh Street. London EC1 V OAT

Copyright © 2005 Alexander Raetsky and Maxim Chetvcrik

The right of Alexander Raetsky and Maxim Chetverik to be identified as the authors
of this work has been asserted in accordance with the Copyrights, Designs and Pat­
ents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system or transmitted in any form or by any means, electronic, electrostatic, magnetic
tape, photocopying, recording or otherwise, without prior permission of the publisher.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.

ISBN 1 85744 378 0

Distributed in North America by The Globe Pequot Press, P.O Box 480,
246 Goose Lane, Guilford, CT 06437-0480.

AU other sales enquiries should be directed to Everyman Chess, Northburgh House,

10 Northburgh Street, London ECtV OAT
tel: 020 7253 7887 fax: 020 7490 3708
email: info@everymanchess.com
website: www.everymanchess.com

Everyman is the registered trade mark of Random House Inc. and is used in this
work under license from Random House Inc.

EVERYMAN CHESS SERIES (formerly Cadogan Chess)

Chief advisor: Garry Kasparov
Commissioning editor: Byron Jacobs
General editor: John Emms

Typeset and edited by First Rank Publishing, Brighton.

Cover design by Horatio Monteverde.
Production by Navigator Guides.
Printed and bound in the US by Versa Press.
 CONTENTS I

Bibliography 4
lntnxluction 5

1 e4 e5 2 lLif3 lLif6

1 3l£Jxe5: The Main Line with 8 c4 7

2 3l£Jxe5: The Main Line with 8 :.c1 34
3 3l£Jxe5: Black Plays 6.. ..id6 47
4 3l£Jxe5: Deviations from the Main Line 69
5 3l£Jxe5: Fourth and Fifth Move Alternatives for White 80
6 3 d4: The Main Line 104
7 3 d4: S .i.d6 and 5th Move Alternatives for Black
... 143
8 3 d4: Fourth Move Alternatives 158
9 3 d4: Black Plays 3...exd4 168
I0 Third Move Alternatives for White 178

Index of Complete Games 189

The Petroff Defen ce

BIBLIOGRAPHY I

Books
Ji'a.rhionable Varia/ion in tbe PetrrJ.ffDye11ce, IUctsky and Chetvcrik (Voronczh 1992)
Modi!Ji 1•aritmt msskoi parlii, IUetsky and Chctvcrik (Voronezh 1990)
Rltssisci.Je Ptlflie, Schwarz (Hamburg 19H6)
JI.Jtssischc f.'erfridigNI'!,. Konikowski and Sicbenhaar (Germany 1992)
The Cochrtmc G01nbil, Matsukevich (Moscow 1994)
TIJc PeiiT!ff,Janjgava (Gambit 2001)
The J>etrq!J Defence, l'orintos and Haag (Batsford 1991)
17:Je PetroffDefence, Yusupov (Oirns 1999)

Periodicals
Cl.�ess 11ifo1711ttnl 1-90
Nen' i11 Chess YeariJOok. 1-72

Software
Chess Assistant 7.1

4
 INTRODUCTION I

'1'/Je e111irr lheol)• �/chm opmi11gs is diiJided i11/o lbe Pellf!lf D�Je11ce tmd the �jerted J>rllf!UDtjellce (1 d4, 1 t4
tllld other li11ts). Alexander Ractsky
-

TI1e Petroff Defence has been one of the most fashionable 'open games' in the last decade,
but it also has a long history. After 1 c4 e5 2 ltlf3, instead of defending his e5-pawn, Black
prepares a counterattack on White's e4-pawn with 2...ltlf6. J .ucena mentioned this idea as far
back as 1479, while in 1512 Damiano analysed 3 ltlxe5 ll'lxe4 and, of course, after 4 tfe2
reached conclusions that were discourah>i.ng for Black. No wonder the symmetrical opening
was forgotten for many centuries after that! However, in I H24 Russian maestro, Alexander
Petroff, found out that 3 ltlxcS could be met by 3 ...d6!, and only after that should Black cap­
ture the e4-pawn. In 1842 another Russian expert, Jaenisch, published valuable analysis in
l'ttltlfllede. The opening was given a name of the Petroff or the Russian Defence Qn Russia,
Germany and Scandinavia).
It is quite narural that two Russian masters write about the Petroff Defence for the pub­
lisher Everyman Chess. Just like ballet, the Russian Defence could be labelled as property of
the Russians. Grandmaster Arrur Yusupov is an outstanding thl:orctical expert of this opening,
while Russian champions Smyslov, Karpov and Kramnik often usc it in practice. Raetsky and
( :hctverik have not made a substantial contribution to the theory of the Petroff Defence so
far. However, it's worth pointing out that our first published articles and booklets were de­
vntL'l.l to this opening in the early 1990s. Alexander Raetsky has considered the Petroff De­
fence to be his f.1vourite opening for more than twenty years and can be proud, at least, of the
l(uantity of his b>amt."S played in the Petroff Defence (about 200, including 60 correspondence
1-�amcs), if not the quality.
Emotional players arc scared off by the S)•mmetrical trend in the Petroff Defence; they arc
afraid of the drawish aspirations of weaker opponents who play White. But what can be more
symmetrical than the initial position in chess, which has not yet been ruined by the notorious
'draw death' despite Capablanca's indications? A more skilful strategic player triumphs in the
Petroff Defence regardless of the colour of his pieces and the position's symmetry. The
healthy strategical foundation of this opening allows Black to defend a lot of systems in the
Petroff Defence even while playing against stronger opponents. Even the lines declared
'doubtful' by theory are normally better than their reputation.

5
The Petroff Defence

Apart from minor alternatives (see Chapter 10), White has to choose between 3 d4 and 3
'Oxe5. The authors believe that after 3 d4, the move 3.../0xe4 is stronger than 3...exd4. Here
the sharp variations like 4 dxe5 .i.c5!? and 4 .i.d3 d5 5 'Oxe5ltld7 6'0xd7 ..i.xd7 7 0-0 �4!?
can replace the popular, solid, but uninteresting alternatives (4 dxe5 d5 and 4 .i.d3 d5 5/0xe5
/Od7 6 'Oxd7 .i.xd7 7 0-0 .i.d6 respectively). After 3 'Oxe5 it is more difficult for Black to
initiate an open battle and the positional niceties of the fight come to the fore.
One final advantage for the fans of the Petroff Defence: you don't have to study numerous
complicated openings like the Ruy Lopez, Two Knights Defence, Italian Game and Scotch
Gamel After 2'0f3'0f6 you are home and dry!
Finally, some acknowledgements. Special thanks go to the founder of the Petroff Defence,
Mr Petroff, and many thanks to our Danish/Scottish friend Jacob Aagaard for his technical
help. Also thanks go to Zoya Nayshtut for her excellent translation into English and also to Mr
Yusupov for his great book on the Petroff Defence- the real bible for people from our 'cast'
(Petroff Defence players).
Play the Petroff Defence and be happy!

Alexander Raetsky & Maxim Chetverik,

Voronezh,
January 2005

6
 CHAPTER ONE I
3ltJxe5: The Main Line
with 8 c4
1 e4 e5 2 �f3 l!Jf6 3 l!Jxe5 d6 4 l0f3
li)xe4 5 d4 d5 6 .i.d3 .i.e7 7 0-0 �c6 8
c4
i\Jter 1 e4 e5 2 tL'lf:\ tL'lf6 3 tL'lxe5 d6 4
li'!fJ tL'lxe4 5 d4 d5 6 .i.d3 Carl Jaenisch sug­
)l;ested the system 6....i.e7 7 0-0 tL'lc6 as far
back as the 19th century. This line is still very
popular today, especially with regard to the
immediate attack on the centre with 8 c4, the
Kubject of this chapter.
Let's see how this line has developed over
the years. The oldest reply, 8....i.e6, is not
very good in view of 9 cxd5 .i.xd5 10 tL'lc3
(hy the way, the li)wenthai-Morphy game
ran be considered a model - sec the notes to
( lame 1l) or 9 l:te I l'Llf6 10 c5. After the
more agJ,rressive development of the bishop
with R ...i.g4, Whitc should prefer 9l'Llc3 to
Kimplifying the play in the ccntre with 9 cxd5
..xd5 to l'Llc3 tL'lxc3 11 bxc3 (again sec the
notes to Game 11).
With 8...tL'lf6 I31ack protects the d5-square
and strengthens his kingside at the same
time. In the case of 9 tL'lc3 .i.e6 (Game 1 1),
White has a pleasant choice between 10 cxd5
nnd to c5, so Black should refrain from the
development of the cH-bishop. If 9...0-0 10
h1 Black appears to equalise by means of
IO dxc4 11 .i.xc4l'Lla5 12 .i.d3 .i.c6. If we
.•.
start with 9 h3 (Game 12), castling is again
effective, but White maintains an trutJaove
after 9...tL'lb4, hassling the d3-bishop.
The ...l'Llc6-b4 raid is more appropriate
when the other knight is still placed on c4,
i.e. 8...tL'lb4 (Games 1-10). One option for
White is simply to ignore this attack: 8...tL'lb4
9 cxJ5 tL'lxd3 10 'ii'xd3 1i'xd5 II l:.c l .i.f5
(Games 9-10). In these games White tries to
take advantage of the hanging position of the
black pieces with tL'lb1-c3 �mmt:diately or
after 12 g4 .i.g6) or 12 tL'le5.
Instead of 9 cxd5 White more often rc­
trl..-ats his bishop to a safe place with 9 .i.e2
(Games 1-8), a lint: that has bccome fashion­
able owing to Karpov. The line 9...dxc4 10
.i.xc4 0-0 (Game 8) leads to a standard posi­
tion witl1 an isolated pawn on d4. Instead
Black usually prefers to maintain pressure in
the centre by means of with ...0-0 and ....i.e6
or ....i.f5. However, the order of the moves
is important here. After 9....i.e6 the queen
exchange 10 ..a4+ 'ilt'd7 11 'ilt'xd7+ �xd7 12
l'Llc3 is interesting, and 10 c5!? followed by
an attack on the quccnside is even more so
(sec the notes to Game 6).
9 ...0-0 is preferable to 9....i.e6 because af­
ter 10 l'Llc3 Black has several possible ways to
develop his light-squared bishop. In Game 7
Black plays 10 ...b6 followed by ....i.b7. One
of the main lines is 10 ....i.c6, against which
7
 The Petroff Defence

White can play 1 1 !.DeS (Games 3-4) and 1 1 Pava.'IOvic, Istanbul Oly mpia d 2000 - and
.l.e3 (Games S-6). here Pavasovic claims an edge for Black after
The most fashionable line at the moment
16...exd3 17 Wb3 'ifd7) 1 S...d3 1 6 �3 dxc2
is 10....i.f5, which was utilised by Kramnik in
1 7 1i'xe2 .l.e6 1 8 llcl1 1Wc7, which Pavasovic
his 2004 World Championship match with
regards as unclear.
Leko. lbis move is discussed in Games 1 -2.
b) 1 1 �3 leads fairly dirccdy to a draw:
1 1 ...dxc4 12 .l.xc4/.Df6 1 3/.Des 'ifxd4 (Bolo-
Game 1 gan pointed out the error 13 ...1.Dc2? 14/.Dxf7
Leko-Kramnik llxf7 IS .i.xf7+ �f8 1 6 .l.e3 when White is
World Ch. (Game 1), Brissago 2004 much better) 14 .i.xf7+ Wh8 1 S I.Lif3 •d7 16
'------• /.Des ..d4 17 I.Lif3 'Wd7 was agreed drawn in
1 e4 e5 2 �f3 �f6 3 �xe5 d6 4 �f3 Bologan-Zamicki, Buenos Aires 2000.
Q)xe4 5 d4 d5 6 .i.d3 &6 7 0-0 J.e7 8 1 1 ...�c3 1 2 bxc3 &6
c4�9 J.e2
Of course saving the bishop is the most
popular option. Alternatives arc studied in
Games 9-1 0.
9 0-0 10 �c3 J.t5
.•.

1 1 a3
This is the natural move but we should
also consider two others:
a) 1 1 �5 cS and now:
a 1) 1 2 dxcS is not a dear error but it does
allow pressure on f2: 12..ixcS 13 I.DxciS
(White must be careful: if 1 3 a.1? then
1 3...1.L!xf2! 1 4 llxf2/.Dc2 1 5 lla2 llc8 1 6/.Dd3
.l.xf2+ 17 I.Lixf2 dxc4 and Black has a clear
advantage) 13...lDxdS 14 ..xd5 'irxdS 1 5
cxdSlLd8 16 i.c4 llfc8 t7/.Df3 h6 1 8 i..d2
I.Lixcl2 1 9 I.Lixd2 i.cl4 with an unclear posi­
tion.
a2) 12 a3 cxd4 13 I.Lixc4 dxe4 14 axb4 f6
1 5 I.Dg4 (15 �t3?1 d3 1 6 .i.xd3 - J.Polgar-
1 3 lZ.e1
After 13 cxdS .xdS White has a few
choices:
a) 14 lle 1 llfe8 transposes to the main
game.
b) 14 .i.e3 l'lfd8 1S :e1 &s 1 6/.Des f6!
(16.....b3? is just a blunder: 1 7 .l.c41 .xd 1
1 8 .i.xf7+ � fB 19l:taxdt /.Dc6 20 .l.ct i.LixeS
21 AxeS .i.g4 22 llde1 .i.f6 23 ll5e4 and
White has won a clear pawn, Movsesian-1
P.Nielsen, Blmdesliga 1 999) 1 71.Dd3 i..d6 1 81
I.Db4 •n is unclear.
c) 14 c4 1i'd6 IS dSI.DeS t6/.Dd4 .i.d7 1 7
a 4 c6!? 1 8 l:tb t (perhaps White shoula prefer
1 8 'ifb3 cxdS 1 9 cxdS 1Wb4 20 'ifxb4 .l.xb4,
when his d-pawn is marginal!)' more of a
strength than a weakness) 18 ...cxdS 19 cxd5
..xdS 20 I.Lif5 .c6 21 I.Lixe7+ ..xc7 22
llxb7 1We6 was level in Van Den Doei­
Schandorff, Esbjerg 200 1.

8

tl) 14 .if4lfr.ts 15 .i.xc7 b6 16 .i.f4 Lc8
17 .i.t.12 .i.d6 18 llc1 h6 19 a4 :C7 20 h3
Wh3 21 .i.b5 (Krakops-Ulescas, Leon 2001).
Now Black cook! have played 2t ...'exdl 22
l:lcxd1 �b3 23 L2 �xd2 24 �xd2 llxc3
with equality.
13 .. .1Z.e8
I\ simple, sensible developing move but
131ack has other options.
13....i.f6 is probably not qui te good
enough to cyualise: 14 .i.f4 �aS 1 5 cxdS
WxdS 16 Wa4 (if 16 .i.xc7 :lacS 17 .i.xa5
Wxa5 18 c4 llfd8 Black has pleasant com­
pensation) 16...b6 t7lDd2 .i.d7 18 'irt.11 and
\Vhitc had an cdbJC in Short-Bologan, Skan­
tlcnborg 2003. This advantage grew after
IK.J:ac8 19 i.d3 .i.f5 20 �4 .i.e7 21 'irh5
i.xc4 22 1txd5 .i.xdS 2.l llxe7 llfe8 24
Jl:ac I �f8 25 llxe8+ llxe8 26 llxe8+ Wxc8
27 .i.xc7 .
13...dxc4 14 .i.xc4 has been well tested:
a) 14....i.f6 IS ..tf4 'ird7 16 ..ta2 llfe8 17
li.\g5!? (this is Ftacnik's suggestion; 17 'irt.12
bS 18 .i.gS .i.g6 19 .i.xf6 gxf6 20 �h4 Wg7
21 f4 toe7 22 ...f2 �f5 23 g4 �xh4 24
'l'xh4 f5 was unclear in Adams-Karpov,
Dortmund 1999) 17...llxet+ 181fxcl .i.xg5
IIJ i.xgS llc8 20 'ird2 h6 21 i.f4 tlX!s 22 d5
b5 and White has a nagging t:dge.
h) 14. ..i.d6 1 S i.gS 'ird7 16 �h4 �a5 1 7
.l.a2 with a further branch:
b l ) Black should avoid 17... b5?1 18 a4! a6
19 ••xb5 axb5 20 �f5 1fxf5 21 i.e7 llfb8
3 liJxe5: The Main Line with 8 c4
22 g41 1Wf4 (even worse is 22. ..1!rd7?! 23
.i.xf7+! Wh8 - the point is 23...Wxf7 24
llxa51 llxa5? 25 1i'b3+ Wg6 26 llc6+ when
Black will soon have to give up his queen -
24 ..ixd6 1i'xf7 25 ..tc5 and White was close
to victory in Karpov-Portisch, Torino 1982)
23 .ixd6 'irxd6 24 'ifo •d7 25 :C2lDc4 26
llacl and White has a strong initiative, espe­
cially since 26...1le8?? loses to 27 lle7!.
b2) 17...i.e6 18 .i.xdi fxe6 19 �0 :aes
20 .ih4 1t'c6 21 1tc2 h6 and Black is just
active enough to hold et:juality: 22 .i.g3 1i'd5
23 a4 �4 24 lle4 a6 25 1i'e2 1i'c6 26 lle1
.i.xg3 27 hxg3 tLki6 28 llxe6 1i'xc3 29 lieS
1i'b3 was now agreed drawn in 1\.dams­
Anand, FlOE World Championship, New
Dell1i 2000.
14 cxd5
14 .i.f4 is discussed in Game 2
14 ....xd5 1 6 ..tf4
White has a minor alternative in 1S �e3
.i.f6 (:tlso fine is IS...�aS!? 16 �d2 'ird7 17
.to .i.d6 18 c4 c6 19 'lfa4 .i.e? with an un­
clear position) 16lDd2 .l:lad8 17 .i.c4 11d7 1 8
�b3 .i.e7 19 a4 .i.d6 20 'lfd2 (Bologan­
Degrnevc, Belfort 2002) ands here Bologa.n
gives 20...�7! 21 .i.f4 b6 as equal.
1 5.. .1Z.ac8
1 8 h3
16 .i.d3 is a serious alternative. 16...1Wd7
17 llbl .i.xd3 1 8 'irxd3 b6 19 d5 .i.f6 2 0 c4
h6 21 h3 lle7 22 llbd 1 lids (if 22...:Ce8 23
llxc7 �xe7 24 l:te1 lidS 25 .icS White has
9
The Petro ff Defence
the more comfortable game) 23 l:lxe7 lbxc7
24 lbes .i.xe5 25 .i.xe5 l:le8 26 .i.g3 lbf5 27
.i.xc7! ..xc7 28 .xf5 'lrxc4 29 d6 and
White's strong passed pawn gave him the
edge in Leko-Anand, Unares 2003.
A few days after this main game Lcko
tried 16 c4 'lre4 1 7 .i.e3 and now:
a) 1 7...'ifc2!? (Kramnik's novelty) 1 8 d5
lbas 19 lbd4 •xdt 20 l:lcxdt (if 20 l:laxdt
Parkin analysed 20....i.d7 21 lbb5 b6!? 22
lbxa7 l:la8 2' lbb5 .i.xb5 24 cxb5 .i.xa3 as
equal) 20....i.d7 21 .i.d2 .i.f6 22 .i.xa5 .i.xd4
23 l:lxd4 l:lxe2 Vz-112 Leko-Kramnik, World
Championship (Game 3), Brissago 2004.
b) Also reasonable is 17 ....i.f6 18 l:la2 b6
1 9 h3lba5 with a further split
b1) 20 .i.d2 lbb7!? (20...�1?1 leads to a
very unpleasant ending: 21 'ifxb1 Lb1 22
l:lxb1 l:lxe2 23 'itft l:lce8 24lbg1 ll2e4 25
.i.xa5 bxa5 26 lle2, Kotronias-Marjanovic,
Kallithea 2003) 21 'ifa4 .i.e6 22 d5 lbc5 23
1i'xa7 .i.d7 when White's awkwardly placed
queen provides Black with compensation for
the pawn.
b2) 20 g4 .i.g6 21 g5 .i.c7 22 lbe5 .i.d6 23
lbxg6 hxg6 24 c5 .i.ffi (Kasparov analysed
24...bxc5 25 dxc5 .i.f4 26 .i.g4 llcd8 27 l:ld2
.i.xg5 28 .i.d7 lieS 29 l:ld4 as slighdy better
for White) 25 .i.g4 l:lcd8 26 l:lae2 1i'c6! 27
1i'c2 (27 cxb6 was immediately abandoned as
a draw in Anand-Adams, Unares 2002)
27...bxc5 28 dxc5 •d5 29 1lrc3 lbc6 is equal
according to Dokhoian.
Brissago (3) 2004.
1 6....i.e41?
Previously 1 6..if6 had been played: 1 7
lbh2!? 'ifa5 1 8 .i.d2 lbc7 (or 1 8. ..l:lcd8 1 9
.i. f3 h6 20 lbg4 .i.xg4 21 hxg4 .i.g5 22
.i.xg5 hxgS - Kramnik-Anand, Wijk aan Zee
2003 - and here Hu?.man gives 23 .cl!
l:txet+ 24 'IVxel 'itffi 25 l:tbt lle8 26 .cl as
clearly better for White) 1 9 lbg4 .i.xg4 20
Lg4 ltcd8 when Anand bclit.-ves White is
slighdy better.
1 7 .i.e3 c!lla 5
Another try is 17 ...l:lcd8 18 lbd2 ..tg6 19
10
.to •d7 20 1i'a4 lbe5 21 '1Vxd7 lbxf3+ 22
lbxf3 ltxd7 23 lbe5 l:ld5 as in Leko­
Bologan, Dortmund (rapid) 2004. Igor Zait­
sev now suggests 24 c4 l:ldd8 25 lbxg6 fxg6
26 d5 �fl 27 �ft l:ld7 28 l:lab 1 b6 29 ltb3
as a way for White to gain an edge.
1 8 c4?!
This allows Black the chance to give up
his queen for a definite equality. A better try
for the advantage is 1 8 lbdz!? .i.f5
(1 R.ixg2? may look worrying for White,
but I. Zaitsev provided the clever refutation:
19 c4 .c6 20 d5 .g6 21 .i.hS .i.f3+ 22
Lg6 .i.xdt 23 .i.f5) 1 9 .t.n 'lid7 20 lbb3
lbxb3 21 1i'xb3 c6 when White has the mer�
est of edge£.
1 8...li»lc41 1 9 .i.xc4 1i'xc4 20 �d2 'ttd5
21 li»le4 ••e4 22 i.g5 1i'xe1 + 23 1i'xe1
.i.xg5 24 ••5 i.f& 25 •••7 c5 26 •xb7
i.xd4 27 :S2
Belov gives 27 l:tdt llb8 28 .d7 l:le2 29
ltxd4 cxd4 30 11Vxc.J4 as equal.
27 ...c4
Black must avoid the back rank tric�1
27.. 1J..c7? 281le2!.
28 Jle2 lled8 29 e4?!
White should have played 29 l:ld2 .i.e5 30
£4! .i.f6 (not 30 ...i.xf4?! 31 llxd8+ llxd8 32
.
'ifc6 hS 33 'ifxc4 .i.g3 34 �ft :Cs 35 1rc3
h4 36 a4 when White has some winning
chances) 31 ltxd8+ l:lxd8 32 .c7 c3 33 a4 g6
34 aS l:ld2 35 a6 lta2 36 �h2 c2 37 a7 after
which the game would have been drawn.

29 . ..c3 30 We4 ..i.b6!
Now Black can think about trying to win.
31 1t'c2
White has to blockade: 31 Wb4?! .i.xf2+
12 'i&i>xf2 c2.
31 ...g6 32 11'b3 l:ld6
·n1e careless 32 ... fld2? would spoil cvcry­
lhin�-f. 33 llc7 llxf2 34 11'xb6 c2 35 llc71 and
White wins.
33 .D.c2 .ia5 34 g4 .D.d2 35 Wg2 .D.cdB 36
J:l.xc3! .ixc3 37 �i'xc3 .D.2d5 38 'tlt'c6 l:la5
39 'iii>g3 llda8 40 h4 ll5a6 41 Wc1 l:la5
42 ft6 .D.xa4 43 h5 .D.4a5 44 Wf47
A disastrous slip. Instead 44 hxj.,>6 hxj.,>6 45
11 g5 46 'ird6 and White hollis the draw.
44 .g51 45 'tlt'f6 h61
..
Now the win is just a matter of rime.
46 f3
The point is that 46 .xh6? ll8a6 traps the
L lucen.
3 li)xe5: The Main Line with 8 c4
46 .D.5a6 47 'tlt'c3 l:la4 48 •ce .D.8a6
•••
49 WeB+ Wg7 50 'tlt'b5 A4a5 51 1rb4
l:ld5 52 'tlt'b3 l:lad6 53 1i'c4 l:ld3 54 Wf2
l:la3 55 Wc5 l:la2+ 56 Wg3 Af6 57 Wb4
l:laa6 58 Wg2 l:lf4 59 Wb2 + l:laf6 60
'tlt'e5 Axf3 61 1i'a1 l:lf1 62 •ca A1f2+
63 Wg3 l:l2f3+ 64 'tlt'xf3 Axf3+ 65 Wxf3
Wf6 0-1
Game 2
Grischuk-Adams
Halkidiki 2002
1 e4 e5 2 l£lf3 ll'lf6 3 ll'lxe5 d6 4 ll'lf3
ll'lxe4 5 d4 d5 6 ..td3 li)c6 7 0-0 ..i.e7 8
c4 ll'lb4 9 .te2 0-0 1 0 ll'lc3 .tf5 1 1 a3
ll'lxc3 1 2 bxc3 li)c6 13 l:le1 l:leB 14 .i.f4
The main altemative to 14 cxd5.
14 ...dxc4 1 5 ..i.xc4 ..i.d6 1 6 AxeS+ 'tlt'xe8
1 7 li)g5 .i.g6 18 .ixd6 cxd6 19 h4 •e7
Anand demonstrnteJ the following clever
variation: 19...M?I 20 h5 .i.xhS 21 'ti'xh5
hxg5 22 'iii'h21 llcB 23 lletlli:Je7 24 .i.d3 g6
25 WxgS llc7 26 llc3 and White is obviously
bener.
20 'tlt'g4 h6 21 ll'lh3
21 h5 has also been tried, but led only to a
draw in Movscsian-Gelfand, Bled Olympiad
2002 after 2t....i.xh5 22 WxhS hxg5 23 lldt
J:lfB 24 .zld3 'ire 1+ 25 �h2 'lt'xf2 26 J:lh3
Wf4+ 27 �g1 Wet+ 28 �h2 'iff4+.
21 1t'e4
•••
This is okay, but an interesting possibility
11
The Petroff Defence
is 2t...1i'f6 22 l:tet �5 23 1Wf3 «i>ffl24 �f4
(Anand-Karpov, Prngue 2002). Anand now
gives 24...1Wxh4!? 25 .i.xf7 1i'g4 26 'ifxg4
.i.xg4 27 .i.d.5 �e7 28 .i.xb7 l:b8 as offering
Black enough compensation.
2211'g3
Black can also defend against 22 �f4 �eS
23 dxeS 1i'xc4 24 exd6 1i'xc3 25 l:tf1 .i.e4!
(not 25...D.d8? 26 �xg6 fxg6 27 d7 1i'c6 28
l:tdl bS 29 hS! gxhS 30 .xhS when White
has a huge advantage, Mortensen­
M.Andersson, Sweden 2003) 26 �6 i.d3 27
'ilfOI fxe6 28l:td1 1i'xa3 29l:txd3 Wet+ 30
'it>h2 1i'c6 31 1i'xc6 bxc6 32 l:l.c3- the ruok
ending is drawn.
22 ll:!a5 23 .ia2 ...d3 24 llf1 •xg3 25
•••
fxg3 lieS 26 hS! .id3
Black avoids the obvious double attack:
26...i.xh5? 27 :5 b6 28 l:.xhS :Xd 29 l:lf5
when White should win.
27 .txf7 + ¢>f8
27...'it>h7?! is too passive: 28 l:.f3 ..ie4 29
l:.e3 d5 30 �f4 �4 31 l:.e2 �xa3 32 .ixd5
i.xdS 33 �d5 and White has good winning
chances.
28 llf2
28 . . .oh7?
A fatal slip. After 28...l:.xc3 29 ..ig(rr �g8
White has no more than a draw.
29 .td5 llf8
The problem with 29...:Xd is 30 l:.f31
J:ct+ 31 Wh2 i.c4 32 2:f7+ Wd8 33 .i.xc4
%lxc4 34 :Xg7 :Xd4 35 g4, when White's
12
kingside pawns decide the game.
30 ll:!f4 .i.h7 31 .ta2 �7 32 ll:!e6 llxf2
33 �xf2 .igS 34 �f8+ �es 35 .bg8
�xf8 36 .idS!
White has kept the extra pawn and now
dominates the kniJ.,rht: Black's position is
hopeless.
36...�e7 37 �e3 ¢>f6 38 ¢>f4ll:lc6 39
.i.xc6 bxc6 40 c4 �e6 41 g4 ¢>f6 42
g5+ hxg5+ 43 Wg4 1 -0
Game 3
Anand-Shirov
Moscow2001
1 e4 e5 2 ll:!f3 ll:!f6 3 ll:!xe5 d6 4 ll:!f3
ll:!xe4 5 d4 d5 6 .id3 ll:!c6 7 0-0 .ie7 8
c4ll:lb4 9 .te2 0-0 1 0 �c3 .i.e&
1 1ll:le5
Apart from 11 .i.e3, which is studied in
Games 5-6, there arc two alternatives to note:
a) 11 .if4 cS 12 l:lc1 .i.f6 13 �bS? (this
expedition fails to some brilliant resources;
White should prefer 13 11'a4 �xc3 14 bxc3
�c6 15l:tab1 'ilfd7 with an unclear position)
13 ...dxc4 14 �7 �5! 15 �xc6 fxc6 16
.i.g3 (Adams-Shirov, Dortmund 1998). Shi­
rov now analysed 16...�xg3 17 hxg3 bS! HI
a4 cxd4 19 �xd4 'tWb6! 20 �f3 (the lx:autiful
point is 20 �xbS? ..xf2+! 21 �xf2 .i.d4
mate!) 20... .i.xb2 21 l:.bt �d! 22 ..c2 �xa4
23 l:.xb2 �xb2 24 '1Vxb21lac8 when Black's
excellent pawns promise a clear advantage.

b) 11 a3 is safe but unthreatening. For ex­
nrnplc, 1 t...lL'lxc3 12 bxc3 tL'lc6 13 cxd5
.lxd5 14 c!Od2 tL'la5 15 .i.d3 b6 16 •c2 h6
17 .i.b2 .i.b7 18 lbe4 b5! 19 a4 (tf 19 .i.xb5
t'5 20 "Wa4 c6 21 .i.c4+ lL'lxc4 22 •xc4+ •d5
2.3 llk12 .i.g5 24 'lfd3 c5 Black's active bish­
ops provide ample compensation) 19...a6 20
uxb5 axb5 21 .i.xb5 fS 22 tL'lc5 .i.xcS 23
(lxcS •d5 24 c6 lL'lxc6 25 •d3 ll:le5 26
Vxd5+ .i.xd5 with an equal position, Lcko­
Krnmnik, Dortmund 1999.
1 1 ... c5
This early break gives White the chance to
ntlvance in the centre. The alternative 11...f6
is the subject of Game 4.
1 2 l0xa4
If 12 .i.e3 cxd4 13 .i.xd4 lL'lxc3 14 bxc3
li:k6 15 lL'lxc6 bxc6 16 .a4 cS 17 ..ie3 rl4
Ulack equali.<;es comfortably acconling to
Yusupov.
12 ...dxa4 1 3 d5 .i.e&
Black must retreat since after 13.....id6?!
14 a3 i.xeS 15 axb4 ..t£5 16 bxc5 White's
pawns are too strong. Baklan-Timman,
Neum 2000 continued 16, ..•c7 17 g3 11fxc5
IK i.e3 •d6 19 c5 •f6 20 �3 ..Lb2 21
J:a41 (Baklan pointed out that 21 lLb1 ..id4
22 'ifxb7 ..ih3 23 .l:tfd1 ..ixe3 24 fxe.l Wc3 is
just unclear) 21..ie5 22 11fxb7 ..th3 23 Act
aS 24 d6 and the pawns promise White an
obvious advantage.
14 a3 �6 1 5 •c2
Anacking the pawn Js White's natural
3 fDxe5: Th e Main Line with 8 c4
plan. 15 f4 was tried in Anand-Lcko, Leon
2001, the game being level after 15...f6 16
tL'lg4 .i.xg4 17 .i.xg4 fS 18 .i.e2 .i.f6 19 l:tbt
i.d4+ 20 Wh1 tL'lc1 21 b4 b6 22 l:tb3 lbes 23
i.e3 i.xe3 24 .l:txe3 llkt6 25 l:tb3 .f6 26
bxc5 bxc5 27 11fa1 ..xa1 28 llxa1 Wfl.
15 f3!? is more enticing. Motylev analysed
15...•c7 16 i.f4! i.d6 17 fxe4 l:tc8 (of
course the idea is 17..ixe5 18 d6) 18 .a4
l:tc7 19 tL'ld3 l:txe4 20 .i.xd6 .xd6 21 ..if3
with a slight edge for White.
1 5...f6
Black threatens the knight at a moment
when it has no choice but to rctteat. Instead
15....i.d6 16 f4 f6 gives White the additional
option of 17 .xe4!? l:te8 (if Black takes the
piece with 17... fxc5 18 fxe5 .l:te8 White has
strong passed pawns and an attack as com­
pensation after 19 .i.d3 rfi 20 c6) 18 .i.h5
Ae7 19 i.fl+ (Morgado assessed 19 lL'lf7
'1Vc7 20 lL'lxd6 l:txe4 21 lL'lxe4 as unclear)
19...Wh8 20 tL'lg6+ hxrfi 21 'IVxrfi i.g4 22 fS!
l:txf7 23 Vxf7 .c7 24 ..xc7 ..ixc7 25 .tf4
and Black's minor pieces cannot find any
activity so White has an edge.
1 6�
The play is certainly complicated, but
White seems to have the better of it.
1 6.....d6
In Topalov-Shirnv, FIDE World Cham­
pionship, Moscow 2001 Black tried 16..id6
17 f4 cxf3 (or 17....txg4 18 .txg4 fS 19 .i.e2
.te7 20 ..ie3 ..i.f6 21 l:tadt and White's ad-
13
The Petroff Defence
vantage is small but defmite) 18 i.xf3 'ifc7
19 g3 i.d7 20 liJt2 bS 21 b3 :ae8 22 'iPg2
i.eS 23 :b1 i.d4 24 i.d2 and White re­
tained the usual nagging edge.
1 7 f3
White tries to destroy his opponent's cen­
tre. Anand also analysed 17 g3 fS 18 i.f4
'ifb6 19liJe3 i.f6 (not 19...g5?1 20 i.eS f4 21
gxf4 gxf4 22 'iPh I! and it is White who at­
tacks along the g-file) 20 h4 ...xb2 21 'ifxb2
i.xb2 22 :a2 when White ha.c; some com­
pensation but no definite advantage.
17 f5
... 1 8 lbf2 .i.f6
Black is more interested in the attack than
mere pawns.
19 fxe4 .i.e5 20 h3 .i.d4 21 e5!7
Similarly, White would rather return the
pawn than allow the fB-rook into the game.
21 .....xe5 22 Wh 1 .i.d7
Black's queen bishop settles for a modest
square. 22 .f4 gains the £5-squarc but only at
..
the expense of surrendering g4: 23 liJd3 i.£5
24 i.g4! (not 24liJxe5?! i.xc2 25 i.d3 i.xd3
26liJxd3 gS 27 h4 h6 28 hxgS hxgS 29 g3 bS,
when White's centre collapses) 24... g5 25
.id2 and White retains the initiative.
23 ltld3 .i.a4 24 •xa4 •xe2 25 .:l.f3
.:tae8
Again this is simple, sensible development.
Black had a tricky try in 25...b5 26 cxbS
liJb4!? 27 axb4 c4. The idea is to kick the
knight and follow with ...1ict+ and ...i.eS+,
but White keeps control and a clear advan-
14
tage with the calm 28 'ifa3.
26 .i.f4
Now this simple move is enough to con­
finn a solid advantage.
26 ...h6 27 .i.d6!
Not 27 :et? ...xcl+ 28 liJxel :Xet+ 29
Wh2 gS and Black escapes with a draw.
27....:tf6 28 lbf4 •e4 29 ltle6 .:texe6 30
dxe6 •xe6 31 .i.g3 .i.xb2 32 .:l.e1 •f7
33 .i.h4 .:I.e& 34 .:Xe6 ••e6 35 •c2 .i.d4
36 •xf5 1i'xc4 37 Wh2 ••2
If Black tries 37...'ifct White wins with 38
'ifd5+Wh7 39 ...e4+Wg8 40.te71.
38 .i.g31
Covering eS from checks ensures the win.
38 1i'd1 39 .:n "tlrb3 40 1ie4 1ib5
.•.
If 40...i.f6 then 41 :xf6! gxf6 42 'ifg6+
forces mate. For example, 42...Wh8 43
'ifxf6+ Wh7 44 ...f5+ and the bishop will
soon join the attack with check.
41 11Ve6+ Wh7 42 'iff5+ Wg8 4a
11Vc8+ �h7 44 .:tf8 1 -o
Game 4
Leko-Grischuk
Wijk aan Zee 2002
1 e4 e5 2 ltlf3 ltlf6 3 ltlxe5 d6 4 l£lf3
lbxe4 5 d4 d5 6 .i.d3 lbc6 7 0-0 .i.e7 8
c4 lbb4 9 .i.e2 0-0 1 0 lbc3 .i.e& 1 1 ltle5
f6 1 2 ltlf3
This simple retreat causes Black the most
problems.

[f 1 2 .i.g4, variation 'b' is the most accu­
r:ue:
a) 12 ... .i.c8 13 .i.xc8 llxc8 l4ltif.3 c6 (or
14...c5 IS 1i'e2 cxd4 16 ltixd4 lle8 17 �e3
i.cS 1 8lLd1 and White has an edge) 1 5 lle1
li_)xc3 1 6 bxc3lDa6 1 7 llbl llc7 1 8 c5 lld7
19 .i.f4 and White was a bit better in Frcssi­
net-Brodsky, Bucharest 200 1 .
b) 1 2... .i.xg4 1 3 ltixg4 f5 1 4 ltie3 (1 4 a3
lbc6 1 5 cxdS ltixd 16 bxc3 'irxJS 17 ltie3
1Wf7 is unclear) 1 4...dxc4 1 5ltixc4 c5 16 'iVb3
�h8 1 7ltieS 1i'e8 (Black has the interesting
nlternativc option of 1 7...cxd41? 18 ltixe4
fxe4 1 9 ltif7+ :Xf7 20 11t'xf7 'iVe8 with un­
balanced play) 18 a3 llk6 1 9ltixc6 1fxc6 20
ll\xc4 1i'xc4 (Minakov-Moq,rado, correspon­
dence 1 999) and here Morgado gives the
l'lJUalising line 21 i.d2 1fxd4 22 'iVxb7 'iVxd2
23 1i'xe7 c4.
12 . �h8
. .
Stepping away from any future trouble on
the a2-g8 diabJOnal.
12...c5?1 is premature: 1 3 i.e3 :C8 14
dx.cS i.xcS 1 S .i.xcS D.xcS 16 1Wb3 aS 1 7 a3!?
(17 lladt is less clear: 1 7...'iVe7 1 8lDa4 i.f7
19 11'e3 llc7 20 a3 ltia6 21 cxdS ltid6 22
1Wxe7 llxe7 23 .i.xa6 bxa6 is very messy but
lllack seems to have compensation, Khalif­
man-Karpov, Denpasar 2000) 17 ...ltixc3
(17.. dxc4 allows White to play a very con­
. structure gives him a small advantage, Gris­
vincing queen 'sacrifice': 18 ltixe4 cxb3 1 9
�xeS .i.f7 20 axb4 axb4 21 i.d 1 and White
has a healthy matc..'t'ial advantage) 1 8 1fxc3
3 et:Jxe5: The Main Line with 8 c4
lDc6 19 b4 axb4 20 axb4 d4 21 11'b2 :Xc4 22
i.xc4 .i.xc4 23 llfd1 when Morgado claims a
clear advantage for White.
A more worthwhile alternative is 1 2..lle8.
For example, 1 3 .i.e3 i.fB 1 4 a3 ltixc3 1 5
bxc3ltic6 16 cxd5 .i.xdS 1 7 c4 .i. f7 18 'iVc2
ltiaS 19 i.d3 g6 20 i.d2 (Morgado suggests
White can gain an edge with 20 h4!? cS 21
dxcS 'iVc7 22 .i.d4 i.xc5 23 i.xf6 1ff4 24
i.g5 1Wg4 25 ltih2 ifd) 26 'iVc3 ltixc4 27
l:lael WdS 28 �r4) 20...c5 21 dS (fimman
assesses the variation 21 1Wa4 llk6 22 d5
lDti4 2l ltixd4 cxd4 24 .i.b4 b6! as equal)
21 ... b5 22 .i.xaS 'iVxaS 23 cxbS LdS 24
lDd2 llad8 2S llfdl �g7 with a double­
edged position, Adams-1imman, Wijk aan
Zl.'C 2001 .
1 3 a3
White has various alternatives:
a) 1 3 cxdS is too simple: 1 3...ltixc3 14
bxc3ltixdS 1S i.J3 cSI (but not 1 S...ltixc3??
16 'IVct with a winning double attack) 1 6 c4
ltib4 17 d5ltixd3 18 11'xd3 i.d61 (of course
taking no c6 would still lose the queen) 19
1Wb3 i.g4 20 h3 .i.hS 21 a4 V2-1/z Anand­
Adams, Dortmund 2001 .
b) 1 3 h3 is a bit too slow: 1 3. .£5 1 4 a3
.
ltic6 1S cxdS (1 5 ltixdS?I was ttied in lvan­
chuk-Ponomariov, FIDE World Champion­
ship, Moscow 2002, but after I5 ...Ld5 1 6
cxdS 'IVxdS 17 'iVa4 i.f6 1 8 lld1 llad8 1 9
i.e3 f4! 20 i.xf4ltixd4 21 ltixd4 .i.xd4 22
i.e3 cS Black was already a bit better)
1S ....txd5 16 .i.f4ltixc3 17 bxc3 .i.d6 Black
has comfortable equality.
c) 13 lle l, however, is a reasonable alter­
native to 13 a3. 13 ... c5 1 4 .i.e3 f5 1 5 a3 f4
(Grischuk assesses 1 S...ltixc3 16 bxc3 llk6
17 cxdS i.xdS 1 8 dxcS f4 1 9 i.d4 as slighdy
better for White) 16 i.d2ltic6 17ltixe4 dxe4
18 dS i.xd51 1 9 cxdS cxf3 20 dxc6 fxc2 21
11'xe2 bxc6 22 i.xf4 and White's better
chuk-Motylcv, FIDE World Championship,
Moscow 2001 .
13 ...�c3 14 bxc3 �
75
Th e Petroff Defence
1 5ll\d2
White must try to conttol c4. Instead 15
cxd5 .i.xd5 16 .i.e3 �5 17 11a4 c6 18 ...c2
�4 gives Black easy equality.
1 5...f5
Making f6 available to the bishop.
15...�a51? is interesting but probably not
quite goo d enough to equalise: 16 cxd5
.i.xd5 (or 16...1ixd5 17 .i.b2 'irg5 18 c4 .i.d6
19 .i.c3 and White has the easier game) 17 c4
.i.f718 .i.b2 and now:
a) If 18....i.d6 White must be a bit careful.
19 ..c2 (supporting c4 is essential; not 19
.i.d3?1 .i.f4! 20 dS c6 and White's centre col­
lapses disasttously) 19....i.g6 20 .i.d3 .L:d3
21 'irxd3 and White has an edge (Belov).
b) 18...5 19 .i.c3 cS 20 dS .i.f6 21 1ic2 b6
22 .i.d3 .i.xc3 23 ..xc3 �b7 24 �f3 llXI6
25 ILleS and the protected passed pawn
makes White's poslbon preferable,
Kasimdzhanov-Adams, FIDE World Cham­
pionship, Tripoli 2004.
1 6 lle1 ..tf6 1 7 a4
Trying to gain space on the quecnside. 17
�b3 has also been tried: 17...b6 (an intrigu­
ing alternative is 17... dxc4 18 ILleS .i.c8 19
.L:c4 1Llxd4!? 20 cxd4 ..xd4 21 'irxd4 .i.xd4
22 .i.e3 .i.xa1 23 D.xa l with equality) 18 .i.f3
.i.gS 19 cxdS .i.xdS 20 .i.f4 D.c8 (or 20...1Lle7
21 i..es c6 22 .L:f6 :Xf6 23 IL!ct �g6 24
00 h6 - Leko-Kramnik, Dortmund 2000 -
and now C.alkin suggests that White gains an
edge with 25 IL!b4 .i.xf3 26 1Wxf3) 21 .i.eS
16
11i'd7 22 .i.xf6 D.xf6 23 �cl :C6 24 :Xc6
1ixc6 25 IL!d3 l%e8 26 h3. Here the game is
level and a draw was agreed in Galkin­
Motylev, Dubai 2001.
1 7.....tg8 18 c5 lieS 1 9 l£if3 h671
This is unnecessary. The direct 19...b6!? 20
cxb6 axb6 21 .i.bS :Xet+ 22 ...xe1 �5 is
l."qUal.
20 llb1 llb8 21 ..td3 llxe1 + 22 •xe1
'ird7 23 ..tf4 lle87
This overlooks a nasty trick. 23 ...i..h7 24
.i.bS lte8 25 •dl restricts White to an edge.
24 .d2 g5
This desperate lunge is forced. If 24...b6?
then 25 cxb6 axb6 26 .i.xh6! wins immedi­
ately.
25 llxb71
Crucially, this undetmines the c6-knight.
25 ...llc8
If 25 ..gxf4 White wins with 26 1i'xf4 i..g5
.
(or 26....i.d8 27 Wxh6+ i..h7 28 .i.bS) 27
'ifxc7 'ifxc7 28llxc7 Q..eko) .
26 h47
This is a clever idea but there was an in­
stant win with 26 .i.bS! gxf4 27 D.xa7 as
Black cannot escape the pin.
26 ...gxf4 27 •xf4 ..tg7 28 •xf5 'ird8?
Now White is wmrung again. After
28.....e8 29 1if4 lie7 30 .i.b5 We6 31 .i.xc6
...xc6 32lLa7 White has only a small advan­
tage.
29 l£ig511 hxg5 30 hxg5 Wf8
This is the only move to save the king.

3 1 1th3+ -*.h6 32 .i.f5
The simple 32 gxh6 l:r.e8 33 llxc7 :Ct+ 34
.tn etlc7 35 llxa7 is also winning.
32 ...J:le8 33 J:lxc7 J:le1 + 34 Wh2 liJe7 35
g41 .i.e6
l.cko supplies the neat winning variation
'5. .l:te6 36 �g2 :a6 37 c6!? llxc6 38 l:r.xc6
. with equality) 2 l ....i.xe2 (2t....i.xg5?! is not
Q'\xc6 39 'irxh6+ 1Wxh6 40 hrxh6.
36 Wxh6+ 1i'xh6+ 37 gxh6 liJxf5 38
gxf5 .i.xf5 39 J:lf7
White's pawns win easily.
39....i.g6 40 J:lxa7 J:lc1 41 Wg3
:Xc3+ 42 �4 .i.h5 43 �5 llf3 44 c6
.tg4 45 a5 J:lxf2 46 a6 llc2 47 llc7 1 -0
P""'"-----.
Callie 5
Shirov-Gelfand
Leon 2001
1 e4 e5 2 liJf3 liJf6 3 lbxe5 d6 4 liJf3
/li)xe4 5 d4 d5 6 .i.d3 .i.e7 7 0-0 liJc6 8
c4 liJb4 9 .i.e2 0-0 1 0 liJc3 .i.e& 1 1 .i.e3
.if5
11 ... f5 is a h'Dod alternative. flor example,
12 a3 etlxc3 13 bxc3 etlc6
and now:
a) 14 cxdS .ixdS 15 c4 .ixf3 16 .ixf3 f4
17 .idS+ �h8 18 .ict etlxd4! 19 l:tb1 (19
i.b2 does not alter the assessment: l9...c5 20
.ixd4 cxd4 21 .i.xb7 llb8 22 .idS .icS is
l"lJUal) 19....ic5 20 l:txb7 f3! 21 .ixf3 etlxf3+
22 gxf3 112-1/z Hubncr-Yusupov, Rotterdam
191'18.
3 liJxe5: The Msin Line with 8 c4
b) 14 111'a4 has also been tried but Black
has no real problems: 14...f415 .i.d2 �h8 16
:ab1 l:tb8 17 l:tfc I dxc4 18 .i.xc4 .i.g4 19
i..e2 a6 20 h3 .ihS 21lbg5?1 (now the tactics
work out in Black's favour; White should
have settled for 21 etlh2 .i.xe2 22 l:txe2 'ifdS
the answer: 22 .i.xhS f3?1 23 .i.xgS 1i'xg5 24
.i.xf3! l:r.xf3 25 1i'xc61 l:txf2 26 l:r.e8+ l:lfB 27
l:r.xb8 9e3+ 28 �h 1 l:r.xb8 29 1i'xc7 and
White should win this ending) 22 etle6 'irdS
23 l:r.xe2 (if 2.1 etlxfB f3 Black is already much
better because 24 gxf3 .i.xf3 25 etle6 1i'f5
gives a winning attack) 23... f3 24l:r.ce1 fxg2!
and Black was clearly better in Rohde-
Scirawan, Estes Park 1986.
12 1Vb3
Increasing the central pressure. 12 l:r.cl is
the subject of Game 6, while a worthwhile
option is 12 etlt:S. flor example, 12....i.f6
(Short assesses 12...etlxc3 13 bxc3 etlc2 14
l:r.c I etlxc3 15 fxc3 .i.e4 as unck-ar) 13 g4
.i.e6 14f4etlxc3 15 bxc3 etlc6 16 .i.f3 .i.xeS
17 dxeS d4 (White has the initiative after
17...dxc418 fS .ic8 19 .i.cS 1i'xd1 20 l:r.axd1
l:r.e8 21 l:r.fcl) 18 cxd4.ixc4 19 dS etle7 20
.i.cS .ixfl 21 �ft 'ird7 22 'lfb3 and
White's bishops and impressive centre pro­
vide ade<Juatc compensation, Short-Anand,
Amsterdam 1993.
1 2...c6
Choosing to bolster the centre but
12...dxc4 is also acceptable: 13 .i.xc4 aS t 4
etlxe4(14a3 can lead to a neat draw by repe­
tition after 14...etld2! 15 .ixd2 .i.c2 16
.ixf7+ l:r.xf7 17 1fe6 .ifS 18 '1Vb3 .ic2 19
'ire6) 14....ixe4 15 a3 a4 16 'ifdl etlc2 17
.J:r.cl etlxc3 18 fxe3 .id6 19 ..tbS 'iff6 20 l:r.c3
'irh6 is unclear, Barua-Manesh, Raipur 2002.
1 3 cxd5
White has various alternatives:
a) 13 c5 etlxc3 14 bxc3 etlc2 15 1fxb7
etlxat 16 l:r.xa1 1fe8 17 'ira6 f6 18 etle1 'ireS
(Kramnik states that White has compensa­
tion after 18 ...1Wd7 19 etld3 .ixd3 20 .ixd3)
19 'ffxc8 .ixc8 20 .i.f4.idS 21 etlc2 l:r.eB 22
17
The Petroff Defence
.i.d3 aS is unclear, Topalov-Kramaik, Til­
burg 1998.
b) 13ll:kS a5 14 g4 .i.e6 1S tbxe4 dxe4 16
a3 f6 17 axb4 fxeS 18 dS cxdS 19 cxdS .i.xdS
20 .i.c4 .i.xc4 21 'ii'xc4+ 'ith8 22 bxaSilxaS
23 'l'xe4 .l:lbS and the simplifications have
led to CXjUality, Topalov-Shirov, FIDE World
Championship 2001.
c) The wild option is 13 g4!? .i.xg4 14 cS
(the point of 13 g4!? is that Black no longer
has the resource ...tbxc3, bxc3, ...tbc2) 14...aS
1S a3 tba6 16 'ii'xb7 tbxc3 17 bxc3 tbc7 18
Zlab1 tllbS with an unbalanced position,
Topalov-Shirov, riDE World Championship
2001.
1 3 ...lDxc3
t3...cxdS 14 l:act aS 1S tLla4llb8 16 tbcs
.i.d6 17llfd1 (Shirov-Topalov, ADE World
Championship 2001) is a safe equalising line.
Now the simplest for 8L1ck is 17..lle8 18
.i.bS.I:le7.
14 bxc3 .!bxd5
1 5 1ixb7
White can play safe with 1S c4 tllxe3 16
fxe3 'ii' c7 17 .i.d3 .i.xd3 18 'lf'xd3 .llfe8 with
an equal game.
1 5 lL'lxc3 1 6 ..i.c4 llb8
••.
If 16 ....te4? then 17 ll:ks llb8 18 1i'xa7
lias 191i'b7llb8 20 11fd7 and White's queen
escapes with a dear extra pawn.
1 7 1txc6 .i.e4 1 8 ..i.xf7+ :Xf1 1 9 1t'xc3
..i.b4
Black also has compensation after
18
19....i.xf3!? 20 gx£3 .td6 21 9c6 llb6 22
'lfe41i'b8.
20 'irc4 .i.d5 21 •d3 :Xf3!? 22 gxf3
...d7 23 l:lfb1
Gelfand mentioned 23 .i.f4!?llfB 24 .i.g3
.l:lx£3 2S 'ii' e2 with a murky position.
23 ...1th3 24 ..i.f4 l:lf8 25 l:lxb4
Or 2S .i.g3 aS 26 a3 .l:lx£3 27 11ft 'irh6!
28 ..e2 (28 axb4 allows a dear draw after
28...llxg3+1 29 fxg3 'ire3+ 30 'l'f2 'l'e4 and
White can only avoid the mate by aUowing a
perpetual check) 28...1rhs 29 llb2 .i.f8 and
Black still has dangerous play- Gelfand.
25 .. .l:lxf4 26 l:lb8+ �7
With 27...llg4+ threatened White seems to
be lost. However...
27 l:lb7+ !
Now i f Black takes the rook, 'irb3+ wins
the bishop.
27 Wg8 2Bl:lb8+ %-Y.z
•••
Game 6
Kotronias-Motylev
Moscow2004
1 e4 e5 2 l£lf3 lL'!f& 3 lDxe5 d& 4 lL'lf3
lL'lxe4 5 d4 d5 6 .id3 lL'lc6 7 0-0 .i.e7 8
c4 lL'lb4 9 ..i.e2 ..i.e&
10 lL'lc3
Now this game simply transposes ro the
main line. However, the move order with
9....te6 gives White interesting additional
options:

a) 10 'i'a# is tempting but Black can de­
fend: 10...'i'd7 11 'i'xd7+ �d7 12 ltk3
J:lhd8! (this is the way to equalise: after
12...ltlxc3 13 lLk5+ �e8 14 bxc3 ltk6 15
cxd5 .ixd5 16 .ic4 Shirov claims an edge
for White) 13 a3 lDxc3 14 lDe5+ �e8 15
bxc3 lDc6 16 lbxc6 bxc6 17 cxd5 .ixd5 18
J:ld �f8 19 .if4 .id6 20 .i.xd6+ cxd6 21
.id3 g6 22 f3ltdb8 231leb1 �e7 and in this
dead level position a draw was agreed, Rau­
sis-Bacrot, France 2003.
b) After 10 cS!? Black must defend 3b'llins t
ideas with 'i'b3.
b1) The natural 10 ...0-0?! runs straight into
!rouble after 11 'ifb3! aS 12 aJ lDa6 13
1i'xh7. The attempt to trap the queen with
13.)tlaxc5 14 dxc5 lDxc5 15 1i'c6 llb8 is
met by 16 .ig51. White already has a pleasant
;tdvantage, but 16... f6?1 just makes things
worse: 17 .if4 .i.d6 18lDd4 .ifl19 b4 axb4
20 axb4 llxb4 21 .ixd6 cxd6 22 lDb5 and
White is winning (Nataf).
b2) tO...aS defends the knight but White
keeps the initiative: 11lDes f6 12lDd3 0-0 13
li:lf4 .ifl 14 .ig4! g6 15 lDe6 .ixc6 16
i.xe6+ �g7 17 a3lDc6 18 lDc31le8 (Nataf
points out the clever trick 18 ...1la6? 19
lLlxd51lDxd4 20 Wxd41lxe6 21lDf411le5 22
1i'xc5Q 19 lDxd5 Lc5 20 dxc5 llxe6 21
i.f4 tDes 22 llct with an cdgt:, Nataf­
Topalov, Cannc..'S 2002.
b3) 10...ltk6 11 .ib5 .if6 (fighting for
control of d1c key e5-squarc; 11...0-0 allows
White to seize ilic initiative in simple fashion
with 12 .ixc6 bxc6 13lDc5 .id7 14 f3lDg5
15 lDc3) 12 Lc6+ (after 12 .if41? 0-0 13
i.xc6 bxc6 14 lDes Wbs 15 Wet .ixc5 16
i.xe5 f6 17 .if4 White was a bit better in
i\dams-Bacrot, Cap d'Agde [rapid] 2003)
12...bxc6 13 11'a4 .id7 14 .if4 0-0 15 lDcJ
(trying to grab ilie c7-pawn with 15 11'a5
llllows Black counterplay: 15...'1'1>8 16 b3
l:te8! 17lDe5 -if 17 Lc7 'itb7 18 .i.e5 .idS
19 1i'a3 'i'b5 White's awkward development
�vcs Black enough compensation -17...11'b7
1 8 f3 lbg5 19 .i.xg5 .ixg5 20 f4 J..f6 21lDc3
3 lC.xe5: The Mein Lin e with 8 c4
lle7 and ilie position is finely balanced)
15.. ll.b8 16 .xa7 (defending ilie b-pawn
wiili 16 llab1 has other drawbacks: 16. ..if5
.
17 llbc1 gS!? 18 .ig3 g4 19 tDes llxb2 20
lDxe4 .i.xe4 21 lDxg4 .i.g7 with an unclear
game) 16 ...1lxb2 17 lDa4 (or 17 ltlxe4 dxe4
18 Wa3 llc2 19 lDe5 .ixe5 20 .ixe5 f6 21
.if4 .ig4 wiili a messy position) 17...1lb4 18
a3 (White must be careful wiili his a4-knight;
instead 18 llab1 ? 'i'bsl 19 Wxb8 llfxb8 and
White was losing in Al-Modiahki­
lordachescu, Dubai 2004) 18.. ll.c4 19 ltlb2
llxd4 20 .i.e3 llc4 21 lDxc4 .ixa1 22 llxa1
dxc4 with an 'unclear' verdict from Ior­
dachescu.
10 0-0 1 1 .te3 .tf5 1 2 llc1 dxc4 1 3
.•.
.txc4
1 3 ...c6
Preparing a retreat to d5 for the b4-knight.
White can achic..-ve an edge after 13...lDxc3:
14 bxcJ lDc6 15 .idJ .i.xd3 16 9xd3 11'd7
171lb1 (instead 17 c4 J..f6 181lfd1 llfe8 19
'lfbt b6 20 h3 lDc7 was only equal in
G.Kuzmin-Hu7man, Rcthymnon 2003)
17...b6 181lfe1 llfe8 19 .if4.
14 lC.e5 lC.xc3 1 5 bxc3 te.d5 1 6 1ff3
Also interesting is 16 1ib3 f6 17 ltlf3 b5
(Ivanchuk suggests that 17 ...1rd7!? leads to
an unclear position after 18 l:lfet �h8 19
.id2 .id6 20 .in lDf4) 18 �e2 �hs 19
.id2lDb6 20 life I 9d7 21 c4 bxc4 22 .i.xc4
lDxc4 23 9xc4 l:fc8 24lDh4 i.f8 and White
has a very faint edge, Kir.Georgiev-lvanchuk,
19
Th e Petroff Defence
Dcbrecen 1992
1 6 .i.e& 1 7 .i.d2
.••
The other way to play is 17 .i.d3 when
Black has three main replies:
b) 17 ....i.f6 18 .i.d2 .d6 19 llfet lladS
20 c4!? is Nijboer's suggestion (20 h4 c5 21
h5 cxd4 22 cxd4 &i:k7 21 h6 lDg6 24 hxg7
.i.xg7 led to a messy position in Nijboer­
foressinet. Leon 2001). After 20. .�7 21
. 20 .i.b3
.i.a5 lieS 22 .i.c3Q)g6 White's space advan­
rnge makes his position preferable.
b) 17...ltlxe3 1S fxe3 .i.d619lL!c4 .i.c7 20
e4 f6 is unclear - Black's bishop pair com­
pensates for White's centre.
c) 17... .i.d6 tS .i.d2 'lh14 19llfet ltlf6 20
a4ltlg4 21 .i.f4 .i.dS 22 .i.g3 lLixeS 23 clxe5
e'h6 (Hubner pointed out the blunder
23...'itxa4? 24 .i.xh7+1 *xh7 25 'lh15+ WgS
26 exd6) 24 .i.e4 .i.xe4 25 "ifxe4 .i.c7 is
equal, Klimov-Smikovski, Toljatti 2003.
1 7 16
•••
Less accurate is 17 ... .i.gS 1S .i.xg5 'itxgS
19 llfel llaeS 20 g3 "iff5 21 "ifdt ltlb6
(White has a promising initiative after
2t..lle7
. 22 .i.d3 1i'f6 23 c4 ltlb6 24lle4-
Topalov) 22 .i.d3 "iff6 23 1i'c21!f> 24 llbt
.i.cS 25 lle2 lie7 26 llbe1 and White had
some prcsswe in Topalov-Akopian. Linares
1995.
1 8�
Or tSlLld3 1i'd7 19llfe1 .i.d6 20 h3 .i.f7
21 .i.b3llae8 22 "ifg4llxet+ 23llxc1 lids
24 Wxd7 llxd7 25 lL!es Le5 26 dxe5 fxe5
20
27 llxc5 ltlf6 and despite White's creative
efforts the position was stiU level in Topalov­
lvanchuk, Novgorod 1996.
18 11'd7 1 9 h3lQf4
..•
Exploiting the loose bishop on c4. Black
could also play the simple 19... b5 2U .i.d3
llfcS 21 life1 .i.d6 22 lL!e3 ltlxe3 23 .i.xe3
.i.d5 24 .e2 'itf7 with equality .
White ambitiously keeps the tension. The
quiet 20 .i.xe6+ ltlxe6 21 lL!e3llfeS 22 ltlf5
.i.fB is level.
20 .i.xb3 21 axb3 lDg& 22 llfe1 .i.d&
.••
23lQe31lae8 24lQf5lQe7
A sharp try is 24....i.f4!? 25 'W'g4 .i.xd2 26
lLlh6+ gxh6 27 11'xd7 .i.xe1 28 'trxb7 lle2
with an interesting and unbalanced position.
25lQxd6 ••d6 26 c4lQg6
The greedy 26...1Vxd4?! allows a deadly
pin: 27 .i.b4llf7 28lle6.
27 .,5 ••d4 28 .i.b4 l0114
Black has to play creatively to survive. If
2S..Jlxcl+? 29llxe1 lidS 30 ..e6+ Wh8 31
.eS+ .!DEB 32 "fif7 Black's weak back rank
costs him the game.
29 'lrh5 g& 30 1ra5
Motylev points out that 30 .c5 'trxc5 31
.i.xc5 llxet+ 32lLet lids 33 .i.xa7 Wf7 is
equal.
30 Jle51
••
Once again 30....1Let+ leads to back rank
problems: 31 llxe1 llf7 32 .i.c5 'lrh2 33
.d8+ Wg7 34 Wh2 (one clever preparatory'

move is required; instead 34 l:le8?1 allows
Black to escape with a draw with 34...'iVct+
JS 'it>h2 �f3+! 36 gxf3 'iVf4+- the king must
!(II to hl and allow a perpetual because mov­
in� to the g-file allows ...'iVgS+ and 1fxc5)
_\4...�£5 35 lleH and White has a dangerous
111tack (Motylev).
31 llcd 1 !
Now White has t o be careful. I f 31 l:lxeS?!
rxeS 32 �xfB 'iVe4 33 �ft 'iVxg2+ 34 �e2
'i'c4+ 35 Wd2 �f3+ 36 �c3 'iVd4+ 37 'iPb4
'.f.>xffi Black's position is preferable: two
pawns and White's exposed king provide
more than enough compensation for the
l'XChangc.
31 ...1rxd1 32 •xeS fxe5
After 32 ...1Vxb3 Black's exposed king al­
lows White to draw with 33 'iVd6 l:lf7 34
l:tc!t+ <l;g7 35 l:lc7 gS 36 :Xf7+ �xf7 37
'i'e7+ '.ti>g6 3M '1Fe4+.
33 .lbd1 llf7 34 lldB+ �7 35 .i.d6 e471
This allows White's bishop too much
Hcope. Instead 3S...�f6 36 llc8 .J:ld7 37
i.xeS+ �f7 is equal.
36 .i.c5
Now White has a tiny edge.
36...b6 37 .i.d4+ ¢>h6 38 .i.e3+ ¢>g7 39
.ig5 �f6 40 g4 �h6
40 .. h6? is a horrible blunder: after 41 �cl
. �aS 15 cS c6 16 lle l �f6 17 �g4 i..cH 18
Black must lose a piece to avoid mate.
41l...li:k7 is reason.-.ble but White stiU has a
pull after 41 b4 cS 42 .J:le8 �6 43 bxcS bxc5
44 �xe4.
3 �xe5: Th e Main L ine with 8 c4
41 b4 llf3 42 lld7 + �f7 43 .i.e3 Jlxh3
44 :Xa7 ¢>16 45 hb6
After the alternative 4S l:la6 �S 46 l:lxb6
�e6 47 bS llh4 48 gS �£5 49 bxc6 �f3+
Black's countcrplay arrives just in time to
save the draw.
45...�5 46 .i.dB+ ¢>e6 47 ¢>g2 llb3
48 lle7+ ¢>d6 49 :Xh7 I.Oxc4 50 .i.g5
1.0&5 51 llg7 ¢>d5 52 .i.f4 lbxg4 53 :Xg6
lbxf2 54 lld6+ ¢>c4 55 ¢>xf2 llf3+ 56
¢>e2 llxf4 57 We3 llh4 58 llxc6+ Wxb4
59 lle6 Wc4 60 :Xe4+ llxe4+ 61 Wxe4
Y.z-%
Gamel
Kasparov-Anand
Unares2000
1 e4 e5 2 I.Of3 �f6 3 I.Oxe5 d6 4 lbf3
lbxe4 5 d4 d5 6 .i.d3 �c6 7 0-0 .i.e7 8
c4 � 9 .i.e2 o-o 1 o &3 b&
1 1 a3
11 cxdS aUows the interesting line
1t..if5!? t21i'b3 aS with unclear play.
11 �S is certainly direct bur Black can
defend: 11...�b7 12 .J:let (12 a3 was tried in
Shirov-Anand, Linares 2000, but led only to
a draw after 12...�xc3 13 bxc3 �c6 14 �f)
cxb6 axb6 19 a4 �4 20 �xf6+ 'iVxf6 21
�a3 �xa3 22 llxa3 �e6 23 'iVb3) and now:
a) 12 ...dxc4 13 �xc4 �xc3 14 bxc3 .!hiS
lS 'iVf3 c6 (Hracek-Yusupov, Bundesliga
21
The Petroff Defence
2000). Here Hracek suggests that 1 6 .LIJ is
worth an edge, the point being 1 6...ltlxc3?!
1 7 ltlxf7! 'ifdS 1 8 l:r.xe7 'ifxf3 19 �h6+ gxh6
20 .i.xh7+ Wh8 21 gxf3 .i.a6 22 .i.xh6 when
White is dearly better.
b) 12 ... c51? (the aggressive approach is
best) 1 3 .i.f3 cxd4 1 4 �xdS f5 1 5 �xe7+
1Wxe7 16 a3 �!? 1 7 �!? hxg6 1 8 1Wxc2
'ifh4 and after some very sharp play the posi­
tion remains unclear.
1 1 ...lL!xc3 1 2 bxc3 l0c6 1 3 cxd5 1txd5
14 �1
The simple 14 .i.f4 is only level after
14....i.d6 1 5 .i.xd6 cxd6 1 6 c4 •as.
14....i.b7 1 5 .i.d3 Aae8 1 6 c4
h is wisest to hit the queen now. If 1 6
'lrc2 Black has two reasonable replies:
a) 1 6...h617 i.h7+ Wh8 1 8 i.e4 (Kas­
parov-Kramnik, I.inares 2000) and here Kas­
parov analysed the equalising line 1 8 ... '1rh5
1 9 �5 i.d6 20 .i.f3 'ifh4 21 �xf7+ Wg8 22
ltlxh6+ Wh8.
b) The dangerous 1 6...'tth5!? is a good rea­
son to avoid 1 6 'ifc2 here: 1 7 J:r.bt �xd4!
(Black targets the h2 square) 1 8 cxd4 {White
must avoid the back rank trick 1 8 �xd4?
i.d6) 1 8... i.xf3 1 9 gxf3 i.d6 and Black has a
threatening attack (Rogers).
1 6...ifd8
If Black again tries 1 6 ...'1rh5 then White is
better placed to meet the threats to h2: 1 7 dS
i.d6!? 1 8 dxc6 l:r.xet+ 19 ..xel i.xc6 20
i.e21 i.xf3 21 .i.xf3 Wxh2+ 22 Wft 'ifh1+
22
23 We2 'ifh4 24 g4 hS 25 i.e3 hxg4 26 i.d5
(Kasparov) when White escapes the attack
and has a good advantage.
1 7 d5 lL!b8
1 8 .c2
Possibly this commits the quL>en too early
Instead 1 8 ltleSI? may be the best choice
Kasparov-H.Olafsson, Kopavogur (rapid)
2000 continuing 1 8...i.f6 1 9 .i.b2 g6 (or
1 9...c6 20 d6 g6 21 cS and White's pawns are
very strong) 20 'ifd2 ti)d7 21 �xd7 i.xb2 22
�xffi .i.xal 23 ltlxg6 l:r.xel+ 24 'ifxe1 hxgt)
25 .xa1 c6 26 dxc6 i.xc6 27 1i'c3 and Kas­
parov was dearly better. After 1 8 ... i.d6 1 9
i.b2 9£6 20 l:r.e3 White is building a strong
attack, the tactical justification being
20...lLe5?! 21 i.xeS i.xeS 22 'ifhs :CS 23
l:r.aet gS 24 h4 when White is already dose to
victory.
1 8...g6 1 9 �5 .i.f& 20 .i.b2 �7 21 f4
If 21 lDg4 then 2t..ixb2 22 Wxb2 hS 23
�3 �cS 24 i.c2 c6 is just unclear.
21 ...i.g7 22 .,2 lL!c5
This is fine but 22... c6!? 23 d6 ltXs 24
llad 1 lle6 with a hard fight ahead wa.o; also
reasonable.
23 .i.c2 •d& 24 1le3 f6 25 � llxe3 26
1txe3 ..i.c8 27 f5 ..i.xf5 28 hf5 gxf5 29
l0f2
The alternative 29 �h6+ also leads tO
equality after 29 .ixh6 30 11fxh6 ltle4.
.
29 ..� 30 lrure4 % -%
.
Here the players agreed a draw. Kasparov

wves the variation 30...fxe4 31 1Wc4 f5 31
Wc2 1Wc5+ 33 Wh 1 .i.xb2 34 ...xb2 1fxc4 3S
We5.
�------.,
Game S
Grischuk-Pavasovic
Istanbu/2003
-------•
1 e4 e5 2 .!Of3 .!Of& 3 .!Oxe5 d& 4 .!Of3
lf)xe4 5 d4 d5 6 ..id3 .!Oc& 7 0-0 i.e7 8
c4 l0b4 9 ..te2 dxc4
Immediately clarifying the siruation in the
c.:cntre to give White an IQP.
10 .bc4 0-0 1 1 l0c3
Also possible is the straightforward I I
li)cS. For example, 1 1 ...c6 12 �3 lbxc3 1 3
hKc3 �5 1 4 1i'b3 (14 1id3 .td6 1 S �b3
WVf6 16 ...g3 .i.�.-6 is also murky) 1 4......c7 I S
.i.xd5 cxd5 1 6 .i.£4 (Sax-Pavasovic, Baden
I 'JIJ9). Now 1 6...'iraS 1 7 llfe1 .1£6 leads to
unck-ar play.
1 1 . . ..!0xc3
Exchanging is narural but retreating is also
pc 1ssible:
a) 1 t ...lbf6 is rather passive. After t2 ltleS
r6 1 3 .i.gS lbfdS 1 4 .i.xe7 Wxe7 1 5 lle1
.iw 1 6 .i.xdS lbxd5 1 7 lbxdS cxdS 1 8 '1Vb3
White's superior minor piece is enough for
n n edge.
b) More active is 1 1 ...�6 12 .i.b3 �f6
and now:
bl) 1 3 h3 (simply preventing ... .1g4)
1 3. . ..i.f5 1 4 .tc3 l:le8 I S a3 lbd3 (1 5...�6!?
3 liJxe5: The Main L ine with 8 c4
is also fine: after 1 6 dS ltleS 17 �xeS �xeS
Black has strong, active piece play) 1 6 llb1
cS 17 dxcS ltle4 (the tricky 1 7...�xb2?! back­
ftres after 18 1lxb2 �xc3 1 9 cxd6 �xb2 20
.i.xf7+! <ifhf7 21 ...dS+ Wf8 22 ...xf5+ ...f6
23 1i'xh7 when White is clearly better) 1 8
�c2! �xb2 (after 1 8...lt)g3?! 1 9 fxg3 l:lxe3
Black appears to be very active but a few
accurate moves from White defuse the dan-
ger: 20 1i'd2 .i.d4 21 �xd4 1Wxd4 22 Wh2
�g6 23 ltldS! :Cs 24 �f4 1We3 25 •xe3
l:lxe3 26 Ld3 llxd3 27 llbe1 and White is
simply a pawn up) 1 9 .xd8 l:laxd8 20 llxb2
.i.xc3 21 l:lxb7 �xeS 22 .i.xcS .i.xc2 and
White still holds a slight advantage, Karpov­
Kasparov, World Championship (Game 41),
Moscow 1 984.
b2) Direct action with 1 3 ltle5 is also
possible: 1 3...�6 (not 13...lbf5? 1 4 lbxf7!
llxf7 1 5 .i.xf7+ Wxf7 16 �3+, a fairly
common tactical theme in this line) 1 4 .1£4
li:lf5 (Black should prefer 14...liJaS!? I S .i.e2
.i.f5 with unclear play) I S lbxc6 bxc6 16 dS
cS 17 lba4 .i.a6 18 lle1 (A.Sokolov­
Ag7.amov, Riga 198S). Now Black should try
l!L.c4 1 9 .i.c2 ltld6 20 lik5 .i.c8 21 .i.e5,
but White stiU has the initiative.
1 2 bxc3 li)d5
1 3 'ttc2
1 3 'lrb3 only wastes time. After l3. ..lbb6
1 4 .1d3 cS I S 'irc2 h6 Black has equalised
comfortably.
1 3 .....ta&
23
The Petroff Defence
It is best to develop this bishop immedi­
ately. Delaying with 1 3...c6 14 .td3 h6 1 5
l:b1 :C8 1 6 �5 would make i t difficult for
Black to complete his development.
1 4 �d3 h6 1 S Ab1 b6 1 6 �h7+ ¢>118 1 7
i.e4 1lc8
Black steps out of the pin rather than
blocking it with 17 ... c6. After 1 8 ltle5 Wd6
19 i..d2 lLcS 20 c4 ti)f6 21 .i.£5 White
keeps a nagging initiative.
1 8 c4 lnf6
Black could have exploited the c-fLic line­
up with 1 8...c6!?. After 19 :C1 .Li6 the posi­
tion is complex.
1 9 i.b7 Ab8 20 i.e& i.d6 21 lld1
Advancing with 21 cS does not achieve
any ck"U after 21 ....te7 22 :Ct bxcS 23
l:xb8 Wxb8 24 dxcS Wb4 25 i..e3.
21 .'W'c8 22 Ab3 .i.fS 23 'irb2 i.d7 24
•.
lOes i.xc6 2S lnxc& 1Wg41
This excellent move is forced. If 2S .. ..z:r.aS?!
then 26 .i.xh6! gxh6 27 dS opens the long
diagonal with decisive effect.
26 llf3
Playing to win material with 26 f3 leads
only to a messy position after 26 ...Wh5 27 g4
ti)xg4 28 ti)xbS ltxb8 29 cS i..xh2+ 30 'itg2
Wh4 31 fxg4 'Wxg4+ 32 'itxh2 Wxdl.
26 ..llbe8 27 lnxa7!?
..
This pawn may seem poisoned, but White
has seen deeply into the position. Also inter­
esting is 27 .i.xh6l? We4 28 �5 i..xeS 29
dxeS tl)g4 30 i..f4 with an unclear position.
24
27 ...1WhS 28 h3 'tWaS
Threatening mate on e I as well as the
knight, but there is a solution.
29 1Wd21 1Wxa7 30 llxf6
This sacrifice is enough for a draw.
30 ...gxf6 31 1Wxh6+ Wg8 32 1Wh4
'tWaS % -%
Black cannot escape the checks.
Game 9
Anand-I .Sokolov
Dortmund 1999
1 e4 eS 2 lnf3 li)f6 3 lDxeS d6 4 lDf3
lnxe4 S d4 dS 6 i.d3 lDc6 7 0-0 i.e7 8
c4 lnb4 9 cxdS lnxd3 10 1txd3 1txd5
This is much more reliable than 10...ti)f6
which gives White the useful option of 1 1
'ifb5+ and now:
a) 1 1 ... i..d7?l I 2 'ifb3 and Black cannot
regain the pawn. Timman-Kovacevic, Zagreb
1985 continued 12...0-0 1 3 tlX3 bS 1 4 i..g5
b4 (now Black's kingside is shattered but
even if 14..llb8 15 :lfet b4 1 6 i..x f6 i..x f6
1 7 ti)e4 i.e7 1 8 tiXS his position is very
unpleasant) 1 5 i..xf6 gxf6 16 ti)e4 �h8 1 7
l:fet a S 18 l:lact .i.bS?! (it wa s already bad
but now Black is lost) 1 9 ti)h4 l:lgR 20 d6�
i..xd6 21 .xf7 .tf4 22 l:lcS l:lg7 23 Wxf6
Wxm 24 ti)xf6 and Black could resign.
b) Black should play 1 t ...Wd7 but White
still has an edge after 12 'Wxd7+ .i.xd7 13
ti)c3 0-0-0 14 l:t:l :Ides 15 ltleS i.. b4 16

.td2llhf8 t 7 llac1 .
1 1 Jle1 .i.f5
12 g4
This prevents Black from blocking the e­
rile with a later ....i.e6, but the weakening of
White's kingside is obvious.
The solid option is 1 2 llk3 t.Oxc3 1 3
1fxc3 and now:
a) The natural 1 3 ... c6?1 allows a beautiful
tncric: 14 .i.h6! .i.e4 (the point is 14...gxh6 I S
:cs 1i'd7 1 6 llaet .i.e6 1 7 dSI cxdS 1 8 :Xe6
fxc6 19 11fxh8+ .i.fB 20 1i'f6 and White wins)
1 5 .i.xg7 llg8 16 llxc4 1i'xe4 17 lle1 .xe1+
I K t.Oxe l llxg7 1 9 dS and White is much
better.
b) After 1 3... .i.e6 Black will be able to cas­
ric without any great difficulty:
b 1) 14 lieS looks aggressive but Black has
1wo acceptable replies.
b 11) 1 4...1i'c4?? is an ugly blunder: lS
W'xc4 .i.xc4 16 b3 .i.c6 1 7 dS and White will
win a piece.
bl2) 14 .....c6 1 S ..e1!? 0-0-0 1 6 .i.gS
.txgS 1 7 t.OxgS llhe8 1 8 llcl ..d7 19 •as
'it>bK 20 t.Oxe6 llxc6 21 llxe6 fxe6 22 •cs g6
2:l g3 .d6 was equal in Milos-Zarnicki. Villa
Martelli 1998.
b 1 3) 1 4.....d7 1 5 .i.gS f6 1 6 lle2 fxgS 17
:acl 0-0 18 llxe6 .i.f6 is unclear.
b2) 14 ..xc7 (White wins a pawn, but
Ulnck will have strong compensation based
on his bishop pair and light square control)
1 4. .i.d6 1 S ..c3 (1 S .c2 leads to similar
. 20....i.h l l 21 �3 .xb3 22 I.Oxb3 .i.c6 23
3 lt:Jxe5: The Main L ine with 8 c4
play, for example t S...0-0 16 .i.d2 .i.f5 17
"eb3 .xb3 1 8 axb3 f6 1 9 .i.c3 �f7 20 lbct2
llfd8 21 t.Oe4 .i.f8 22 b4 b6 23 /.Og3 .i.d3
and Black has adequate compensation, Hiib­
ner-Smyslov, Vdden 1 983) 1 S.. O-O 1 6 .i.d2
.
llfc8 17 'iVe3 h6 te .i.c3 bS (it is not clear
that this advance achieves anything; the sim­
ple 18 ...lle8!? 19 ..d2 'l'bs would have given
Black reasonable compensation) 1 9 a3 aS 20
'ii'd2 �3 21 dS! ..xdS (there is nothing
better: 21 ....i.xdS? loses neatly to 22 .i.xg7!
�xg7 23 lbct4!, threatening both the queen
and t.OfS+) 22 .i.xaS ..xd2 23 .i.xd2 llc2 24
.i.c3 .i.cS 2S I.Od4 .i.xd4 26 .i.xd4, Rogt:rs­
Volkmann, Bled Olympiad 2002. White has a
clear extra pawn, but the presence of oppo­
site-coloured bishops gives Black drawing
chances.
1 2 .i.g6 1 3 lbc3 .!Llxc3 1 4 11fxc3 f&
. . .
Since ....i.e6 is no longer an option, Black
must fmd a different way to escape the pin
on the e-file, and here he prepares to unpin
with ...�fl. An interesting alternative is
14 ...Wd6!?, for example 1S .i.gS f6 1 6 .i.h4
c6 I 7 .i.g3 1i'd7 with an unclear position.
1 5 11fxc7
Taking the pawn docs not force an advan­
tage, but neither do the alternatives:
a) 1 S .i.f4 lld8 1 6 l:te3 �f7 17 ..xc7 lld7
18 ..c3 l:lhd8 19 a3 .ite4 20 I.Od2 (the solid
way out is 20 l:tdt .i.xf3 21 l:txf3 ..e4 22
l:te1 .xd4 with dead equality - Varavin)
25
The Petroff D e fen ce

:C3 h5 and Black has plenty of compensa­ 'ilig2 1i'c4+ with a perpetual check.
tion, Varnvin-Egin, Kaluga 1998.
b) 1 5 b3 �fl 1 6 1i'xc7 llhe8 17 "lfc4
..xc4 18 bxc4 b6! (fixing then c-pawn as a
target; Motylev points out that 18 ...:Cc8?! 1 9
c5 b6 20 .i.a3 bxc5 2 1 llact gives White a
dear advantage) 19 .i.b2 llac8 20 l:lacl .i.d6
21 l:lxe8 l:lxe8 22 c5 .i.f4 (Zhang Pengxiang­
Motylev, Shanghai 2001). Motylev now sug­
gested that White should play 23 l:let llxel+
24 tfue1 bxc5 25 dxc5 �e6 26 lDg2 .i.d2 27
lDe3 �d7, assessing the position as equal.
1 5 0-0 1 6 1lxe7
•••

'This leads to sharp play. Black has the

usual good compensation after 16 ..c3 .i.d6 23 eba+ .J:lf7 24 Ad1 'tig4+ 25 �1
1 7 �3 11'xb3 1 8 axb3 .i.fl 19 lle3 l:lfd8. Wg7 26 h3 •e4 27 1rd5 'tfc2 28 Ad2
1rc1 + 29 Wg2 lle7 30 1rd8 'tfc6+ 31
'tfd5 1rxd5 + % -%

Game 10
Nyysti-M.Rychagov
Gothenbutg 2003

1 e4 e5 2 li:lf3 li:lf6 3 lt:lxe5 d6 4 li:lf3

lt:lxe4 5 d4 d5 6 .i.d3 lLlc6 1 0-0 .i.e7 8
c4 lLlb4 9 cxd5 li:lxd3 1 0 1rxd3 'tfxd5 1 1
lle1 .i.f5 1 2 �5

1 6 •xf3 1 1 llxg7+ Wh8 18 .i.h6

•..

•xg4+
White's mating threat is demonstrated by
18....i.e4?? t9 l:lxh7+.
1 9 •g3 'tfxd4 20 .lbg6!
White very wisely heads for a draw. In­
stead 20 l:lxb7? allows Black a deadly attack
on the g-file with 20...l:lg8 21 l:le1 ..d5!
(2t....i.ffi! fails to make the most of Black's
chances: 22 llxfl l:lxg3+ 23 hxg3 llg8 is only
unclear) 22 llee7 11fdt+ 23 <i>g2 .i.e4+ 24
l:lxe4 l:lxg3+ 25 hxg3 11fd5 26 l:tbe7 f5 27 'This is cet"tainly a threatening-looking
l:t7e5 "lfxe4+ 28 llxe4 fxe4 and Black should move, but Black has convincing replies.
win this ending. 1 2 o-o-O
..•

20 hxg6 21 11h3 Wg8 22 .i.xf8 .lbf8

••• Alternatively:
Another drawing path is 22. 11Vxb2 23
.• a) Black must avoid the losing blunder
lld1 llxffi 24 Wh6 11Vc2 25 l:ld7 1fb 1+ 26 12....i.h4? 1 3 g3 lDxg3 (there is no hope in

26

retreating: 13 ....te7 14 Wf3 g6 1 5 g4! and
White wins a piece) 14 1Vf3! lbe4 15 lbc3
�xc3 16 bxc3 Wxf3 1 7 lbxf3+ i..e7 1 8 i..a3
Ac6 1 9 .i.xe7 Wxe7 20 d5 and White wins
cas�y. Zuidema-Barendregt, Amsterdam
1 966.
b) 12 ... f6?! is more complicated but still
bnd: 1 3 lbc3 (13 Wf3 is only enough for
Cl(uality after 1 3...g6 L4 lbc3 lbxc3 1 5 bxc3
1i'xf3 1 6 lbx£3 c;ild7) 1 3 ...lbxc3 1 4 1fxf5
�b5 1 5 ..g4 l£lxd4 (of course Black would
like to take the piece with 1 5... fxe5 but after
1 6 llxe5 Wxd4 1 7 llxc7+ Wxe7 1 8 .i.g5+
�d6 1 9 lld1 White is dearly better) 1 6 l£ld3
and the pin causes Black immense difficulty:
bl) 16 ...Wf7 1 7 l£lf41 1ff5 18 llxe7+!
�xe7 1 9 1fxg7+ �d6 20 .i.e3 llhg8 21 1ff7
cS Qf 21 ... b6 then 22 Ld4 1Vxf4 23 .i.xb6!
clears the d-file and continues the attack) 22
1i'xb7 llgb8 23 Wa6+ llb6 24 Wa3 and
White's attack was obviously progressing weU
In Jonsson-Schandorff, Panormo 2001 .
b2) If 1 6...lbc2 White must find 1 7 lbb4!
�xb4 1 8 .xb4 c5 1 9 Wg4 �f7 20 i..h6!
gxh6 2 1 llad1 with a massive attack, De Fir­
mian-Plaskett, Copenhagen 1 985. 17 llxe 7+
leads to a draw after 1 7...Wxe7 1 8 1Ve2+ Wf7
19 'irxc2 llad8 20 lbe 1 1i'd1 21 1Vxc7+ Wg6
22 1Vg3+, as the pins force White to setrle for
n perpetual.
c) A reasonable alternative is 1 2...g6 1 3
1'13 (not 1 3 g4?1 lbxt21 1 4 W n 'eht+ 1 5
�xf2 i..h4+ when White is i n trouble)
1 l .. Wxd4 1 4 lbc3 1i'xe5 1 5 .i.f4 (or 1 5
lDxe4 0-0-0 1 6 .i.f4 1i'a5 1 7 llact lld7, when
the game is unclear) 1 5 •a5 1 6 b4!? 1i'a3!
...
(lhis is the best way to contain White's initia­
tive; 1 6 ...1Wxb4?1 is obviously bad after 1 7
ltld5, while 1 6. .ixb4 is not a clear mistake
hu1 after 1 7 llxe4+ .i.e6 1 8 .te5 White's
initiative is very dangerous) 17 lDd5 .x£3 1 8
!(X f3 i..dS 1 9 fxe4 .te6 20 lbxc7+ i..xc7 21
j_xc7 (the game has fizzled out to equality)
2 l ...f6 22 a3 llc8 2l llact <t>f7 1/z- 1/z Be­
liavsky-Smyslov, Re�o Emilia 1 986.
1 3 .,3 g6 1 4 g4
3 l0xe5: The Main L ine with 8 c4
1 4 lbc3 is a solid option. However, after
14...lbxc3 15 bxc3 Wx£3 16 gxf3 Black has
no problemS. For example, 1 6 ...llde8!? 1 7
lbxf7 llhfB 1 8 l£le5 .i.h3 with compensation.
1 4...-*.h4
1 5 �3
After this White seems to be in trouble.
FoUowing 15 lbc3 there arc many paths bu t
White cannot force more than equality:
1 5...lbxc3 1 6 bxc3 .i.e6 17 g5 (or 17 .xd5
.i.xcl5 1 8 g5 f6 1 9 gxf6 .i.xf6 with a balanced
game) 1 7.. .llhe8 1 8 c4 .xd4 1 9 llb1 c6 20
.i.f4 (if 20 lbxc6 Black bails out to a draw
with 20 ...Wxt2+1 21 1Vxf2 .i.xf2+ 22 Wxf2
bxc6) 20...1Wc5 21 lbxc6 lld31 22 llxe6 (after
22 Wxd3 Black can force a draw with
22..Wxf2+ 23 Wh t .i.dS+ 24 Wxd5 l'lxet+
25 llxe1 .xet+ 26 Wg2 '1Ve2+ 27 Wh3 .ft+
28 Wxh4 1i'xf4+; instead 22 lle3 continues
the fight: 22...llxc3 23 lbxa7+ Wxa7 24 .i.xe3
Wb8 25 We4 .i.xgS 26 i..xgS and with oppo­
site bishops and both kings vulnerable, the
position is totally unclear) 22... fxe6 23 ..xd3
Wxt2+ 24 Wht .xf4 25 lbxa7+ �b8 26
lbc6+ Wc7 27 lba5 lidS 28 1Ve2 lld2 29
llxb7+ Wd8 30 llb8+ Wc7 lf2-1/z Wahls­
Yusupov, Germany 1 992.
1 5 10xf2
.•.
This neat trick turns the game in Black's
favour. Another attractive choice is
1 5...llhg8!? 1 6 .ic3 hS 17 h3 hxg4 18 hxg4
.i.e6 when Black has a dangerous initiative.
1 6 'lfxd5 l0h3+ 1 7 �2 llxd5 1 8 gxf5
27
The Petroff Defence
.i.xe1 1 9 �xe1 lle8 20 �f3 llxf5 21
Wg3?1
White should prefer 21 tik3, but after
21 ...�f4+ 22 .i.xf4 lLf4 Black is still obvi­
ously on top.
21 �1 1 22 lbbd2
•..
The point is that if 22 �xgl then 22. ltft
.
23 l!Xi2 lLgl+ 24 Wf3 Aeel 25 tDb3 f5
leaves White hopelessly tied up.
22 �2+
•.•
Anothc:r way to win is 22...:C3!? 23 'iPf2
.l.ex£3+ 24 Wxgt l:tg5+ 2s Wht Af2.
23 Wg2 g5 24 �b3 g4 25 �1
There is also little hope aftc:r 2S �eS
�xd4 26 �xd4 AfxeS 27 �g3 £5.
25 .. .c!wcc1 26 llxc1 lle2+ 27 Wg3 llf1
Perhaps evc:n simpler was 27 ...hS 28 tileS
Aft 29 lilcd3 £5.
28 c!Lig2 llf3+ 29 Wxg4 llxb3 30 c!Lif4
h5+1 0-1
28
11tis little trick enables Black to save both
rooks, so White resigned.
Game 1 1
Shirov-Gelfand
Astana 2001
1 e4 e5 2 c!Lif3 c!Lif6 3 c!Lixe5 d6 4 c!Lif3
c!Lixe4 6 d4 d5 6 .i.d3 .i.e7 7 0-0 c!Lic& 8
c4 c!Lif6
8....tc6 possibly commits the bishop too
early:
a) 9 cxdS is interesting: 9 ....i.xdS 1 0 W
lilxc3 11 bxc3 0-0 1 2 i.f4 .td6 13 .txd6
Wxd6 14 lL\g5 h6 (14. . £5 1 S c4 .i.xg2! 1 6
.
Wxg2 1Wg6 1 7 f4 h6 1 8 dS llkl8 1 9 h4 hxgS
20 hxgS was a bit better for White in l..owen­
thai-Morphy, London 1 8S8) 1 S 1Whs .i.xg2
16 .i.h7+ �h8 17 �xg2 1Wd5+ 1 8 .te4
1Wxg5+ 1 9 1WxgS hxgS 20 llab 1 llab8 21 l:l.bS
and White has good compensation.
b) 9 Act is natural and probably best
9...lilf6 to cS!? 0-0 I 1 lilc3 (preventing
....i.g4 with 1 1 h3!? may be best: 1 t ...tDb4 1 2
lilc 3 lilid3 1 3 1Wxd3 h 6 1 4 tiles i s good
enough for an edge) 1 1 ....i.g4 1 2 .te3
.i.xcSI? (now Black has counterplay) 1 3 dxc S
d4 14 .txd4 (1 4 h3 i..x f3 1S 1Wxf3 dxc3 1 6
lladl tDcS 1 7 'Wf5 ll:lxd3 1 8 Axd3 'fle7 1 9
llxc3 llad8 is equal) 1 4...tilxd4 1 S Lh7+
lilih7 1 6 .xd4 ..Lf3 17 1Wxd8 llfxd8 18
gxf3 ll:lg5 with a level position, Short
Smyslov, Hastings 1988/89.

Also possible is 8 ..ig4 9 ltic3 (9 cxd5 is
loo simple an approach: 9...Wxd5 10 ltic3
lt:\xc3 1 1 bxc3 0-0 12 IZ.c l ..i.xf3 13 Wxf3
Wxf3 14 gxf3 ..i.d6 1 5 :b1 lLb8 1 6 ..i.e3 b6
17 f4 t'iJe7 1 8 c4 c6 1 9 Wg2 :rds and Black
had equalised in Kotronias-Langrock, Ham­
burg 2001) 9...t'fJxc3 (9...t'iJf6 transposes to
H ...t'fJf6 9 t'fJc3 ..i.g4) 1 0 bxc3 0-0 1 1 IZ.c1
tlxc4 1 2 ..i.xc4 Wd6 13 :b1 ltab8 14 :b5
nnd White has hi� usuaJ edge.
9 .!003
Black is okay in the variations after 9 cxd5
i;)xd5. For example, 10 t'iJc3 0-0 11 :e1 t'iJf6
1 2 a3 ..i.g4 1 3 d5 t'iJd4 14 ..i.xh7+ t'iJxh7 1 5
Wxd4 R.xf3 1 6 gxf3 R.f6 1 7 1i'd3 Wd7 18
.if4 t'iJg5 �t is important to improve the
badly placed knight immediately; instead in
Svidler-Ovetchkin, Tomsk 2001 Black played
IH..1lfe8 when 19 t'iJc4 stranded the knight
on h7 - White was a touch better after
19...b6 2.0 :e2 c5 21 llact R.d4 22 d6) 19
.lxg5 ..i.xg5 20 IZ.c4 WfS and White's weak­
ened structure means that Black has no prob­
lems.
9 .i.e6
..•
9...R.g4 exerts less influence on d5, so it's
luwcaJ for White to continue 10 cxd5. For
example, 1 0...t'iJxd5 1 I ..i.e4 R.e6 1 2 Wd3
�cb4 1 3 Will c6 14 a3 t'fJa6 1 5 t'fJxdS ..i.xdS
J (, l'.lc1 t'iJc7 17 R.gS R.e6 1 8 ..i.xc7 1r'xe7 1 9
li]c5 0-0-0 20 Wc2 l:hd4?1 (20. . 1r'd6 21 .
Jlnd1 is better, after which White has only a
Hmall advantage) 21 .i..xc6! 'iVd6 22 ..i.d7+!
3 lDxe5: Th e Main Line with 8 rc 4
R.xd7 23 t'fJxf7 'ttfB 24 t'iJxh8 Wxh8
(Sutovsky-Huzman, Tel Aviv 2000) . Now
White should have continued 25 WcS :<12
26 'ttc3 :d5 27 1Z.c7 with a dear advantage.
1 0 c5
Ensuring a space advantage at the expense
of taking the pressure off Black's centre.
With the bishop already on e6, 10 cxdS
seems less logical but it is b)• no means bad.
For example, lO.. .t'fJxdS 1 1 IZ.c1 0-0 12 .i.e4
.i.f6 1 3 a3 h6 14 1r'd3 t'fJce7 1 5 .i.d2 c6 1 6
ltad1 llc8 17 h3 t'iJxc3 1 8 bxc3 b S 1 9 .i.f4
..i.dS 20 t'iJd21 llc8 21 1i'g3 Wh8 22 R.c21?
(dearing the way to centraJise the knight;
instead 22 Wg4 R.c6 23 'i'hs t'iJg81 24 .i..b I
rfJ 25 'irf3 .idS 26 1r'g3 .ig7 was equaJ in
Movscsian-Yusupov, Batumi 1999) 22....d7
23 t'iJe4 and White's position is mildly prefer­
able.
1 0.....tg4 1 1 ..tb5 0-0 1 2 ..txc& bxc& 1 3
h3 ..txf3 1 4 11'xf3 Ae8 1 5 ..te3 1ib8
Black can provoke many exchanges with
1S ...t'iJe4 but after 16 t'fJxe4 dxc4 17 1r'xe4
..i.xcS 1 8 •xc6 .ixd4 19 ltad1 1tf6 20 'irxf6
.ixf6 21 b3 White's better structure is a sig­
nificant factor.
1 6 .:ab1 .i.f8 1 7 l:lfe1 eb4
1 7 ...t'iJe4 still does not help. Gelfand
anaJyses 18 .if4 t'iJxc3 19 bxc3 llxct+ 20
llxe 1 Wb2 21 lle2 as dearly berter for White.
18 a3
1 8 Wf4!? l:tac8 1 9 lle2 also keeps an edge.
18 11'c4 1 9 l:lbc1
..•
29
The Petroff Defence
Again there was a good alternative: 1 9
1i'g3 lie? 20 lled1 g6 21 .i.gS .i.g7 22 1i'f4
lObS 23 '1t'd2 and White has a pull.
1 9 .%lab8
••
If 19...lle6 then 20 ..g3 reminds Black of
the weakness of c7.
20 11ff5
Now Black is able to fix b2 as a weakness.
20 b4 was required to keep a small advan­
tage: 20...a5 21 bxa5 :b3 22 l0b11; or
20...1i'b3 21 -*.g5 'lt'xa3 22 .txf6 gxf6 23 b5
cxb5 24 1Wg4+ ..thB 25 l0xd5 and White has
a strong attack.
2o...a3 21 .t.us 10&4
The bold 21...11'xb2 is also possible. 22
.i.xf6 gxf6 (but not 22. ltxe1+ 23 l:lxe1
.
11'xc3 24 :e3 'lfcl+ 25 Wh2 'lt'bt 26 1i'd71
gxf6 27 flg3+ .i.g7 28 'lt'g4 'lt'g6 29 'lt'f4
when Black has to surrender his queen) 23
1Wd3 .i.h6 24 llxe8+ :xeS 25 :c2 1i'xa3.
Notkin analysed this variation but under­
standably stopped now with the assessment
of 'unclear'.
22 ll)xe4 dxe4 23 .llxe4 11fd3 24 Ace1
.llxb2
It's clear that White has lost control of the
position.
25 11ff4
Or 25 11'f3 :xe4 26 1t'xd3 rlxel+ 'Z7 Wh2
h5 wilh a murky, unbalanced position.
25 .1lxe4 26 11fxe4 11fxe4 27 Axe4 f6 28
•..
.i.f4 Wf7 29 .i.xc7 Aa2
29..ltb3 was another way to draw after 30
30
.i.d6 Ld6 31 cxtl6 :b?.
30 .i.d6 .:Xa3 31 .i.xf8 ¢»xf8 32 Ae6 Ad3
33 .llxc6 Ad1 + 34 �2 .llxd4 35 .:CB+
�e7 36 c6 Ac4 37 Ac7+ Wf8 %-%
Game 12
Topalov-Anand
Wijk aan Zee 2003
1 e4 e5 2 lL!f3 lL!f6 3 ll\xe5 d6 4 ll'lf3
lL!xe4 5 d4 d5 6 .i.d3 lLlc6 7 0-0 .i.e7 8
c4 ll'lf6 9 h3 lLlb4
9._0-0 is also reasonable: 10 l0c3 tlxc4
(10...ltlb4 t t .i.e2 dxc4 1 2 .i.xc4 transposes
to the main game) 1 1 -*.xc4 l0a5 1 2 .i.d3
.i.e6 13 llet l0c6 (13...c5 liquidates the cen­
tre but following 14 dxc5 i.xc5 1 5 i.g5
White's superior activity is still annoying) 14
a3 a6 1 5 .i.f4 rle8 (t5...1i'd7?! exposes the
queen to a series of tactics: 16 lOeS l0xe5 t 7
dxeS ltlds 1 8 l0xd5 .i.xd5 1 9 •c2 g6 20
:Sdt c6 21 .i.h6 :rds 22 e61 fxe6 23 .i.xg6!
and White is much better, Kasparov-Karpov
World Championship [Game 481, Moscow
1 985) 1 6 :C1 lOds 17 l0xd5 .i.xd5 and
Black has comfortable equality.
1 0 .i.e2 dxc4 1 1 .bc4 0-0 1 2 lL!c3 c6
1 3 Ae1
1 3 a3 is a logical alternative, for example
t 3...ltlbd5 14 llet .i.e6 1 5 .td3 h6 1 6 .i.d2
rle8 Qt is best to avoid all the :Xe6 ideas
with 1 6 ...1Wc81; after 1 7 .c2 :cs 18 �e2
.i.d6 19 l0e5 a complicated struggle lies

11hcad) 17 llxe6!? (White sacrifices the ex­
change to weaken the light squares around
Hinck's king) 1 7 ... fxe6 1 8 �g6 l:lffi 19 1i'e2
�7 20 .Jlct .td6 21 .tbt .Jle8 (In Bologan­
llo�entalis, Belfort 1 998 Black tried
2l.. .•e7?1 but after 22 ltleS .i.xeS 23 dxeS
Q')f<.I5 24 ltlxd5 ltlxd5 25 1i'e4 g5 26 1i'g6+
11rg7 27 1i'xe6+ *h8 28 h4 White had a dan­
Aerous attack) 22 ltle5 c5 23 llle4 cxd4 24
ltlxf(r+- ..xf6 25 ltlg4 and here Bologan
1tnres that White has good compensation for
the material deficit.
13 ltJbd5 14 Wb3
.•.
'Inc other try is 14 �g5 .i.e6 15 WbJ and
nnw:
a) 1 5...1i'b6 is far too obliging. White eas­
Ily achieves a preferable structure after 1 6
�xd5 cxd5 1 7 'irxb6 axb6. Following 1 8
.ib.l h6 19 �f4 llfc8 20 ltles g5 2 1 .i.h2
J..b4 22 .Jle2 ltlc..-4 2.l .tZkl3 .*.fB 24 J:l.ae1 .i.g7
(nnt 24...lbd6?1 25 .txd6 .i.xd6 26 .llxe6!
fxc6 27 .Jlxe6 when White will win far too
many pawns) 25 .*.e5 .*.xeS 26 ltlxc5 �g7
27 lbd3 White had an enduring Cllgc in
I .cko-Adams, Dortmund 2002.
b) 1 s....JlbS 16 J:.ad 1 .Jles 11 .th4 h6 t R
i.g3 .*.<.16 1 9 .tcS lllxc3 20 bxc3 .*.xc4 (an­
olhcr convincing eyualiscr is 20...lbd5!? 21
'i'c2 b5 22 .*.d3 a6) 21 1i'xc4 llldS 22 'irbJ
hS 23 c4 bxc4 24 1i'xc4 .Jle6 25 .Jlct .llb6 26
i.xd6 .llxd6 with a balanced position, Lutz­
Dautov, Bundesliga 2004.
14 l0b6 1 5 ..id3 ..ie6 1 6 •c2 h6
..•
3 l0xe5: The Main Line with 8 c4
1 7 1lxe6!?
Boldly sacrificing the exchange is the the­
matic continuation. The safe liner is 1 7 a3
ltlbdS t 8 ltla4 (18 .i.d2 led nowhere in Gcl­
fanc.I-Yusupov, Istanbul 2000: after 18....Jlc8
19 ltla4 .Jlc7 20 .Jle2 .i.d6 2 1 l:.ac t ltlhS 22
.i.h7+ �h8 23 ..i.. £5 lllh f4 24 .i.xf4 ltlxf4 25
.Jle3 tlld S 26 .Jl3e2 ltlf4 27 .Jle3 a draw was
agreed) t 8...ll:ld7 19 .i.d2 .Jle8 (Bologan­
Kasimdzhanov, Pamplona 2002). Here Bolo­
gao suggests 20 llc21? '1Vc7 21 .Jlae1 ltlf4 22
.i.xf4 'irxf4 23 ltlc3 'ire? 24 dS cxdS 25 ltld4
with an edge for White. Instead, 23...�f8
aUows 24 llxL-61 fxc6 25 ..tg6 .Jled8 26 .Jlxe6
with a big advantage.
1 7 . . .fxe6 1 8 ••2 1lrd7 1 9 ..i.d2 ..i.d8
If 1 9....Jlad8 20 .lle 1 .J.d6 21 lllc4 ltllx15
White can claim an edge.
20 lOe4 ltJbd5 21 l0xd8 •xd6 22 Jle1
.llad8 23 a3 JZ.fe8 24 ..i.b 1 c5 25 lfle5
cxd4 26 1i'd3 ltJe3
After 26...11'b6 Anand suSbrests 27 b4 li:k7
28 .J.a2 to keep up the pressure.
27 ltJf3 lOadS
27...ltlf51? is a solid option, but after 28
.i.a2 ltlds 29 llc4 Topalov still claims com­
pensation.
28 ..i.xh6! ltJf4
The bishop cannot be taken: 28...gxh6??
29 9g6+ �£8 30 •xh6+ �g8 31 .Jles and
White wins.
29 ..txf4 1i'xf4 30 ..i.a2 JZ.d6 31 h4
A good alternative is 31 'it'c4!? ltlds 32
31
The Petroff Defence
/t)xd4 lZ.fB 33 l£lf3 with perfectly adequate
compensation.
31 �6 32 .*.c4 et7 33 ..,3 d3 34
•.•
Ad1
Topalov starts to lose his way - it was
probably time for 34 lZ.xe6. Following
34...:dxe6 35 .t.xe6+ �h8 36 l£Je5 'l'hs the
position is unclear, one possible continuation
being 37 l£lf7+ ..xf7 38 .ixf7 lZ.et+ 39 �h2
d2 40 ..xb7 dt .. 41 ..c8+ �h7 42 ..£5+
with a draw by perpetual check.
34 d2
.••
35 11'b4?1
The d-pawn had to be taken, even though
Black is a bit better afu:r 35 ltxd2 llxd2 36
/t)xd2 ..xh4 37 Le6+ �h8.
35 .Aed8 36 l0g5
••
Now it is too late for 36 llxd2?!. After
36 ...�1+ 37 Wh2 ltxd2 38 /t)xd2 Wet the
pin is deadly.
36 •c2 37 '*.b3
•••
32
Or 37 �3 11rxb3 38 .ixb3 lZ.b6 39
.ixe6+ llxe6 40 /t)xe6 lZ.eS 41 lZ.xd2 ltxe6
and Black should win.
37 •c6 38 g3 1tb& 39 .!Oxe& •xb4 40
•••
axb4 .llc8
Since the discovered check is harmless, it
is clear that Black is winning.
41 �1
41 lDcS+ docs not help: 41 ...Wh8 42
lDxb7 lZ.b6.
41 ...Wtla 42 lOQ5 Ad4 43 lLln+ �h7 44
lOgS+ Wtl6 45 lLln+ wg6 46 li)e5+ �h7
47 l0f3 .Jlxb4 48 .!Oxd2 AdS 49 .*.c2 +
�hB 50 b3 lZbd4 51 �e1 b5
52 lDb1
The queenside pawns are also decisive af­
ter 52 /t)f3 lZ.e8+ 53 Wft ltxdl+ 54 .i.xdl aS.
52. . .Ae8+ 53 �1 llxd1 + 54 .*.xd1
lDe4 55 .*.c2 a5 56 lDB3 lDd2+ 57 �2
llb8 58 g4 b4 59 �b1 l0xb1 60 .i.xb1
AcB 81 .*.f5 llc3 0-1
 3 /i)xe5: The MBin Line with 8 c4

Summary
The line 6... .i.e7 7 0-0 liX6 8 c4 has been enduringly sruuied, but it is still very popular. Strate­
lo{ic understanding of the variation is the key to succL"SS - there is no real need to memorist: the
det:uls.
Although 8...lbf6 looks quite reasonable, at the moment aD the attention L� on 8 ...lbb4.
White should really retreat his bishop to e2 - Games 9-1 0 show that 9 cxdS does not lead to
�uccess against accurate black defence. After 9 .i.e2 the bishop fianchetto to b7 (Game 7) may
well not equalise (see the note to White's 1 8th move). In the case of 10 ... .i.e6 t i lDeS neither a
counterattack in the centre with 1 1 ...c5, nor ousting the centralised knight with 1 1 ...f6 shakes
off aU Black's operung problems. Mobilisation on the qut:enside with 1 1 .i.t:J seems to give
lc:ss chances for an advantage after both t t ...fS and 1 t....i.f5.
With the knight on b4 it is logical to develop the bishop to f5 as soon as possible. After 9
.le2 0-0 10 lbc3 .i.f5 1 1 lbeS, t l...cS is quite effective, while in the event of 1 1 a3 lbxc3 1 2
bxc3 llX6 the complicate<.!, strategic play has not revealed any appreciable advantage for White
�o far.

1 e4 e5 2 /i)f3 /i)f6 3 /i)xe5 d& 4 /i)f3 /i)xe4 5 d4 d5 6 ..td3 ..te7 7 0-0 Nc& 8 c4 /i)b4
8...lbf6
9 llX3 - Ga111e 1 1; 9 h3 - Gtllllt: 12
8 ..te2
9 cxdS lbxU3 10 1Wxd3 1i'xd5 t l llet .i.fS (D)
1 2 g4 Ga111e 9; 1 2 lDc5 - Ga111t 10
-

8...0-0
9...dxc4 - Ga111e 8
10 lllc3 ..tf5
10.. b6 G1111�e 7
. -

I O.. ..i.c6 (D)

t t lbes
1 1 ...cS - Ga111e 3; 1 1 ... f6 - GtJJ1Je 4
1 1 .i.e3 .i. f5
1 2 1Wh3 - Ga1ne 5; 12 llct - Gt�nJe 6
11 a3 �xc3 1 2 bxc3 t;)c& 1 3 lle1 lieS (D) 14 cxd5 - Ga111t I
14 .i.f4 Gtlll�t 2
-

1 1 . ..tf5
. . 10 ..te6
. . . t3. . . :es

33
 CHAPTER TWO
3ltJxe5: The Main Line
with 8 lle1
1 e4 e5 2 l0f3 lOt& 3 l0xe5 d& 4 .!Df3
l0xe4 5 d4 d5 6 .i.d3 .i.e7 7 0-0 .!De& 8
l:le1
After 1 e4 eS 2 li)f3 li)f6 3 lilieS d6 4 li)f3
li)xe4 5 d4 dS 6 .i.c.J3 il..e7 7 0-0 liJc6, it's true
that the immediate attack on the dS-pawn with
8 c4, as covered in the previous chapter, is in
the foreground of modem practice. Howt.-ver,
8 :Ct, developing the rook and attacking the
knight on e4, is a serious alternative. Even
Bilguer considered the 'pawn sacrifice' 8 .il_g4
..
to be the best reply, and he indicated that it
was unfavourable for White to grab and hold
on to the pawn: 9 .i.xe4 dxe4 1 0 lLe4 .i.xf3
1 1 gxO?! (1 1 •xn li)xd4 12 1i'd3 maintains
equality) 1 1 ...fS 1 2 llf4 0-0. Instead White
should advance his c-pawn with the aggressive
9 c4 (Games 1 3- 1 5) or the modem 9 c3
(Games 1 6-18).
The intensive attack on the centre with 9
c4 virtually forces Black to defend the dS­
pawn with 9 .li)f6. If White retains pressure
..
in the centre with 10 li)c3 (Game 1 3), the d4-
pawn fall-; and Black equalises with either
1 0...li)xd4 or 10 ... .i.xf3 11 .xf3 li)xd4. This
is the reason White often releases the tension
in the centre with 1 0 cxdS (Games 14-1 5).
Pillsbury introduced I O...•xdS, but after 1 1
llJc3 neither the retreat 1 t...1i'h5 nor the
queen swap I t....i.xO 1 2 li)xd5 il..xdt leads
34
I
to equality (the latter in view of 1 3 li)xc7
�d7 1 4 .i.t4Q. Black can take on dS with his
queen without harm if he precedes this with
l O... il..xO 1 1 .xf3 (Game 1 4), while in
Game 1 5 Black takes the strategic decision to
occupy the square in front of the isolated
pawn with 1 0. .li)xd5.
.
While 9 c4 looks aggressive, 9 c3 appears
distinctly restrained. However, comparing the
consequences of the lines leads to unex­
pcctt..od results. 9 c4 leads to the quick reduc­
tion of tenc;ion in the centre with approxi­
mately equal chances, while the conse­
quences after 9 c3 arc unproven. White
strengthens his d4-pawn and resumes his
threat to capture on e4. Moreover, after c2
c3 White's queen has a clear way to b3 and
after 9... f5 it is Jjkcly to take this road. 1 0 h3
has not become popular, although it looks
logical to force the bishop on g4 to make a
decision before the queen leaves for the
flank. 1 0 c4?!, trying to take advantage of the
weakening of the a2-g8 diagonal at the cost
of a tempo, is unjustified: 1 o ..dxc4l 1 1 .i.xc4
.
•d6! and .0-0-0 (Mikhalchishin). to li)bd2
..
.d6!? 1 1 c4!? (Game 1 6) makes some sense
because ... d5xc4 does not work, while the
manoeuvre li)d2-ft deserves consideration
after 1 0 li)lxl2 0-0. However, this plan is less
appeaJjng after 1 0 li)lxl2 •d6 in view of

Ul11ck's possibility of castling long.
An early 'iVb3, attacking b? and d5, is cov-
r!rcd in Games 1 7 1 8. A comparatively rare
· Now this capture on f3 is safer. lO...dxc4
J>lnn after tO 'Wb3 is 10 ...1i'd6, preparing
lJUccnsicle castling. Ths gives White both the
lypical possibility of 1 1 �bd2, 1 2 1i'c2 fol-
lowed by the lunge with the b-pawn, and a
rnradoxical idea of 1 1 �fd2!? followed by f2-
1'3. In the main line with tO 1i'b3 0-0 1 1
t1�bd2 Black has two main options: the king's
removal from the dangerous a2-g8 diagonal
n2-g8 with 1 1 ...';;,h8 (Game 1 7) offers a sacri-
nee of the b-pawn, while Botvinnik proposed
prot1.:cting b7 with 1 1 ...�5 (Game 1 8).
Game 13
Shirov-Vusupov
European Team Ch., Batumi 1999
-------•
1 e4 e5 2 li'Jf3 li'Jf6 3 li'Jxe5 d& 4 li'Jf3
�xe4 5 d4 d5 6 .i.d3 liJc& 7 0-0 .i.e7 8
.l:le1 ..llg4 9 c4 li'Jf6
Trying to win a pawn immediately is a
mistake: 9...i.xf3?1 1 0 'if'xO �xd4 11 ..d
li:\15 12 1ih3! �fd6 1 3 cxcl5 �f6 1 4 i.g5
�.:'lxd5 1 5 �c3 �fB 1 6 llxe71 �xe7 1 7 �5
�)xdS (this gives up material and admits that
It has all gone wrong; if Black tries 1 7 ... f6
rhcn 1 8 �xf6! gxf6 1 9 i.xf6 W f7 20 •o is
fnr too strong) 1 8 i.xd8 llxd8 1 9 �4 �f6
20 1t'd4 and White was much bcttet· in Lo­
hron-Handoko,Jakarta 1 983.
1 0 l0c3
3 liJxe5: The Main L ine with 8 lie 1
1 0 cxd5 is the subject of Game 1 4.
1 0 .i.xf3
..•
is less accurate after 1 1 i.xc4 0-0 12 clS �aS
1 3 i.cl3 c6 1 4 h3 i.hS 1 5 i.gSI? cxclS 16 :es
i.xf3 (not 16 ...d4? 17 i.xf6 i.xf3 1 8 'ii'x f3
i.xf6 19 lidS clxc3 20 :Xd8 llfxd8 21 bxc3
with a big advantage for White, J.Polgar-
Kamsky, Las Palma.-; 1 994) 1 7 'if'xO �c6 1 8
i.xf6 i.xf6 19 llxdS, when White's pieces
are sufficiently active to force an advantage
even in this symmetrical position.
However, Black should seriously consider
10 ...�xd4 here: 1 1 cxd5 i.xf3 (1 1 ...c5?! is
inaccurate: 1 2 i.bS+ WfB 1 3 .ie2 and White
has a useful k·ad in development) 12 gxO
c5!? 13 dxc6 �xc6 14 i.bS 0-0 15 '5'xcl8
i.xd8 16 .i.xc6 bxc6 17 i.c3 i.b6 1 8 .J:r.ad l
llfd8 and the game was completely level in
J.Polgar-Karpov, 1-loogt:Veen 1999.
1 1 1fxf3 lt'lxd4 1 2 'W'd1
White can also try 1 2 'iVg3, when 1 2...�e6
is a little better for White after 1 3 .i.f5 0-0 14
cxd5 �xd5 1 5 �3 �f6 1 6 i.e3. However,
Black is okay after 1 2...dxc4 13 i.xc4 (Kar-
pov pointed out the losing blunder 13 'iVxg7?
�0+1 14 �hl llg8 1 5 'ifxf6 �xc1) 1 3...0-0
1 4 .tg5 i.d6 1 5 �4 h6 1 6 i.xf6 1i'xf6 1 7
1i'xf6 gxf6, Lobron-Karpov, Hanover 1 983.
1 2 li'Je& 1 3 .i.f5
.••
After 1 3 cxdS Black defends successfuUy:
1 3...�d5 1 4 i.b5+ c6 1 5 lllxd5 cxb5 1 6
.i.f4 �xf4 1 7 llxe7+ (not 1 7 �xe7?! �6 1 H
35
The Petroff Defence
tllf5 0-0 and Black keeps a useful extra
pawn) 17...¢'£8 18 lies •d6 19 %15 lids 20
lOe3 ..xdl+ 21 %lxd1 :lxd l+ 22 c!tlxd1 etle6
23 %lxb5 b6 24 ttlc3 q;e7 and unswprisingly
a draw was agreed in Kasimdzhanov­
Yusupov, Essen 2001 .
1 3 dxc4
•.•
Much riskier is t3...d4 14 .ixe6!? (this is
more dangerous than 1 4 ttle2 d3 1 5 tC!f4
ttld4 16 .ixd3 0-0 1 7 .ie3 .ib4 1 8 ltft c5
19 tlle2 ttlg4, which was unclear in J.Polgar­
Shirov, Prague 1999) 1 4... dxc3 (or 14... fxe6
15 tC!bS e5 16 %lxc5 a6 1 7 etlxd4 c5 1 8 .ih6!
and White has a vicious attack) 15 ..xdS+
:lxdS 1 6 .i.f5 cxb2 17 .ixb2. White has
good compensation: the bishop pair, a lead in
development, and the annoying pin on the e­
fale.
1 4 ...84+
14 c6
••.
14.....d7?1 sees the queen stepping into
ttouble: 1 5 ..xc4 lids 1 6 .i.g5 and White has
n dangerous initintive. For example, 16 ... h6
17 .i.xe6 fxc6 lS .ixf6 .i.xf6 1 9 %lad1 ..cS
20 %lxe6+ 'i\lf8 2 1 ltxd8+ 11'xdS 22 ttlds and
White's attack crashes through.
1 5 -*.xe6 fxe6 1 6 •xc4 ll:ld5
Black's simplest plan is 1 6 ...0-0 17 ..xc6+
:n 18 .i.g5 1l'd6 1 9 1Wxd6 .i.xd6 with (..'llSY
equality.
1 7 llxe6 0-0
There is no time for 17...tC!xc3? because
of 1 S .i.g5 etld5 19 .i.xe7 c!tlxe7 20 :lae 1 %1£8
36
21 llxc7+ 11'xc7 22 .rlxe7+ Wxe7 23 �+
�f6 24 11'xb7 %lfc8 and White has good
winning chances.
1 8 -*.e3
1 8...-*.f6
Another option is 18 ...Wh8, for example
19 etlxd5 cxd5 20 ..d3 .i.f6 21 :ldt 1Wa5 (or
21 ....i.xb2 22 ..xd5 ..xd5 23 llxd5 b6 24
llc7 �g8 25 g3 and White's activity gives
him a smaU endgame advantage, Magem­
Lal.ic, Seville 2000) 22 �3 :ladS 23 .id2
..c7 24 .i.b4 llfe8 and White again has a tiny
edge, this time due to his better structure.
1 9 -*.c5 J:lf7 20 J:lee1
This is more conttoUed than 20 %ld6. For
example, 20.....c7 21 .rlc1 .i.xc3 22 bxc3
llc7 23 lidS+ %lxd8 24 %lxe7 b5 25 �3
(after 25 %lxc7 bxc4 26 %lxc6 c!tlb4! 27 %ld6
%lxd6 28 .ixd6 ltJxa2 Black's dangerous a­
pawn gives him rhe edge) 25 ...'iff4 and Black
had serious coumerplay in Magem-IUescas.
France 2000.
20 .. ixc3 21 bxc3 h6 22 ha711'
This trick gains White a small advanrage.
22... llK:71 23 J:le7 •d5 24 1i'xd5 �xd5
After 24...cxd5 25 :Xf7 q;xf7 26 .id4
%lxa2 27 :lbl c!tlb5 28 .ie5 Yusupov still
believes White is a bit better.
25 J:le8+ J:lfB 26 llxf8+ Wxf8 27 -*.c5+
�8 28 c4 lDf4! 29 -*.d6
After this Black equalises. White could
keep pressing with 29 .ie3! ttld3 30 %le2
c!tlb4 31 f3 %lxa2 32 Axa2 c!tlxa2 33 .i.d2 Wf7

'4 Wf2 �e6 35 «li'e3 Wd6 36 Wd4 (Yusu­
pov).
29...liJd3 30 lte2 :dB 31 i.c7 ltd4 32
g3 % -%
Ga1f/e 14
Sax-Vusupov
Thessaloniki Ofyntpiad 1984
_______________.
1 e4 e5 2 liJf3 lilf6 3 liJxe5 d6 4 liJf3
�xe4 6 d4 d5 6 i.d3 liJc6 7 0-0 i.e7 8
J:le1 i.g4 9 c4 liJf6 1 0 cxd5
10 i.xf3
••.
This is much safer than 10 ...'iFxd5 1 1 lllc3
.ixf3 Qf 1 t ...'iFh5 1 2 i.b5 ..txf3 13 gxf3 �fB
14 i.xc6 bxc6 15 i.f4 White's easier devel­
opment is worth a slight advantage) 1 2 lC!xd5
i.xd1 1 3 lllxc7+ Wd7 14 i.f41 ..tg4 (Kavalck
nnnlvsed 14...lC!h5 15 i.f5+ Wd8 1 6 i.c5
li'lx�S 1 7 lllxa8 lllc6 1 8 .llaxdt i.d6 1 9 Zl.e5!
as winning for White) 1 5 d5 lCJd4 1 6 lllxa8
l:lxa8 1 7 i.c51 .i.f5 (or 1 7 ... ..tc5 1 8 llect
.ib6 19 llc4 and White still has a clear ad­
vantage) 18 i.ft lllc2 1 9 i.bS+ Wd8 20 d6
ll'lxct 21 llxe l i.e6 22 dxe7+ Wxe7 23 .i.d4
h6 24 a4 (Kavalck-Toth, Haifa 1976), when
rhc tactics have burned out but White's
llishop pair still ensures a small advantage.
1 0.. lllxd5 is studied in Game 1 5.
. continued 1 8.....ib4 1 9 lled1 g6 20 i.h3
1 1 •xf3 •xd5 12 •g3
White has two other moves:
a) 1 2 .,3 lllxd4 (also playable is
12 ....xd4 13 lllc3 l:td8 14 i.b5 0-0 15 i.xc6
3 liJxe5: The Main Line with 8 :e 1
bxc6 16 .i.g5 l:tfe8 with an unclear position)
13 lllc3 'iFd7 14 'iFxd7+ �d7 1 5 .i.e3 llle6
1 6 l:tad1 .i.d6 17 .i.f5 We7 1 8 lllb5 Zl.hd8 1 9
lllxd6 cxd6 20 h 3 (White has obvious com­
pensation for the pawn) 20...b6 21 g4 h6 22
.i.d4 l:tac8 23 .i.c3 g6 24 .i.c2 hS 25 f3 1/z-1/z
Kasparov-Karpov, World Championship
(Game 28), Moscow 1984.
b) 12 'iFxdS should not be dangerous:
12 ...lllxd5 13 lllc3 0-0-0 (Black is well ad-
vised not to chase material with 1 3.. llldb4 14
.
.i.e4 lllxd4 because after 15 .i.e3 lld8 1 6
..ixd4 l:r.xd4 17 a3 lllc6 18 lllb5 White will
win back the pawn and force an advantage)
14 .i.e4 i.b4! 1 5 i.d2 lC!f6 1 6 i..xc6 bxc6 17
..ie3 lCJds 18 a3 (Kamsky-Karpov, Linares
1 994 continued 1 8 %tact lllxd 1 9 fxe3 cS 20
l:tft f6 21 llf5 Zl.he8 and Black had an edge,
which grew considerably after 22 dxcS?
llxe3) 1 8...i.xc3 19 bxc3 lC!xc3 20 llact
lC!bS 21 llxc6 lC!xd4 22 Zl.c4. Despite Black's
extra pawn Karpov assessc..-s the position as
el(Ual.
1 2...11fxd4 1 3 lilc3
After 13 'iFxc7 one of BL1ck's options is to
force e']uality with 1 3...lllg4!? 1 4 'iff4 'ifxd3
1 5 'ifxg4 0-0.
1 3 0-0 14 liJb6 .g4
...
Black has a sound alternative in 14 ...1i'b41?
I S i.gS llfd8 1 6 'iFh3 '!fl 1 7 lllxc7 IL1cR 1 8
i.xf6 .i.xf6 1 9 llld s 'iFd6 20 lllxfM 'ifxf6,
achieving easy equality.
1 5 11fxg4
1 5 lllxc7 Zl.ad8 1 6 'ifxg4 lllxg4 1 7 i.e2
lC!ge5 1 8 i.f4 ..if6 is also level.
1 5...liJxg4 1 6 i.f5?1
Ths is inaccurate; 16 ..te2 lC!f6 17 lllxc7
llac8 l8 lllb5 ..ib4 is simply eCJual.
1 6 �f6 1 7 liJxc7 lladB 1 8 i.e3 a61
•••
A later game, Kamsky-Khalifman, FIDE
World Championship, Las Vegas 1999,
a6 (or 20....i.d6 21 lllbS ..ie.S 22 lllxa7 lllxa7
23 .i.xa7 Zl.xdl+ 24 llxdl l:ta8 25 -'.c5
..ixb2, again with equality) 21 i.h6 llxdl+ 22
Zl.xd 1 l%d8 23 llxc.l8+ lllxd8 24 i.g5 i.e7
37
The Petroff Defence
with a level position.
1 9 1lac1
Perhaps White should try 1 9 :ed1 .i.d6
20 .i.b6 .i.e5 21 llab1 (after 21 l:txd8 :Xd8
White has some problems since 22 l0xa6?
fails to 22...:d5) 21 ..J:lxd t + 22 :xd1 .i.xb2,
when White's activity is enough to hold the
balance.
1 9....tb41 20 llf1 lbd4!
Ths keeps up the pressure. 20..id2?! al­
lows White to complicate matters with 21
l:lxc61 bxc6 22 .i.c5.
21 llc4?1
Now Black's advantage is serious. After 21
.i.xd4 l:lxd4 22 l:lfdt llfd8 23 llxd4 :Xd4 24
�fl .i.d6 Black has very little.
21 ...o!bxt5 22 llxb4 l:td7! 23 :C1
Black's advantage is confmned by the fact
that White cannot grab on b7: 23 llxb7?
loses to 23...l0xe3 24 fxc3 llc8 25 llcl l0e8,
while the line 23 .i.f4? lieS 24 llxb7 li)d4! is
also hopeless.
23 . .llc8 24 llbc4 l:tcd8 25 h3 lbxe3 26
.
fxe3 �8 27 e4 �7
28 l:tb4
If 28 c5 then 28 ..l0d5 29 l£lxd5+ llxdS 30
. a) 1 2...i.h5? is just a blunder: 1 3 .i.xh7+!
llg4 g6 and Black has good chances of win­
ning the rook ending.
28...l:td1 + 29 llxd1 llxd1 + 30 �2 Wd6
31 e5+?
Now it is all over. Also hopeless is 31
l:lxb7 'ii>c6 32 lla7 �b6. but White can put
up more resistance with 31 �e2 llat 32 lft!5
38
l0xd5 33 cxd5 b5.
31 Wxe5 32 lDa8 b5 33 a4 o!bd5 34
•••
llb3 bxa4 35 llb7 l:tb1 36 �3 a3 0-1
Gan1e 1 5
Psakhis-Cooper
Pott Erin 2003
1 e4 e5 2 o!bf3 lDf6 3 lbxe5 d6 4 o!bf3
o!bxe4 5 d4 d5 6 .td3 lbc6 7 0-0 .te7 8
lle1 .tg4 9 c4 lDf6 1 0 cxd5 fud5 1 1
o!bc3 0-0 1 2 h3
Twenty years earlier Ps.1khis had tried 12
.i.e4 .i.e6 (after 1 2..l0f6 Zak assessed 1 3 d5
l0b4 1 4 a3 l0xe4 15 llxc4 .i.x£3 16 'iVx£3
li)a6 17 b4 as clearly better for White) 13
'it'd3 h6 14 .i.h7+ (if 14 'lt'b5 Black draws
with 14 ...�b4!? 1 5 'lt'xb7 l:lb8 1 6 '1t'xa7
liaS) 1 4...�h8 1 5 .i.fS l0cb4 1 6 -.,1 .i.xfS
17 .xfS l0f6 18 .i.f4 l0bd5 19 .i.e5 c6 with
an equal position, Psakhis-Schussler, Tallinn
1 983.
1 2....te6
Ths is the solid option, but two others
moves are worth mentioning:
�xh7 14 lbgS+ .i.xg5 1 5 'lt'xhS+ �g8 1 6
.i.xgS l0 f6 1 7 .i.xf6 'it'xf6 1 8 llkls '1t'xd4 1 9
lt.d1 br6 20 Axd4 gxh5 2 1 llh4 and White
had a huge advantage in Yemelin-Ka?.akov,
St Petersburg 1996.
b) More serious is 1 2... -*.x£3, although
White should have a slight edge. For exam-

pic, 1 3 'i6'xf3 ll:ldb4 14 .ibl l:e8 I S dS .id6
I (, l:xe8+ 1fxe8 17 .ie3 lbes 1 8 'i6'e4 ll:lg6
1 9 a3 ll:la6 20 i.c2, Dolmatov-Yusupov,
1 1nm7.c 1 98 1 .
13 a3 l/Jxc3
Black has also tried the immediate
l l..i. f6, for example 1 4 &4 .t.r.; 1 S Wb3
�b6 1 6 dS .ixe4 17 .i.xc4 ll:le7 1 8 .i.e3
li.)cc8 (the greedy 1 8 ...lLlexdS docs not work:
I IJ .ixb6 lLlxb6 20 .i.xb7 l:b8 21 l:tad1 lL!c.I7
22 'il'bs l:te8 2.1 :Xd7 l:txel+ 24 lLlxet •e8
25 �f3 l:txb7 26 ...xb7 1fxd7 27 ..xa7 and
White has good winning chances) 1 9 l:tacl
�16 20 .i.bt 'i6'd7 (Shirov-I.Sokolov, FIDE
World Championo;hip, Las Vegas 1 999).
Now Ivan Sokolov suggests that 21 .i.gS!?
i.xgS 22 lLlxgS g6 2.1 ...g3 gives an edge to
White.
14 bxc3 ..i.f6
1 5 Ab1
This is natural, but playing IS 'i6'c2 g6 and
only then 1 6 l:tbl is also sensible. Now if
Black plays as in the game with 1 6 ... b6?!,
White has 1 7 .i.xg6! hxg6 1 8 l:txe6 fxe6 1 9
.x�:,r6+ �h8 20 llbS with a crushing attack.
Instead Vasiukov-B.Vladimirov, USSR 1981
continued 16 ....i.dS 1 7 lLld2 b6 1 8 lLlc4 lOc7
1 9 lLle5 .i.g7 20 c4 .i.b7 21 .i.b2 and White
hnd a pleasant advantage.
15 b6 1 6 ..i.f4
.••
White could also try 16 ...c2 h6 1 7 .i.f4
.idS 1 8 .i.h7+ <t>h8 1 9 .i.e4 with a promis­
ing position.
3 l/Jxe5: Th e Main L in e with 8 Ae 1
1 6 .. .'�h8?!
Black could have limited his disadvantage
with 16 ...lL\e7 1 7 lLlcS lLlc..iS 1 8 .i.d2.
1 7 ..i.e4 ..i.d7 1 8 l/Je5
Now Black has to make a concession.
1 8. . ....i.xe5 1 9 dxe5 •ea 20 ..i.g5 •e6
After 20 ..h6 White docs not need to re­
.
treat: 21 .i.f6! l:tb8 22 9h5 �g8 23 lte3 and
the attack is devastating.
21 �5 h6 22 ..i.f61 �g8 23 Ae3 l/Je7
24 Ag3 l/Jg6 25 Axg6 fxg6 26 •xg6 1 -0
Game 16
Leko-Yusupov
Dottmund 1998
1 e4 e5 2 liJf3 li.:lf6 3 l/Jxe5 d6 4 lDf3
lDxe4 5 d4 d5 6 ..i.d3 ..i.e7 7 0-0 lDc6 8
Ae1 ..i.g4 9 c3 f5
This is more aggressive than 9. .ll:\f6 lO
.
.i.gS 'i6'd7?! (Black should play 10...0-0 but
White's position is preferable after 1 1 lbbd2
.i.hS 12 Wb3 lLlas 1 3 'i6'c2 .i.g6 1 4 lLleS) 1 1
lL!bd2 (White could also choose to shatter
Black's kingside with 1 1 .i.xf6: White is bet­
ter after 1 1 ...gxf6 1 2 lLlbc12 0-0·0 13 .i.bS)
1 1 ...0-0-0 1 2 'i6'a4 h6 1 3 .i.h4 g5 1 4 .ig3
.i.xf3 (or 14 ... i.d6 1 S .i.bS and Black is still
in trouble because 1 S ...�b8? loses to 16 lL!eS
i.xeS 17 .i.xeS) 1 5 lLlxf3 g4 1 6 lbest lL!xeS
1 7 i.f51 1i'xf5 1 8 l:txeS ...d3 (l!L1i'd7 1 9
'i6'xa7 'i6'd6 20 l:te3 is also winning) 1 9 l:txe7
:d7 20 l:te3 1i'a6 21 .xa6 bxa6 22 .i.eS 1 -0,
Keres-Alexander, Hastings 1 9S4/S5.
1 0 lDbd2
1 0 h3 i.xf3!? offers a promising pawn
sacrifice: 1 1 gxf3 (if White declines the offer
with 1 1 ...xf3 0-0 1 2 ..i.£4 .i.d6 the game is
unclear) l t ...lLlf6 1 2 i.xf5 0-0 13 Wd3 i.d6
14 .igS llJc7 1 5 .i.e6+ �h8 1 6 lLld2 lL!g6 17
.i.f5?l (this lets Black take control; White had
to try 1 7 lL!b3 even though Black has com­
pensation after 17 ... .i.f4 1 8 •rs .i.xgS 1 9
'ifxgS 1i'd6) 1 7....i.f41 1 8 .i.xf6 (or 1 8 .i.xg6
i.xg5 1 9 ..i.f5 lL!hS and Black's occupation
of f4 will be very threatening) 1 8......xf6 1 9
39
The Petroff Defence

.ixg6 .txd2 20 :ez .i.f4 21 .i.xh7 'Wg5+ 22

�h1 1ih5 0-1 Enklaar-Dvorctsky, Wijk aan
Zee 1 975.

1 1 ...0-0?1
Now castling short is a mistake - Black
has two better moves:
1 0.....d6 a) 1 t ...ltlxd2 12 .i.xd2 dxc4 (not
Possibly preparing to castle long. Instead 12...ltlxd4? 13 lihd4! .i.xdt 1 4 ltlxf5 1fc5 1 5
1 0... 0-0 is also reasonable, with the foUowing: b4, when White wiU have a decisive material
a) t 1 '1Fb3 transposes to Game 17. advantage) 1 3 i.xc4 0-0-0 14 .i.c3 .i.f6 with
b) 1 1 lbft with a further split: an unclear position.
bt) t t ....i.d6 is tricky but probably inaccu­ b) 1 1 ...0-0-0!? leads ro sharp play: 12 cxd5
rate after 1 2 ll)g3 (not 12 lbe3? .i.xh2+1, ltlxd2! (once again 12 ...ltlxd4? is a mistake:
when 13 �? is imposs.ible because of 1 3 lihd41 .txd1 14 ltlxfS ..eS 1 5 ltlxe7+
13...l£lxf2 1 4 ..c2 ltlxd3 1 5 1Wxd3 .i.xt3 ..xc7 1 6 .i.xe4 and White's extra material
winning for Black - the problem with 1 6 ensures a clear advantage) 1 3 dxc6 ltlxf3+ 14
gx f3 is 16...1ih4+, winning the rook) 12.....f6 gxf3 i.h3 1 5 cxb7+ Wb8 16 .tn i.xfl 17
1 3 '1Fb3 and White is a bit better. D.xft i.f6 with a very mes.c;y position (Leko).
b2) t t .. .th4 12 g3 .tgs 13 .txg5 l£lxg5
. 1 2 cxd5 c!Dxd4
14 .i.e2 l£lh3+ 1 5 �g2 f4 16 ltl3d2 .i.xe2 17 White a bit better after 12...l£lxd2 1 3 dxc6
1fxe2 ltlg5 1 8 1Wg4 fxg3 19 hxg3 ..f6 with l£!xf3+ 1 4 gxf3 i.hS 1 5 cxb7 llab8 1 6 i.c4+
an unclear pos.ition, Ljubojevic-Hjartarsson, ll.f7 17 'irb3.
Belgrade 1989. 1 3 .*.xe4l fxe4 1 4 1llxe4
1 1 c41?
The other option is 1 1 ltlft. For example,
1 1 ...0-0-01? 1 2 ltk3 h5!? 1 3 h3 g6! 14 hxg4
(mstead 1 4 i.c2 i.xt3 1 5 i.xf3 .th4 1 6 1le2
lbe7 1 7 l£lf1 i.f6 1 8 1lc2 'itb8 19 a4 1lh7 20
aS g5 was unclear in Adams-Makarychev,
Oviedo 1992) 14...hxg4 15 .txc4 (White
must avoid 1 5 ltld2? because of 1 5...11fh2+
1 6 'itft ltlxf21 17 'iPxf2 .th4+ 18 Wc2 1ldc8
with an immense attack) 1 5...dxc4 1 6 ltlc4
..f6 and Black has good compensation - 1 7
.i.gS? loses to 1 7...1fg7 1 8 Le7 1ih6 19
l£lh4 l£lxc7 20 g3 g5.

40

14 .i.xf3
..•
Leko assessed 14...�xf3+ 1 5 gxf3 .i.xf3
H i �xd6 .i.xdl 17 �xb7 l:labB 18 l:he7
:xb7 1 9 d6 as clearly better for White.
1 5 1i'xd4 1fb4 1 6 1t'xb4 ..txb4 1 7 .l:le3
.i.h5 1 8 .l:lb3 .i.a5
White also has a clear plus after 1 8...a5 19
a."\ .i.g6 20 f3 .i.xc4 21 fxe4 .i.d6 22 .i.d2.
19 ..te3 ..tf7 20 .!Lig5
Or 20 llxb7 .i.xdS 21 :lbS .i.xc4 22 l:lxaS
n6 23 :lc1 and, despite the opposite-coloured
bishops, �'hitc has an obvious advanta�:,>c.
20 ..tb6 21 .!Lixf7 .l:lxf7 22 ..txb6 axb6
.••
23 d6! cxd6 24 .l:lxb6 .l:lc7 25 �1 .l:la6
26 .l:lb4
The direct 26 :Xa6 should also win. For
example, 26 ... bxa6 27 b4 :lc2 28 a4 :lb2 29
bS axbS 30 aS l:lc2 31 a6 :lc8 32 a7 :laB 33
.Ji!c2 etc.
26 �7 27 a3 d6 28 .l:ld1 �6 29 .l:lb5
•..
lld6 30 'it>e2 'it>e5 31 'it>d3 .l:ldc6
32 �3
Or 32 :lel+ �d6 33 l:le8 :lcS 34 l:lxcS
l:.xcS 35 l:lb8 and White's advantage is t:vi­
tlcnt.
32 .l:lc5 33 .l:lxc5 .l:lxc5 34 'it>d3 .l:lc4 35
..•
lle1 + Wd6 36 .l:le8 .l:lf4 37 f3 .l:lh4 38 h3
b6 39 .l:lb8 'it>c5 40 .l:lc8+ 'it>d6 41 .l:lf8 g5
42 .l:lf6+ 'it>c5 43 b4+ 'it>b5 44 .l:lf5 �a4
45 .l:lxg5 'it>xa3 46 .l:lxd5
46 bS �b4 47 l:lxdS is even more straight­
forward.
46 .Axb4 47 .l:lh5 llb3+ 48 �e4 .l:lb4+
••
3 '!Jxe5: The Main L ine with 8 .l:l e 1
49 'it>e3 .l:lb3+ 50 �4 .l:lb4+ 51 'it>g3
White prefers to avoid the pawn ending
51 WgS ltbS+ 52 �h6 1lxh5+ 53 WxhS even
!hough he wins neatly after 53 ... b5 54 f4 b4
ss 5 b3 56 f6 b2 57 fl b1W ss f89+ Wa2
59 Wa8+.
51 b5 52 h4 .l:lb1 53 .l:lxh7 b4 54
.••
.l:la7 + �b2 55 �4 .l:lh1 56 'it>g5 b3 57
h5 'it>c3 58 .l:lc7 + �d4 59 .l:lb7 Wc3 60
h6 b2 61 g4 b11i' 62 .l:lxb1 .l:lxb1 63 h7
.l:lb8 64 �6 .l:lb6+ 65 �g7 .l:lb7+ 66
¢>h6 .l:lb6+ 67 Wh5 .l:lb8 68 g5 1 -0
Game 1 7
lvanchuk-Shlrov
Dortmund 1998
1 e4 e5 2 ltlf3 .!Lif6 3 d4 ltlxe4 4 .!Lixe5
d6 5 .!Lif3 d5 6 ..td3 &6 7 0-0 .i.e7 8
.l:le1 .i.g4 9 c3 f5 1 0 'iJbd2 0-0 1 1 'irb3
11 •.• Wh8
Sacrificing a pawn. 1 1 ...�5 will be stud­
ied in the next game.
12 ..xb7
White can opt to include 12 h3 .i.hS be­
fore playing 13 Wxb7. For example, 13 ...l:lf6
1 4 9b3 l:lg6 (or 14 ...g51? 1 5 .!?lfl l:lbB 1 6
Wc2 with a murky position) 15 .i.e2 .i.d6
(1 5 ... .i.h4?! is the wrong way to build up the
attack: 1 6 l:ln .i.xf3 1 7 o!?lxf3 and White is a
safe pawn up because 17....i.xf2+?! fails to 1 8
llxf2 .!?lxf2 1 9 � f2 Wd6 20 o!?lgS! l:lf8 21
9a3 Wd8 22 .i.f4 h6 23 o!?lf3 lle8 24 .i.d3,
41
Th e Petroff Defence
Katpov-Korchnoi, 6th matchgame, Moscow
1 974) 1 6 ltk5 I!Dxe5 17 .i.xh5 ltxg2+! 1 8
�xg2 .gS+ 19 �ft 1Wh4 20 I!Dxe4 (not 20
ltxe4? 11'xh.J+ 21 Wg1 �! and Black wins)
20...1i'xh}t 21 �g1 I!Df3+ 22 Lf3 .i.h.2+
and Black has a perpetual check (O'Kelly).
1 2...1lf6
1 3 ...,5
The other retreat 1 3 'tl'b3 also allows
Black counterplay: 1 3 ...ltg6 (1 3...ltb8 14 .a4
simply transposes to the game) and now:
a) 1 4 .ib5 I!Dxd2 1 5 I!Dxd2 .i.d6 1 6 g3 (or
1 6 .ixc6 Lh.2+1 17 Wxh.2 1Wh4+ 1 8 �gl
lth6 1 9 f3 •xel+ 20 .!Dn lthl+l 21 �ht
.xft+ 22 Wh2 .i.xf3 and now White should
accept that Black has a perpetual check be­
cause if 23 'ii'c2? .ie4 24 .d2 :ciS Black has
a deadly attack) 1 6 ...ltk7 1 7 .i.d3 h5 1 8 .!Dfl
h4 and Black has a dangerous initiative,
Peshkov-Raetsky, correspondence 1985.
b) 14 g3!? 1:lb8?! 1 5 .c2 .i.d6 1 6 b41
(Movsesian pointed out that the imprecise 16
.ift ..f6 17 .i.g2 ltfB would allow Black un­
deserved compensation) 16.JWf6 17 b5 ltk7
18 � .ixeS 19 dxeS and White was clearly
better in Movsesian-Kroeze, Nethedands
2001 . More accurate is 1 4...•d6 1 5 .ib5 ltb8,
when Black has some compensation.
1 3 1lb8 1 4 1ra4 i.d6 1 5 h3
•..
Or 1 5 .ibS I!Dxd2 1 6 .i.xd2 .ixf3
(1 6...ltk7?! is too passive: 17 �5 a6 18 .if1
:xb2 1 9 .tgs .i.h5 20 .ixf6 gxf6 21 •d7 is
much better for White, Gagunashvili-
42
Bayramov, Batumi 2001) 1 7 .ixc6 tlg6 1 8 g3
.i.e4 and Black's kingside play provides ob -
vious play for the pawn.
1 5....th5 1 6 .i.e2 llg6
1 7 Wf 1 17
Trying to escape the pressure along the g­
file. In this complex position White has two
other main options:
a) 1 7 I!Dxe4 leads to very unclear play after
1 7 ... fxe4 1 8 li)gs :Xgs 1 9 .ixgs •xgs 20
.i.xhS lik71 21 .d 1 g6 22 .i.g4 h5 23 .i.d7
lids 24 .i.b5 I!D5.
b) Curiously, 1 7 lbcs allows Black a
choice of forced draws: 17 ...�e5 1 8 dxe5
.ixe2 19 exd6 ltxg2+! 20 Wxg2 .gS+ 21
�h 1 I!Dxf2+ 22 Wh2 f4! 23 lOft � 24
hxg4 1Wh4+ 25 Wg2 1i'xg4+ with a perpetual
check, or 1 8...•g5 1 9 .if1 I!Dxf2! 20 exd6
.!l)xh.}t 21 Wh1 I!Df2+ (lvanchuk).
1 7 -*.14 1 8 �3 .i.xc1 1 9 llaxc1 •d6
.•.
20 1lc2
The ambitious 20 .ib5?1 backfires after
20.. .:Xg2! 21 Wxg2 ..g6+ 22 Wh2 .ixf3 23
ltgt 1i'd6+ 24 .l:[g3 �7 25 :n hS, after
which Black's attack is very strong.
lvanchuk analysed 20 I!Dh4 as leading to a
draw: 20...•f4 21 I!Dxg6+ hxg6 22 f3 l0c5! 23
1i'xa7! I!Dxf3! 24 1i'xb8+ Wh7 25 .ixf3 .ixf3
26 :Xe4 .ixe4+ 27 �e2 .g3 28 .l:[g1 .d3+-
29 Wf2 f4 when to avoid mate White must
play 30 .!DeS and allow perpetual check.
20 11ff4 21 ,j_d1 1le8 22 libl2
.••
It is vital to support f3. If 22 �5?! then

22... ..i.xf3 23 ..i.xf3 'Wh2 24 g3 .1Lg31 25
.lxe4 'ti'xh3+ 26 We2 .J:g6 and Black's attack
i� venomous.
22 1lge6 23 liJxe4 fxe4 24 g3! 1tf5 25
•..
t;)g1 .i.xd1 26 llxd1 e3?1
This is the wrong approach. 26...h5 leads
In an unclear position after 27 c4 e3 28 cxd5
1i'xd5 29 'it'c4 liJe7. However, the best line is
1hc simple 26...l:tf6 27 l:te2 l:tcf8 when Black
has a definite edge.
27 �2 1We4+?!
One slip follows another... After 27 ...l:tffi
2flf3 fi:k7 29 l:te2 White would have b�o.-cn
only slightly better.
28 f3 1Wf5 29 h2 1lg6 30 llde1 1Wf4 31
g4 h5 32 'lrb51 hxg4
A better try is 32 ... l:tcd)!? when lvanchuk
analysed 33 'lrxd5 hxg4 34 hxg4 l:tx�:,t4+ 35
fKg4 'ifxg4+ 36 �h 1 l:th6+ 37 l:th2 llxh2+ 38
Wxh2 'irh4+ 39 liJh3 'it'xe1 40 'it'hS+! Wg8
3 liJxe5: The Main L ine with 8 lle 1
41 'it'e8+ �h7 42 'it'xc6 'ti'd2+ 43 'iVg2 e2 44
liJgS+ 'it'xg5 45 'it'xg5 e11i' 46 'it'hS+, with
White having reasonable winning chances.
33 hxg4
After 33 fxg4? Black can draw with
33...Wfe4+ 34 tll f3 l:txg4+! 35 hxg4 'ifxg4+ 36
'ith l 1i'xf3+ 37 .llg2 c2 38 �gl .e3+ 39
'itht •n.
33 ...llh6 34 1txd5 /:i)e7 35 •e5
White had a faster win with 35 'ild7! l:tb8
36 'ifxe7 llxb2 37 l:lxb2 llh2+ 38 Wfl l:txb2
39 l:lxe3.
35 ....-xe5 36 dxe5 /:i)d5 37 �g3 lie&
38 /:i)h3 llxe5 39 /:i)f4 /:i)xf4 40 �xf4 lla5
41 a3 llb5 42 �xe3 llcb6 43 b4 liaS 44
Jla1 c5 45 llb2 cxb4 46 llxb4 llba5 47
a4 llc5 48 Jla3 1-0
Game 18
Anand-Gelfand
Moscmv 2004
1 e4 e5 2 .!Llf3 .!tlf& 3 c!Dxe5 d6 4 .!tlf3
o!Llxe4 5 d4 d5 6 .i.d3 .i.e7 7 0-0 lbc& 8
l:te 1 .i.g4 9 c3 f5 1 0 'lrb3 0-0
A major alternative here is lO...'ifd6 with
the idea of ca!ltling lJUeenside. White has
three main answers:
a) Taking the pawn with 1 1 'lrxb7 should
certainly be analysed, but Black has good play
after 1 1 ...l:tb8 1 2 'ifa6 ..i.xf3 1 3 gxf3 0-01? 1 4
fxe4 (not 1 4 ..i.bS?! l:tb6 1 5 'lra4 .i.h4 with a
dan�rous initiative; for example, 1 6 fxe4?
43
The Petroff Defence
allows a winning attack after 1 6....txf2+! 17
�£2 9xh2+) t 4...fxe4 1 s .tn .th4.
b) 1 1 .!Dfd2!? is interesting but also poten­
tiaUy risky: 1 1 ...0-0-0 1 2 f3 i.h4 (the tricky
12 ....!De5?! fails to 13 i.xe4 dxe4 1 4 fxg4
.!Dxg4 1 5 .!Df1 and White is clearly better -
Timman) I 3 :n and now:
bl) The creative 1 3...i.h3?! is refuted by
accurate defence: 1 4 ..c2 11'g6 1 5 .!Db3 :h£8
16 .!Da3 llde8 17 i.f4! {but not 17 Wht??
.!D£2+! t 8 llxf2 i.xg2+! 0-1 Ivanchuk-Anand,
Reggio Emilia 1 989) 17 . .:e6 18 .!DbS and
.
White has a clear advantage.
b2) 1 3 ...i.f2+!? 1 4 :xf2 lOx£2 1 5 Wx£2
..xh2 1 6 .!Dn 11'h1 1? 1 7 fxg4 fxg4 and the
situation is impossible to assess with any
confidence: White has a large material advan­
tage but a dangerously exposed king.
c) 1 1 .!Dbd2 is the safest option. 1 1 ... 0-0-0
12 ..c2 (12 .lb5is less clear, for example
12..if6 13 1i'a4 :he81? I 4 i.xc6 bxc6 I 5
l0xe4 :xe4 16 :xe4 fxe4 17 .!Des i.xeS 1 8
dxeS ..cS 1 9 i.e3 'irbs with an unbalanced
position) 12..Wb8 13 b4 i.f6 1 4 bS .!De7 1 S
.!Des i.xeS 1 6 dxeS 1Wg6 1 7 .!Db3 lbgs 1 8 f4
.!De4 19 i.e3 and White had an edge in S7.e­
lag-Naumann, Rostock 2002.
1 1 lbbd2 lba5
1 2 Wc2
12 1Wa4 is a very important alternative.
Play continues 12....!Dc6 1 3 i.bS and now:
a) White has an easy time after 13 ....!Dxd2?!
14 .!Dxd2 '1Fd6 I S h3 i.hS 16 .!Db3 i.h4 17
44
.!DeS i.xf2+ (White is also clearly better after
the alternatives 17...f4 1 8 :e6 9d8 19 l0d3
f3 20 g3 and 17 ...lbd8 18 .tn b6 19 .!Ddl
.!De6 20 1tb4! - Anand) 1 8 Wxf2 ....2 19
i.xc6 bxc6 20 11'xc6 f4 21 ..xdS+ Wh8 22
1WxhS f3 23 11'xf3! :Xf3+ 24 Wxf3 and White.
has cleverly consolidated his advantage, An·
and-Kramnik, Tilburg 1998.
b) 1 3 ...i.h4! with a further branch:
.I
bl) 14 g3 (this is usual) 14 ... i.f6 t S Let
bxc6 1 6 1i'xc6 :es 17 .!Des i.xeS 18 dxe5':
lbgs 19 f4 (Kramnik mentions 1 9 c4!? d41�
20 WdS+ 1WxdS 21 cxd5 llad8 as still giving
compensation) 1 9....!Dh3+! (White has an;
edge after 1 9...:C6?!, for example 20....!De4?1;
21 l0xe4 dxe4 22 .i.e.� a6 - Shirov-Kramnik.j
Belgrade 1 999 - and now 23 :OCt i.f3 24J
:c2 would have clarified White's advantage).;
20 Wg2 :b8 21 c4 (21 .!Db3!? :b6 22 -�
d4! 23 c41 •c8 is messy) 2t ...dxc4 22 lbxc4
1t"d3 23 tbe3 (White has to watch out fen'
mating trick.o;: 23 i.d2? :eds 24 .tc3 •cZ+
25 Whl ..a4!) 23 .. 1E.xb2+! 24 .i.xb2 ..dz..+;
25 Wht l0f2+ 26 Wg1 .!Dh3+ 112-lfz Anand�
Kramnik, Wijk aan Zec 1999.
1
b2) 1 4 :n !? is also worth considering;
14....!Dxd2 1 S .!Dxd2 f4 1 6 i.xc6 bxc6 11,
.xc6 .gS 18 c4 (18 g3 allows Black to fore �
a draw with 18...fxg3 19 hxg3 .txg3! 20 fxg�
1i'e3+ 21 Wg2 9e2+) 18...1lad8 19 cxdS :d�
20 1Wc3 llg6 21 .!Df3 Whs 22 .!Des f3 �
lbxg6 hxg6 24 .i.e3 .i.c8! is totally unclear�
j
Koziak-Motylcv, Lvov 1999.

1 2....i.d6
This is sttongcr than 1 2...c5 1 3 llk:S .i.hS
(not 1 3 ...cxd4?1 14 �xg4 dxc3 1 5 bxc3 fxg4 1 6
ti)xe4 dxe4 17 .i.xe4 i.c5 1 8 i.xh7+ �h8 1 9
.ic3 when White defends easily) 1 4 t3 cxd4
1 5 fxc4 fx<:4 (1 5...dxc3 is tricky but White
keeps the initiative with 16 cxd5!? cxd2 17
.ixd2 11Vxd5 1 8 i.xaS 11Vxa5 t9 li)d7 1lfe8 20
lieS) 1 6 �e4 dxe4 17 .i.xe4 h6 (Anand­
Uoi<Jb>aO. Wijk aan Zee 2004) and here White
can gain a large advantage with 18 b41 dxc3 19
!Jxa5 1id4+ 20 �hl 11Vxe5 21 .i.a.�!.
Black has also tried 14....i.h4 but White
3 lbxe5: The Main Line with 8 lle 1
keeps an edge after 15 g3 �g3 1 6 hxg3
i.xg3 1 7 lle2, whereas 1 5 fxe4 i.xel 1 6 cxfS
c4 17 i.fl .i.xd2 18 .i.xd2 �c6, Moro­
zevich-Adams, Dorttnund 2002, is rather
unclear.
1 3 b4
This is reasonable but perhaps White
should prefer 1 3 �5!? .i.xeS 1 4 dxe5 11Vh4
1 5 llfl with a pleasant edge.
1 3... ..i.xf31 14 l'Oxf3 l'Oc4 1 5 ..i.xc4
Or 1 5 a4 c6 1 6 11Vb3 b5 17 axb5 cxbS 1 8
lla6 llf6 with unclear play.
1 5 dxc4 16 ..i.e3 J:leB 1 7 J:lad1 •te 1 8
•••
lbe5 ..i.xe5
This is the obvious choice but 1 8.. .Axe51?
is also interesting, 1 9 dxe5 '1Vxe5 20 g3 1ixc3
leading to a messy position.
1 9 dxe5 •xe5 20 .i.d4 'MI5 21 a4 •d7
White also has compensation for the
pawn after Black's two other options:
21 ...1i'a6 22 f3 �6 23 .i.c5 and 21 ...11Vc6 22
f3 �6 23 b5 'W'd7 24 .i.eS llad8 25 11Vf2.
22 ..i.xa7 rod& 23 J:lxe8+ •xeS 24 .i.c5
•xa4 25 •xa4 J:lxa4 26 .i.xd& cxd6 27
Wf1 .IZa& 28 J:ld4 b5 29 J:ld5 J:la3 30
J:lxd& J:lxc3 Y.z Y.z
-
45
Th e Petroff Defence

Summary
Nowadays the active continuation 9 c4 docs not seem to offer White an advantage in the
opening. It can be recommended to fans of the isolated quecn's pawn, which arises after
9 ... l0f6 10 cxd5 l0xd5. However, we should not forget that 1 0... .A.xf3 1 1 1i'xf3 l0xd4 (Game
1 4) has proven to be a reliable way of equalising.
Despite initial appearances, 9 c3 is one of the sharpest system� in the Petroff Defence. Of
course, opposite side castling after 9. . . f5 1 0 'irb3 1Vd6 and ...0-0-0 leads to many complica­
tions. After 1 0 ...0-0 11 l0bd2 l0a5 Black maintains material equality but loses control over the
important eS-square, and following 1 2 "it'c2 .i.d6 (or 1 2. ..c5) 1 3 lOeS White's chances are pref­
erable. The line with 1 1 ...'�h8 (Game 17) is extremely intricate and every move must be care­
fully analysed. After 1 2 1Wxb7 Black's queenside becomes (.."VCO more vulnerable but the offside
position of White's queen exposes his king to danger. Black has .. .llf8-f6-g6 at his disposal
increasing his initiative on the kingside, and despite many games it is still impossible to give an
exact assessment of this position.

1 e4 e5 2 lDf3 lDf6 3 lDxe5 d6 4 .IL!f3 lllxe4 5 d4 d6 6 i.d3 i.e7 7 0-0 .IL!c6 8 .lle 1
i.g4 (D) 9 c4
9 c3 f5
10 l0bd2 - Game 16
10 1Wb3 0-0 1 1 li.:lbd2 (D)
1 t ...lOas - G0111e 18
1 1...Wh8 - Game 1 7
9••. lllf6 (D) 10 cxd5
10 l0c3 - Go1ne 1 J
1 0...i.xf3 - G0111e 14
10...l0xd5 - Come 15

B. . . i.g4 1 1 lllbd2 9. . .li:Jf6

46
 CHAPTER THREE
3ltJxe5: Black Plays 6 . . . J..d 6
1 e4 eS 2 c!Llf3 I.Llf& 3 c!LlxeS d6 4 I.Llf3
c!Llxe4 5 d4 dS 6 ..Q.d3 ..Q.d&
The line I e4 eS 2 lDf3 lDf6 3 lDxeS d6 4
lDf3 lDxe4 5 d4 ciS 6 i.d3 .i.d6 is sometimes
called the symmcttical variation. This is a
pretty formal name as the symmetry is ruined
lJuite quickly, and it's actually more accurate to
call it the Marshall Variation. At the lx.-ginnin�
of d1c 20th century Marshall often playt:d the
J..,'lUllbit 7 0-0 .i.g4?! 8 c4 0-0 9 cxdS f5, a line
that is nut quite correct, and eventually White
players learned how to dampen Black's dan­
�-,rcrous initiative (sec Game 1 9).
·n,e tabiya of the system arises after the
moves 7 0-0 0-0 8 c4 c6. Black has strcllbrth­
ened the <.IS-pawn. which lx.>Came vulnerable
;�ftcr ....i.d6. 9 'iVc2, attacking the kni1,rl1t on
e4, is covered in Games 1 9-2 1 . Defending the
knight with 9...f5, as played in the historic
J..,>amc Williams-Staunton, London 1851 is
certainly possible (see the notes to Grune 19),
but these days most Black players prcfer
Krau7.e's su&�,rcstcd pawn sacrifice 9 ...lDa6!?.
'I hen to .i.xe4 dxe4 11 'iVxe4 (Game 19) gives
Black active play as compcm;ation for the
pawn after either 1 1.. .lle8 or 11...lDb4!? (Raet­
sky's b'llmes are important here). Nowadays,
j.,>randmasters, including the elite, prefer to
prcvent the possibility of ...lDa6-b4 with 10 a3,
;�ftcr which Black continues with 1O...lle8,
I
IU ....i.g4(Game 20) or to... fS (Game 21).
Attacking the e4-knight with 9 lle I was
unpopular for many years, but it has become
quite fashionable rccendy. Defending the
knight with 9....i.f5 and the !,>ambit continua­
tion 9....i.�4 arc discussed in Game 22, while
9. .:es is the subject of Game 23.
.
Despite the obvious transposition after 9
lDc3 lDxc3 10 bxc3 .i.g4 1 1 cxdS cxdS and 9
cxdS cxdS I0 lDc3 lDxc3 11 bxc3 .i.g4, this
docs not mean that there arc no signitlcant
differences berween 9 lDc3 and 9 cxdS. After
9 lDcJ (Games 24-25) Black can proceed
with 9...lDxc3 10 bxc3 dxc4 1 1 .i.xc4 .i.J..,r4,
when the sharpest line is 12 'ii'd3 lDd7 13
lbgs lDf6 1 4 h3 i.hS 1 5 f4 h6 1 6 g4 as
played by Capablanca (Game 25).
The theory is very far advanced in the line
of 9 cxdS cxdS 10 lDc3 lDxc3 11 bxc3 .i.g4,
and here White manages to develop a strong
initiative on the kingsidc. After 12 llb1 the
game branches: 12 ...b6 is studied in Game
26, while 12 ...c!'Dd7 is the subject of Games
27-28.
Gamc 19
Burkov-Raetsky
Correspondence 1985
1 e4 eS 2 c!Llf3 I.Llf& 3 c!Llxe5 d6 4 c!Llf3
47
The Petroff Defence
l'Dxe4 5 d4 d5 6 .i.d3 .*.d6 7 0-0
White has a couple of minor alternatives:
a) 7 lbbd2 f5 8 �5 0-0 9 0-0 c5 to c4!?
(the passive 1 0 c3 allows Black to create
strong counterplay after 1 o...cxd4 1 1 cxd4
..i.xeS!? 1 2 dxeS llk6 1 3 lbf3 f4, for example
14 .txe4?! dxe4 1 5 ..xd8 l:lxd8 16 tbgs ..i.f5
17 ..i.xf4 h6 18 lbh3 - Shafranska-Raetsky,
Budapest 1 991 - and here 1 8...g5 gives Black
an edge) to...cxd4 1 1 tben (not 1 1 tbdf3?
1fc7 1 2 .if4 g5!, when Black is a1ready al­
most winning) 1 1...dxc4 1 2 lbxe4 fxe4 1 3
.1xe4 d 3 and the position i s complex.
b) Playing 7 c4 before castling allows
Black an additional option: 7 ....tb4+ 8 lbbd2
0-0 9 0-0 .txd2 1 0 tbxd2 (also possible is 1 0
.i.xd2 .1g4 1 1 .te3 li)c6 1 2 h3 .thS t 3 llc1
l:le8 14 a3 dxc4 15 .txc4 'irf6 16 .1e2 -
Short-Adam.-;, Wijk aan Zee 2000 - and here
Adams suggests 1 6 ...lbg3! 17 fxg3 l:lxc3 1 8
g4 .tg6 1 9 ..d2 lte4! leading to an unclear
position) to._li)xd2 1 1 .i.xd2 dxc4 12 .txc4
.i.e6 (1 2...Wxd4 t 3 .tb4 ..xdt 14 ltaxd 1
l:le8 15 l:lft:1 l:lxet+ 16 l:lxel .te6 17 .i.xe6
fxe6 18 l:lxe6 li)c6 is also equal) t3 .1xe6
fxe6 1 4 .g4 1Wd7 1 5 l:lfe 1 l:lf6 1 6 .tg5 l:lf5
17 d5! l:lxd5 1 8 llad1 tbc6 is level. Tiulin­
Raetsky, correspondence 1985.
7 0-0
•••
7....tg4?! is premature after 8 c4 0-0 9
cxd5 f5 10 h3 .th5 1 1 tbc3 tbd7 12 lbxe4
fxe4 1 3 .i.xe4 li)f6 1 4 .1f5 Wh8 and now:
a) 1 5 Wb3 tbxd5 Qf 1 5....txf3 1 6 ..x£3
48
li)xd5 17 'tth5 li)f6 18 1Wh4 White is much
better - Cordel) 1 6 .tgS .te7 17 .i.xe7 li)xe7
1 8 .i.e4 .i.xf3 19 .1xf3 li)f5 20 11fxb7 l:lb8
21 ..xa7 lbxd4 22 .lg4 l:lxb2 23 l:lad1 l:lb6
24 1Va4 and White has small advant!lh>c,
Bernstein-Marshall, San Sebastian 1 91 1 .
b) 1 5 g41 (this is even stronger) 1 5...lbxd5
1 6 .1e61 .tfl 1 7 lbg5 .txL-6 18 lbxe6 ..f6 1 9
li)xfB ltxf8 20 11fd3 and White is much bet­
ter, Gipslis-Christiansen, Gausdal 1 992
8 c4
This is the critical try, but 8 l:lc1 is also
reasonable. Now 8... .tf5 9 c4 c6 transposes
to 8 c4 c6 9 l:le 1 .if5, while 8...l:lc8 9 c4 c6
transposes to 8 c4 c6 9 :C l l:le8. This leaves
8....tg41? as the independent try: 9 .i.xc4
dxe4 10 l:lxc4 f5 1 1 l:le1 llk6 1 2 c.1 'iii'hH 1 3
lbbd2 11'£6 14 h3 .ths 1 5 ltlc4 l:lae8 1 6 :n
(Black also has compensation after 16 ltxe8
l:lxe8 1 7 ltlxd6 ..xd6 18 .td2 f4) 1 6... £4 1 7
lbxd6 cxd6 (White was a bit better after
17 .....xd6 1 8 b3 b5 1 9 .tb2 •ds 20 a4 a6 21
axb5 axb5 22 .ta3 l:l£6 23 .1c5 in Kulaots­
Rozentalis, Cappelle Ia Grande 2004) 18 b3
g5! 1 9 .1b2 l:lg8 20 g4 (not 20 d5?! .ixf3 21
'lrxO g4! 22 hxg4 ltle5 23 "irh3 ltxg4, when
White is in serious trouble) 20 ... fxg3 21 fxg3
l:le3 22 g4 ..£41 23 lbe 1 l:lxh31? 24 l:lx£4 gxf4
25 Wg2 l:lh4 26 �f2 l:lgxg4 27 ltlo l:lh3
with an unclear position (Kulaots).
8 c6 9 1tc2 tl)a&
..•
9 ... f5 is a major alternative here. Play
continues 1 0 lbc3

and now:
a) lO. ..lba6 1 1 -.,3 �h8 12 cxd5 cxd5 1 3
lbb5 (13 .i.xa6 can lead to a dmw after
1 3...bxa6 14 1i'xd5 l:lb8 1 5 �5 .i.b7!? 1 6
lbt7+ l:lxf7 1 7 1i'xf7 lbxc3 1 8 bxc3 .i.xh2+!
1 9 �2 Wh4+ 20 Wgt ..txg2! 21 Wxg2
1Wg4+) 13....i.b8 14 lle1 ..te6 15 a3 .i.gs 1 6
lbc3 ..Ll6 17 ..t ft g5!? 18 1Wxb7 g4 is un­
ck."llr, Volchok-Borisov, correspondence
1 984.
b) JO ...Wh8 1 1 cxd5 cxd5 12 ltlxd5
.ixh2+ 13 �xh2 1Wxd5 1 4 .i.c4 ..d6+ 1 5
tbcs lbc6 16 .i.f4 1Wf6 (not 16...lbxd4? 1 7
1Wd3 and White i s winning, but perhaps
Black shoukJ tty 16 ...ltlxe5 17 .i.xeS Wh6+
1 8 Wg1 .i.e6 1 9 dS i.g8 with unclear play)
17 ltlxc6 bxc6 1 8 i.eS Wh6+ 19 Wg1 ltld2
(19...lbg5?! allows White a strong advantage
after 20 f4 ltle4 21 i.d3, Macieja-Meijers,
Istanbul Olympiad 2000) 20 l:lfel (Macieja
pointed out 20 1Wc1 ?! ltlf3+1 21 gxf3 f4 22
ltc1 .i.h3, when White is in trouble)
20...ltlxc4 21 'ilxc4 and White has an edge.
1 0 J.xe4
1 0 a3 is studied in Game 20.
10 dxe4 1 1 •xe4 lilli4
..•
This is strong but Black has a good alter­
native in 1 t ...l:le8 12 1Wd3 (or 1 2 1Wc2 lbb4
1 3 -.o3 .i.f5 14 lba3 aS and Black has excel­
lent compensation) 12....i.g4 1 3 .i.gS 1Wd7 14
ltlbd2 h6 1 5 i.e3 (White should probably
prefer 15 .i.h4, but Black stiU has good play
after 1 5...lbc5 1 6 1Wc2 ..tf5 1 7 'l'c3 ltld3)
3 ltl xe5: Black Pla ys 6 . . . .i.d6
1 5... £5 1 6 ltlb3 .i.xf3 1 7 gxf3 1Wf7 18 f4 ltlb4
19 'ilc3 Whs with a dangerous attack,
Kruppa-Rozentalis, Lvov 1985.
1 2 /l\g5
White's two other possible options also al­
low Black good counte�play: 12 ltlc3 l:le8 1 3
'libt 'lfJ 1 4 tlk4 .i.g4, or 12 .lle1 l!fl 13 .ih6
.if5 14 ..e2 l:le8 1 5 1Wxe8+ 1fxe8 16 llxe8+
llxe8 1 7 lbc3 ltld3.
1 2 ... f5 1 3 ••2 f4 14 lD84
White wants to avoid the nasty pin that
arises after 14 ltlf3 ..tg4. For example, 15 a3
Lf3 16 gxf3 Wh4 17 'A'eM �h8 18 1Wg4
(other tries lose by force: 1 8 ..xd6 lbc2 19
lla2 'irh31 20 .ixf4 llad8, and 18 axb4 l:lf6
19 1Wd7 llg6+ 20 Wh1 :cis 21 1Wf5 llg5 22
1ic6 �1.51 23 lla3 llffJ 24 1We4 .g5)
18...1Wxg4+ 1 9 fxg4 ltlc2 20 lla2 lbxd4 and
Black was slightly better in Kuznetsov­
Raetsky, correspondence 1 985; or 1 5 ltla3
1if6 16 c5 .i.c7 17 1ic4+ ltlds and Black's
position is preferable.
14 f31 1 5 gxf3 J.h3 1 6 lle1 •c7
..•
1 7 a3
Black is also much better after other tries:
a) 17 lbxd6 1fxd6 18 �hl 1Wxd4.
b) 17 ltlg3 llae8 1 8 .ie3 (Black has as
strong attack after 1 8 1Wxe8 llxe8 19 llxe8+
..tf8 20 lba3 1Wf7 21 lle3 hS) 1 8...1Wt71 19
l0a3 lbd31 20 lleb1 lbf4 21 ...<12 �! and
White is in terrible trouble.
1 7 ... ..bh2+ 1 8 �1 lOa& 1 9 ll\g5 .if5
20 ll\c3
49
The Petroff Defenc e
20 tik6? leads t o a quick loss foUowing
20...W'f7 21 cJilxh2 D.aeS 22 d5 'WbS+ 23 cJilgt
11fg6+ 24 cJilh2 .i.xe6 25 dxe6 llf5.
20. . h6.
Black can also tty 20... .i.f4!? 21 tlk6 .i.xe6
22 ..xe6+ Wh8, when his attack continues.
2 1 lilge4
Playing to block the e-file. After 21 o!L!e6
1i'f7 22 o!L!xg7 Wxg7 23 Wxh2 l:lae8 Black
has exceUent play.
21 .Dae8
..
Also promising is 2t....i.g61? 22 �e3 llf5
23 Wg2 l:a£8, when Black has a powerful
attack. For example, 24 o!L!d2? l:lh51 (instead
24 ....i.h5 25 llh1 llxf3?? 26 1l'e1! was sud­
denly winning for White in Khramov­
Raetsky, correspondence 1985) 25 l:lht .i.f5
and White has no hope of defending.
22 .f1
White could try 22 .i.xh6!? but Black still
has compensation after 22..lle6 23 .i.e3
.i.f4.
22 ...1t'f7 23 1t'g2?
Now White is completely losing. After 23
Wxh2 W'hS+ 24 Wg1 .t.h3 25 1l'e2 Wg6+ 26
.i.g5 o!L!c71 27 ..e3 lbe6 at least White has
chances to fend off the attack.
23...i.b8 24 i.d2
Or 24 f4 :C6 25 dS cxd5 26 cxd5 llg6
and the attack is too strong.
24...•5+ 25 �g1 i.h3 26 lilg3 i.xg3
27 •xg3 :Xe1 + 28 :Xe1 :Xf3 29 •2
.g6+ 30 �1 i.g4 31 Jlg 1
50
31 ..g2 wouW have aUowed a neat finish:
3t...llh3+ 32 Wgt .i.f3!.
31 ...•f5 32 1t'e5 llh3+ 33 �g2 1Vf3+
34 Wf1 llh1 35 .g3 .th3+ 36 �.1
1t'xg3 0-1
Game 20
Kasparov-Shirov
Wijk aan Zee 2001
1 e4 e5 2 lilf3 �f6 3 li)xe5 d6 4 lilf3
li)xe4 5 d4 d5 6 i.d3 i.d6 7 0-0 0-0 8
c4 c6 9 1t'c2 lila6 10 a3
1 0 ....tg4
This and 10. .£5 (Game 21) are the usual
.
moves but 10..lle8 is also reasonable. Play
continues 11 o!L!c3 and now:
a) 1t....i.f5 is a solid line: 12l:le1 h6 13 cS
i.e? 14 .i.d2 .i.a5?1 Onstead Anand assesses
14...o!L!xd2 15 l:lxc8+ Wxe8 16 'it'xd2 .i.xd3
17 11'xd3 as cquaQ 15 .i.f4 .i.xc3 16 bxc3
�c7 17 h3 l£le6 18 .i.h2 (Anand-Kasimdz­
hanov, Hyderabad 2002) and here White is
slightly better after 18....i.h71 19 llab1 b6 20
lDes.
b) tt....i.g4 121£lxc4 dxc4 13 .i.xe4 .i.xf3
14 .i.xf3 'iVh4 15 g3 11'xd4 16 .i.g5 1i'e5 17
..d2 11'f5 18 �2 o!L!cs 19 l:ad1 (According
to Nataf, 19 'irxd6 is also equal after
19...1i'xg5 20 £4 ..f5 21 b4 l£lc6 22 1i'd7
l:ab8 23 l:ladt lled8 24 11'e7 :C8 25 1Vh4
lLif8) 19 . 1£le6 20 .i.e3 l:ad8 21 ..c3 .i.cS 22
. .
b4 llxd1 23 llxd1 .i.xe3 24 .xe3 aS! and

Black had equalised in Nataf-Vo17.hin, Stock­
holm 2001.
1 1 �5
11 cS .i.b8 12 l£les gives Black more
chances: 12...Le5 13 dxeS �exeS (but not
l3...lt)axc5?! 14f3 �xd3 15•xd3, winning a
piece) 14 .i.xh7+ 'iPh8 15 b4 1Wh4 16 .i.d3
�xd3 17 •xd3 Whs 18 lle1 .i.fS 19 Wg3
ftk7 20 �2lbe6 21 �b2 b6 and Black has
sttung counterplay, Jobava-Mamedyarov,
Plovdiv 2003.
1 1 .....th&
This is a major decision for Black as the
alternatives are very interesting:
a) 11...�f5 12 �e3 �acS 13 c.lxcS �xeS
14 cxdS 'ii'xdS 15 llk3 �xc3 16 bxc31lad8
17 .i.d4 �c3! lH �xfS (after 18 .xc3?
Motylev assesses 18...�xd3 19 �xg7 �xft
20 llxft llfe8 21 .i.h8 f6 22 'ti'xf6 lld7 as
clearly advantageous for Black) 18. ..Wxd4 19
lLct gp 20 1i'xc31i'xc321 llxc3gxfS is very
llrawish, Smimov-Motylev, Russia 2004.
b) 1t....i.xe5 leads to heavily analysed
complications after 12 dxeSlbacS
and now:
b1) Black has an easy position after 13
cxdS ?! 'ii'xdS Qess exact is 13... cxd5 14 f3
llc8 15�e3- not 15 fxe4? dxe4!with a clear
advantage - 15... d4 16 �xe4 .i.e6!? 17
.i.xh7+ ¢'h8 with an unclear position) 14
.i.xe4 �xe4 15 llk3 (if 15 f3 �cS 16 lldl
1i'b3! Black has tremendous play) 15...lthc3
16 Wxc3llfdK.
3 f0xe5: Bisek Plays 6 ... �d6
b2) 13 f3 �xd3 14 1i'xc.I3 �cS ! 15 1i'd4
�b3 16 Wxg4 �xal 17 .i.h6 g6 (if Black
inserts the moves 17 ...'1Fb6+ 18 Whl before
playing 18...g6 then 19 Wf4f6 20 .i.xf8 llxf8
21 cxdS cxdS 22 Wd2 promises White a
strong initiative) 18 llk3 (an important
choice) 18-.'1Fb6+ t91lf21lfe8 20 Wf41i'c7
(in Shirov-Leko, Linares 2000 White was a
bit better after 20...f5 21 cxdS lDb3 22 e6
cxdS 23 �xdS 1i'xe6 24 �c7 1i'c6 25 lDxe8
llxe8 26 g4Q 21 lle2 lle6, which Shirov as­
sesses as unclear. Instead of 18 ru, 18 cxdS
is very messy after 18...:C8 19 �2 Wb6+ 20
Wh1 llxeSI 21 llxa1 1i'xb2 22 lld1 .ll xdS.
Perhaps the best move is 18 �xf8 as White
has an edge after 18...9xf8 19 cxdS 'ii' cS+ 20
Wh l cxdS 21 e6 f5 22 1i'a4.
1 2 cxd& cxd& 13 �c3
A good, simple developing move. Taking
the pawn with 13.i.xc4dxc414Wxe4aUows
Black sufficient countcrplay: 14..lte8 15.i.e3
(or 15 �f4�xeS 16 dxeS llks 171We3l0d3
and Black has excellent compensation)
15...lDcs 16 1i'c2 t:&6 17 'ii'rs Wh4!? 18
�c.l2 �g6 19 1i'g4 1i'xg4 20 �xg4f5 21 �c4
�f8 22 dS fxbr4 23 dxe6 .llxe6 with an I.:<Jual
position, Potkin-Rozentalis, Bad Wiessee
2003.
1 3 ...�xc3
This may seem natural, but Black should
seriously consider 13...Le5. For example,
14dxcS lDacS 15�e4lDxd3 16 'iixd3�gp
17 'ilb3.i.xe4 18 Wxb71Wh4 19 1ib3 d4 20
51
Th e Petroff Defence
1i'g3 1i'xg3 21 hxg3 l:lfc8 and Black had suf­
ficient compensation in Leko-Shirov, Linares
2004.
14 bxc3 �h8
Black is clearly worse after 1 4 ... ..ixe5?! 1 5
..ixh7+ �h8 1 6 dxc5 g6 1 7 'itd2! �h7 1 8
'ii'h6+ 'iPg8 1 9 ..igS f6 20 ..ixf6 l:lxf6 21 exf6
1i'xf6 22 £31 (Dokhoian).
1 5 f4
A natural move but White has a more
threatening choice in 1 5 .ixh7!?. After
1 5...g6 the bishop is trapped but 1 6 1i'd2
�h7 1 7 9'h6+ �g8 18 ..tg5 ..ie7 1 9 ..txe7
1i'xe7 20 llae1 1i'c7 21 l:le3 creates a dan­
gerous attack.
1 5 ....txe571
Now White's rooks arc too strong. In­
stead 15 ...f6!? 16 li:l£3 1i'd7 17 llb1 l:lab8 is
unclear.
1 6 fxe5
White would even have a slight advantage
after 16 dxe5 lLic5 1 7 .i.c3 li:lxd3 1 8 .xd3
b6 1 9 h3.
1 6 .tg6 1 7 a4 ...d7
•.•
Or 17 .....txd3 1 8 Wxd3 f6 19 .i.a3 l:lf7 20
l:lf5 1i'd7 21 l:laft Wxa4 22 .i.d6 and the
attack continues.
1 8 .ia3 llfe8 1 9 .ixg6 fxg6 20 11fb3
20 b6
•..
Against 20...ll:lc7, Dokhoian supplies the
neat winning line 21 1i'xb7 :CbS 22 e6! 1i'd8
23 1i'xa8! l:lxa8 24 e7 Wg8 25 l:r.f8.
21 .id6 lfx;7 22 llf3 llac8 23 llaf1 h6
52
24 .c2
24 llf7 is also winning for White: 24...1i'c6
25 "ifc2ll:le6 26 Wxg6 11f'xc3 27 .i.e7! 'lfxd4+
28 Wh1 1i'xe5 29 ..if6 1i'c4 30 ..ixg7+ �g8
31 1i'g3 h5 32 .i.eS+ 1i'g4 33 1i'f2 (Dok­
hoian).
24•..•g4
There is no way to save the pawn on g6,
for example 24....z:r.e6 25 l:lf7 1i'c6 26 .: 1 f6!.
25 Ag3 'trh5
The rook ending after 25...•e4 26 1i'xe4
dxe4 27 .i.xc7 ltxc7 28 :Ct is easily winning
for White.
26 JDI3 ...g5 27 llg3 'trh5 28 .ixc7
llxc7 29 llxg6 1th4 30 h3 •xd4+ 31
cxd4 llxc2 32 llf7 J:lg8 33 J:ld6 J:lc4 34
J:lxd5 J:lxa4 35 J:ldd7 lla1 + 36 �2 1la2+
37 Wf3 Wh7 38 e6 �g6 39 d5 J:lc8 40
llc7 J:le8 41 g4 a5 42 llxg7+ Wf6 43
J:lgf7+ �5 44 1lf5+ �d4 45 e7 1 -0
Game 21
Grischuk-Gelfand
Wijk aan Zee 2002
1 e4 e5 2 lL!f3 lL!f6 3 li)xe5 d6 4 lL!f3
lL!xe4 5 d4 d5 6 .id3 .id6 7 0-0 0-0 8
c4 c6 9 •c2lLla6 10 a3 f5 1 1 &3
White has a sound alternative in 11 c5
.i.b8!? (1 t....i.e7 12 ll:lc3 lLic7 transposes to
11 lLic3 ll:lc7 12 c5 ..ie7) 12 lLic3 lLic7 1 3
ll:le2 (White should try 1 3 ..tf4!? ll:le6 1 4
.i.xb8 l:lxb8 1 5 lLic2 1i'f6 16 b4 g 5 1 7 'iVb2
with a slight edge) 1 3...lLic6 1 4 b4 1i'f6 1 5
..tb2 ..td7 16 llae1 ..tes 1 7 'lfcl .i.h5 1 8
ll:le5 ..ixe2 (also promising is 1 8...li:lxd4!? 19
.i.xd4 ..txe2 20 li:l£3 ..ixd3 21 .i.xf6 l:r.xf6
with excellent compensation - Piket) 19
.i.xe2 o!Dxd4 20 l2Jct7 o!Dxe2+ (not 20...9'h4?!
21 o!DxbB o!Dxe2+ 22 .llxe2 Lxb8 23 £3 .!Df6
24 l:lfe1, when White had plenty of compen­
sation, Moro7.evich-Piket, Wijk aan Zee
2002) 21 .D.xe2 'lff7 22 o!Dxb8 l:r.axb8 23 £3
.!Df6 24 1i'g5 ltbe8 and Black is slightly better
(Morozcvich).
1 1 &7
.•.

12 .lle 1
White has plenty of alternatives:
a) 12 b4 a5 13 b5 .i.d7 14 llb1 (or 14 b6
c!frc, 15 cxd5 cxd5 16 �d5 llc8 17 'W'a2
lt.?h8 with compensation - Svidler) 14...�h8
15 c5 .i.e? 16 bxc6 bxc6 17 .if4 lbe6 18
.ic5.i.e8 19 llb6 'W'c8 20 �2.i.g5! (White
has excellent compensation after 20.....td8 21
11 .Lb6 22 cxb6) 21 �b3.i.hS 22 a4 'ireS is
unclear, Svidler-Morozevich, Krasnoyarsk
2003.
b) 12llk2 � 13 b4 .id7 14 cxd5 cxd5
IS �3 .ie8 16 llk5 .i.e? (if 16...�xd4,
A.Sokolov claims an edge for White after 17
..xdS+ .if7 18 'W'xd4 .i.xe5 19 'ii'xe5 'W'xd3
20 .i.b2 �f6 21 �f4) 17 .i.e3 .ih5!?
(17 ....if7 is inferior. 18 f3 .i.xe5 19 dxe5 d4
20 fxc4 dxe3 21 .ic4 �4 22 'W'a2! �xe2+
2.1 'W'xe2 'W'e7 24 :Xf5 Lc4 25 'ii'xc4+ �h8
26 'W'c3 and White was clearly better in
A.Sokolov-Frcssinet, Val d'lscre 2002) 18 f3
lL\4g5 19 llfd 1 f4 20.i.£2 and White has only
a small advantage.
c) 12 c5 .ie7 J3lbc2 with a branch:
cl) 13....i.f61? 14 b4 &1> 15 .ib2 'ii'c7 16
l:lad1 rffi 17 'ii'b1 'W'g7 18 .i.at .id7 was
unclear in Tseshkovsky-Motylev, Ekaterin­
hurg2002.
c2) 13...�c6 is a viable alternative: 14ltlf4
.if6 15 �xe6 .i.xe6 16 .if4 'W'e8 Oess
promising is 16....i.f7 17 .icS .ih5 18 .ixf6
1i'xf6 19 lDc5 llad8 20 f3l0g5 21 .llfellbc6
22 '1Vf2 f4 23 b4 .llfe8 24 11fh2 .lle7 25 l:le2
3 liJ.xe5: Blac k Plays 6 . .. �d6
llde8 26 llac1 when White had a definite
edge in Gipslis-Raetsky, Berlin 1993) 17 .i.e5
1fh5 18 life I llae8 and Black has managed
to complicate the game.
1 2 ...�8 1 3 b4 .td7 14 .tb2liJ.e6
This is an ambitious move. Black could
equalise easily with 14...dxc4 15 .ixc4 �xc3
16.i.xc3�5
1 5 cxd5
This is probably best White is in trouble
after 15ltle2 �6g5 16 �5.i.xcS 17 dxc5 f4
18 ltld4 (or 18 f3 �x£3+1 19 brxf3 1ig5+ 20
ltlg.'i fxg3 and Black's attack crashes through)
18...£3, when Black has a dangerous initiative
(Gelfand).
1 5 ...cxd5 16liJ.xd5 llc8 1 7 •d1 liJ.6g61
·n1is keeps up the initiative. After 17....i.c6
18 �3 �3!? 19 .i.xc3 L£3 20 1txf3 :Xc3
21 ltlxf51 �xd4 22 11fg4 �f5 23 .ix£5
White escapes with an edge.
1 8 liJ.e5
After t8lbc3ltlxf3+ 19 'W'xf31th4120 g3
1th6 the attack continues (Gdf.1nd).
1 8 ...lLlh3+ 1 9 gxh3 •us+ 20 �1
Tlus is forced. 20 l0g4? loses immediately
to 20...ltlxf2!.
20 ...�xe5
In such a position it is not surprising that
Black has more than one possibility:
20...'ilh4!? 21 lle2 (after 21 'it'c2 .ixe5 22
dxe5 .ic6 23 .ic4 1ixh3+ 24 �g1 llKl2! 25
1txd2 'it'g4+ 26 �hl 1i'xc4 Black has a pleas­
ant advantage) 2t...1fxh3+ 22 �g1 .ixe5 23
53
Th e Petroff Defence
�xeS 23 dxe5 Wg4+ 24 �f1 11t'h3+ when
White should accept a draw (Gelfand).
21 dxe5 .*.c6
This is promising, but so is 21 ...�e6!? 22
c!De3 Wf4! 23 :e2 1fxh2 24 Wet Whl+ 25
c!Df1 c!DgS 26 .llc2 :cd8 27 f4 ..xh3 (if Black
wishes he can force a draw with 27 ...c!Df3+ 28
W£2 �d5 29 �g3 lDel! 30 Wxel 1ff3+ 31
Wh2 1lt'h 1 +) 28 fxg5 :xd3 29 1fe2 1lfd8
with excellent compensation (Gelfand).
22 �c1 1
A superb defensive move. After 22 c!De3
:Cd8 23 1fc2 :Xd3! 24 9xd3 1ff4 Black has
a terrific attack. Por example, 25 .!&II l.lkl2+
26 Wg1 c!Df3+ 27 Wf1 1fxh2 28 We2 1fxh3
and Black wins (Gelfand).
22 ...11114 23 �xe4 fxe4
24 �e3
White continues to defend stoutly. 24
c!Df4? loses rapidly to 24.. Jixf4 25 �xf4
54
1i'xh3+ 26 �gl e3, while 24 ltaz?! e3! 25
:xe3 1fc4+ is not much better.
24...Wxh3+
Now the game must end in a draw. Black
could play on with 24...:0!? but after 25 b51
�xbS + 26 �gl .llxh3 27 �f4 :d3 28 �g3
Wd8 29 1fg4 l:txd5 30 1fxe4 the position is '
simply unclear.
25 �1 hd5 26 'irxd5 'irg4+ 27 Wf1
1ih3+ 28 Wg1 1tg4+ 29 wn % - %
Game 22
Shirov-Piket
Wijk aan Zee 2001
1 e4 e5 2 l0f3 l0f6 3 l0xe5 d6 4 l0f3
l0xe4 5 d4 d5 6 .*.d3 .*.d6 7 0-0 0-0 8
c4 c6 9 1le1
9 -*.fS
••.
9. .:es is examined in Game 23. Black's
.
two other options make life easier for White:
a) 9... f5 1 0 lDc3 Wh8 1 1 1Vb3 lDa6 1 2
cx.d5 cxd5 13 c!Db5 �b8 1 4 �d2 .t.c6 1 5
llact :f6 (or 15...c!Dxd2 t 6 c!Dxd2 :e8 1 7
c!Df1 �f4 1 8 lDe3 and White has a pull) 1 6
�g5 lDx.gS 17 c!Dxg5 .tgS 18 c!Dt3 .tf4 1 9
:c2 1lb6 20 1ia' 3 and White was a bit better
in Naiditsch-Meijers, Senden 1999.
b) 9....ig4 and now:
b1) tO c5? leads only to trouble:
1 0....ixh2+! 1 1 Wft (or 1 1 Wxh2 c!Dxf2 and
Black is better because 1 2 ..e2? loses to
12...lfud3 1 3 ..xd3 .ixt3) 1 1 ...5 12 lDc3

.!l:Xt7 13 �xe4 fxe4 14 .i.xe4 J.c7 15 .i.g5
lilf6 16 .tc2 11'd7 17 1l'd3 �4 and Black
was obviously better in Yemelin-Ractsky,
Rostov on Don 1993.
b2) The simple 10 .i.xe4 dxt.-4 11 Le4 is
best. Following 11... f5 White has two op­
tions:
b21) 12 l:e6 gives Black chances after
12...11'd7 13 :Ct lL!a6 14 11'b3 (or 14 lilc3
l:ae8 15 l:xe8 l:xe8 16 .i.d2 J.b8 with b>t>Od
attacking possibilities) 14...J.xf.3 15 11fxf.3
ltae8 16 J.e3? {White can hang on to equal­
ity after 16 J.d2 J.b4) 16...�b4 17 c5? (after
17 lLla.3 lDd3 18 :n J.xa3 19 bxa3 f4 20
'ii'd1 •f5 21 J.d211'e4 Black is much better,
but there is no forced win) 17.. .ru 18 cxd6
f4 19 J.d2lilxe1 20 11'b3+ *h8 21 � f.3
0-1 Kosteniuk-Raetsky, Bicl 2004.
b22) 12 l:e 1 (again the simple move is the
best) 12...J.xf.3 13 11'xf.3 11'h4 14 h3 ..xcl4
1511'b31ib6 16 11'xb6 axb6 17lL!c3lbd7 18
ltdt l:f6 and White had a tiny edge in
Ehlvest-Mamedyarov, Moscow 2002
1 0 1rc2
This is probably the best of a wide choice:
a) 10 cS .tc7 11 � l:eR 12 1l'c2lbd7
13lL!xe4 dxe4 14 .i.xe4 .e7 15 �5lL!f6 16
f3 (Socko-Skato;hkov, Cappelle Ia Grande
2004) and hert: Skatshkuv suggests that
16 ...lilxe4 17 fxe4 h6 18 ...f2 .tg6 is slightly
bcrtcr for Black.
b) t O eb3 �6 t1 cxd5 cxd5 12 �3
and now:
3 li)xe5: Bisek Plsys 6... i.d6
b1) 12...J.c6 l3 a3 lilc7 14 ...c2 f5 15
liJc5 ltc8 16 .if4 g51 17 .td2 (also worth
considering is 17 J.e3 lila6 18 f.3 lL!xc3 19
bxc3 .txe5 20 dxe5 �5 21 .td4 with coun­
terplay - Gelfand) 17...�6 18 1i'd1 .i.xeS 19
dxeS �acS 20 .ic2 �xd2 21 9xd2 d4 22
�4lilb3 23 .txb3 .txb3 24 c6! was unclear
in Frcssinet-Gclfand, Cannes 2002
b2) Also interesting is 12...�b4 13 .i.b11?
(Black has enough compensation after 13
.i.xe4 dxe4 14 .ig5 1ih6 15 �xe4 .i.c6 16
'ild l �LacS) 13....ie6 14 a3 �6 15 1Wxb7
�5 16 ...a6lL!xc31 (not 16...lilb3?117lilxe4
dxe4 18 .ixe4 �xa1 19 .i.xa8 �c2 20 l:te2
�xd4 21 �d4 J.cS 22 .ie3 .txd4 23 lld2
and White was comfortably better in Palac­
M.Ivanov, Cannes 2004) 17 bxc3 �b3 18
.ig5! J.xh2+ 19 '1Ph1 ...c7 20ltxe6!? fxd) 21
11fxe6+11'f71 22 9xf7+ :Xf7 23 .i.a2!?lilxa1
24 .ixdS and after all the complications the
position remains unclear.
c) 10 �3 �c3 11 bxc3 .txd3 12 1l'xd3
dxc4 13 'lfxc4lLld7 with another branch:
ct) 14 .igSI? 11'c7 15 .te7!? (15 11'b3 is
less enterprising: 15.. 1lfe8 16 .lab1 h6 17
.i.e3 b6 with a dull equality) 15...�b6 16
.i.xd6 9xd6 17 'iib3 l:tacS 18 c4 'flc7 19 a4
lLld7 20 l:tab1 b6 21 g3 1l'd8 22 aS 11'f6 23
axb6 axb6 24 *g2 and White had a marginal
edge in lvanchuk-Gelfand, Lvov 2000.
c2) 14 1tb3 has also bet."' tried: 14...1i'c7
15 c4 :rea 16 .ib2 h6 17 g3l:tad8 18 �4
.i£8 19 11'c2 (perhaps a bertcr try is t9l:tad1
1Wb6 20 .ic3 with unclear play) 19...1fa5! 20
lL!g2ltxe1+ 21 l:txe1 bS 22 .ic3 b4 23 .i.a1
�b6 24 91>3 cS 25 dS Wa6 26 �e3 lte8 27
l:tct 11fc8 and Black was a bit better in Ad­
ams-Morozevich, Dortmund 2002.
1 0 ...i.g6
Solidly defending the bishop. 1o.. .lLxl7 is
also playable: 11 �3 lLldf6 12 cS (12 cxdS
cxdS 13 �3 :Cs 14 "xb7 ltb8 15 'W'xa7
l:e7 16 1i'a6 lte6 17 1l'a7 l:e7 is simply a
draw- Piket) 12....ic7 13 b4 J.g6 14 .i.b2
�c3 15 11fxc3 .ixd3 16 1i'xd3 �e4 17 l:te2
l:e8 18 l:taet l:e6 19 �cS 11'h4 20 J.c1 (20
55
The Petroff D e fence
f3? loses after 20..llh6 21 fxe4 'lfxh2+ 22
wn Wht+ 23 Wf2 llf6+ 24 tnn �g3+! 25
Wxg3 �>6+ - 1-Jaba) 20...f6 21 ti)f3 'iVh5 22
g3 llaeS 23 Wg2 g5 was unclear in Lanka­
Haba, Hamburg 2003.
1 1 c5 J..c7 1 2 l0c3
1 2 l0f6
•••
Alternatives are not so good:
a) 12...tnxc3 13 bxc3 tnd7 14 �xg6 hxhr6
15 �g5 lDf6 16 &5 and White has a pleas­
ant initiative - Pikct.
b) 12...f5 13 'lfb3!? and now 13...:tt7 al­
lows a trick: 14 ti)g5! b6 (the point is
14...tnxg5? 15 �xg5 Wxg5 16 :teS+ J:tf8 17
l:lxfB+ �xf8 18 'ifxb7 and White wins) 15
�xe4 fxe4 16 lDxt7 �xt7 17 cxb6 axb6 18
lDe2 .f6 19 �e3 tnd7 20 llac I with a clear
advanmge, Anand-Morozevich, Monte Carlo
(blindfold) 2003. Instead, 13... b6 14 cxb6
axb6 15 tnxe4 fxe4 16 �xe4 �xc4 17 :Xe4
gives White a small advantage.
1 3 J..g5 lbbd7 1 4 r!Lle5 .bd3 15 1txd3
1tc8 1 6 f4
If 16 b4 lDxe5 17 dxe5 tnd7 18 f4 f6 19
exf6 tnxf6 20 g3 �dB Black's stands well.
1 6 .. .i.a5! 1 7 1tg3 �8 18 e.4 .i.xc3
Preparing ...tne4
. The immediate
18...lDe4!? is also possible: 19 lle3 tnxe5 20
fxe5 ..i.xc3 21 bxc3 f6 22 exf6 gxf6 23 �h6
llt7 24 c4 1Ve6 25 cxd5 cxd5 26 :lfl l:g8 27
h3 with an unbalanced position - Piker.
1 9 bxc3 �e4 20 l:le3
White couJd try 20 �c7 tnxc5 21 fxc5
56
lies 22 �d6 tnxc3 23 llc3 lDe4 24 :th3 h6
25 :tft with reasonable play for the pawn.
20. . .�xg5 21 •xg5 h6
Trying to win a piece with 2t...f6? loses to
22lDg6+! hxg6 23 1Wxg6lDxc5 24 :te7.
22 -.,5 �f6 23 �xf7+ �h7 24 �g5+
Wg8 25 •s6 hxg5 26 l:le7!
Black defends easily after 26 fxg5? lbe4 27
:th3.f5.
26 ...�8 27 l:lae1 1tg4
Black would be mated after 27.-gxf4? 28
llxg7+! tnxg7 29 l:te7.
28 fxg51th4
The safest line is 28. .1Wf4! 29 h3 'ifg3 30
.
�h1 'ifxc3 31 :XeS Lc8 321lxc8 'ifct+
33 Wh2 1Wf4+ with a draw (Piket).
29 g3 .3 30 llxb7 .,5 31 ••6+ 1
11rxe6 32 l:lxe6 l:ln?!
32...J:tf31? 33 llxc6 J:txc3 is still unclear.
Now Black is in trouble.
33 llxn �xn 34 llxc6 l:lb8 35 l:la6 l:lb7
36 h4 fi:jc7 37 l:ld6 �5 38 g6+ h7 39
llxd5 �c3 40l:le5+ � 41 c61 l:lb1 +
There is no way to save the game: 41.. llc7
42 h5 l:lxc6 43 g4 llc6 44 gS+! �c7 45 h6
gxh6 (45...lbe2+ 46 :txe2 :txe2 47 hxg7 is a
beautiful finish) 46 g7 l:lxc5 47 dxe5 �t7 48
e6+ Wxg7 49 !,rxh6+ and a pawn will queen
(Piket).
42 �g2 l:lb2+
42...tnb5 doesn't help: 43 J:ld5 :tb4 44 a3
:tb2+ 45 �h3 <iii'e6 46 lld7 and White wins.
43 Wf3 1-0

Game 23
Kovalev-Rozentalis
Glogow 2001
1 e4 e6 2 l'Of3 ll)f6 3 l'Oxe5 d6 4 l'Of3
ll)xe4 5 d4 d5 6 �d3 �d6 7 0-0 0-0 8
c4 c6 9 .lie 1 .lieS
10 ll:lc3
This is the most logical of a range of pos­
sibilities:
a) 10 ..c2 ..i.g4 11 lLic5 ..i.xeS 12 dxe5
lfr.t6 13 ..i.xe4 dxe4 14 'ilxe4 f5 15 'ilc2lLib4
16 ..c3 lLid3 and Black's huge lead in dcvcl­
' 1pmcnt provides ample compensation for
the pawn.
b) 10 lL!bd2 ..i.t'!i 11 c5 (Black equalises
easily after t 1 'ifc2 lLixd2 t2 ..i.xd2 i..xd3 13
1rxd31lxc1+ 14 l:lxe1 dxc4 15 ..xc4lLia6 16
.!DgS ..d7) 11...i..f8 12 Wb3 b61 13ll\xe4 (or
13 cxb6 axb6 14ll\fl ..i.d6 with an unclear
position) 13 .....i.xc4 14 ..i.xe4 dxe4 15 ll:x:S
16'd5! 16 cxb6 axb6 and Black has equalised.
·n1e point is that 17 Wxb6?! is a mistake after
17... £6 18 Wb7 (Ulibin-Raetsky, Makhachkala
1993). Black should continue t8...lLia61 19
�d7 ll\b4t 20ll\xf8 (or 20ll\b6 Wbs 21 a4
'ii'a6 and Black is clearly better) 20...ll\c21 21
iLh61? ..f7 (of colU'SC not 21 ...gxh6?? 22
lZ'Ixh7 and White wins) 22 lik!7 ll\xct 23
Axel llcd8 with a smaU advantage.
c) I 0 CXliS should not be dangerous:
IU ..cxdS II �3 ll\xc3 121lxe8+ Wxe8 1 3
. 1 3 ...1i'd7 is another option: 14 c5! ..i.c7 15
3 l'Dxe5: Bla ck Plays 6 ... �d6
bxc3 ..i.g4 14 ..i.e3 .dB?! (Black should play
to exchange White's better bishop; after
14.....d7 15 h3 ..i.fS 16 i..xfS WxfS the game
is leveQ IS llb1 ll\d7 16 h3 ..i.x£3 {after
16 .....i.h5 White can simply win a pawn with
17 llxb7 ll\b6 18 g4 i..g6 19 i..xg6 hxg6 20
ll\g5 ..i.c7 21 'iWD) 17 ..xf3 (White's bishop
pair ensure an edge) 17...ll\b6 1 8 a41lb8 19
a5ll\a4 20 i..d 2 ..xa5 21 l:lb5 •c7 221lxd5
.1Zd8 23 llb5 ll\b6 24 h4 a6 25 l:lg5 lieS?
{White had a promising attack, but this al­
lows a quick kiU) 26 .J:lxg7+1 1-0 Timofeev·
Raetsky, Abu Dhabi 2004 {26.. .'�xg7 27
..i.h6+! leads to mate).
1 0 l'Dxc3 1 1 bxc3
...
Also playable is 1 1 l:lxe8+ ifxe8 12 bxc3
..i.g4 t 3 ..i.d2 (with the idea of Wb t and a
double attack) 1 3...h6 (t3....d8?! mio;ses the
threat: 14 cS ..i.c7 15 Wb1 i..xf3 16 Wxb7
iDd7 17 gxf3 .J:lc8 18 llet g6 19 h4 and
White was much better in Raetsky-Varlamov,
correspondence 1985) 14 Wb3 dxc4 15
i..xc4 (Black can draw after 15 .xb7 cxd3
16 IZet 'ilfd7 17 'ilxa8 ..Lf3 18 gxf3 1Wh3 as
there is no good way to avoid perpetual
check) 15...b5 16 l:le1 1Vf8 17 i..x£7+1 ..x£7
181lc8+ ..i.f8 19ll\e5 Wxb3 20 axb3 ..i.hS 21
.J:ld8 and White has sufficient compensation
for me piece: the pins are nlmost impossible
to escape.
11 �g4 12 .i.g5 Jlxe1 + 13 1rxe1
.••
1 3 1rc8
•••
57
The Petroff D efence
�h4!? h6 16 .i.d2 'lidS? (now Black is in
desperate trouble; instead t6...liJa6 17 'lfbt
Zle8 18 g3 Wc8 19 f3 .i.h3 20 IllES is only
slightly better for White) 17 f4l .i.c8 18 Wg3
(surprisingly White is already winning by
force) 18...b6 (after 18...1i'f8 Anand analyses
a beautiful winning line: 19 %let .i.d8 20
lt\g6! fxg6 21 1i'xg6 .i.e? 22 �7+ Wf7 23
.i.g6+ Wf6 24 .i.hs .i.fS 25 g4l .i.xh7 26 g5+
hxg5 27 fxg5+ WfS 28 Zlc:S mate) 19 Zlc:t
bxc5 20 dxc5 'iff8 21 .i.c:31lla6 22 .i.d4 g5
23 ...f2 and Black is defenceless and resigned
in Anand-Piket, Wijk aan Zce 2001.
1 4 lDh4
This prepares a kingside advance. Whi.te
has tried two other possibilities:
a) 14 .i.e7 .i.f4! 15 .i.h4 �7 16 .i.g3
.A.xg3 17 hxg3 .i.xf3 18 gxf31£lf6 (Black has
equalised) 19 WeS dxc4 20 .i.xc4 Wd7 21
Wf4 ZleB 22 11fd2 hS 23 %let :Xet+ 24 'lfxel
bS Vz-112 Movsesian-Rozentalis, Neum 2000.
b) Perhaps White's strongest line is 14 cS,
for example 1 4....i.f8 15 'ltb 1! h6 (after
15 ..g6 16 �eS .i.e6 17 f4 �7 18 �f3
.
White has a promising attack - Rw.entalis)
16 �eS .i.e6 17 .i.f41£ld7 18 a41? .i.e? 19 aS
.i.d8 20 'i'b4 and White's easier devclop­
ment promises an edge, Degraeve-Rozentalis,
Montreal 2002.
1 4...lt:la6
If 14... h6 then 15 J:.e7 .i.e? 16 .i.a3 and
White still has pressure.
1 5 c5 .i.fB 1 6 h3 .i.e& 1 7 f4!
58
1 7 h6?!
•.•
This is too weakening. Black should play
17...f6, for example 18 fS fxgS 19 fxc6 �xeS
20 dxcS .i.xc5+ 21 Whl gxh4 22 1i'xh4 g6 23
%let and White obviously has compensation
for the pawn but Black can hope to defend.
1 8 f5lt:lxc5
Black prefers to sacrifice a piece rather
than take one: 18...hxg5 19 fxe6 Wxc6 (the
problem is that 19...gxh4? loses to 20 cxf7+
'ii?xf7 21 Wxh4, when the king is doomed) 20
'lfxe6 fxe6 21 �f3 .i.e7 22 Zlb1 and White
has a clear advantage - Rozentalis.
1 9 dxc5 hxg5 20 fxe6 .i.xc5+ 21 �1
1be6
If Black tries to recapture the piece Whi.te
gains a d<.-adly attack: 21...gxh4 22 Wxh4 g6
23 exf7+ Wg7 24 %let 'ifd7 25 J:.xg6! Wxg6
26lle5.
22 11xe6 fxe6 23 li:lf3 g4 24 hxg4 �7
25lt:le5+?!
Now Black's central pawns have a chance
to roll. White can keep control with 25 g3
ri;e? (25. id6 26 Zlfl Wc7 27 �5 is also!
.
unpleasant) 26 Wg2 'ii?d6 27 Zlcl with good
winning chances.
25 .. .'�87 26 g3 �d6 27 lt:lf7+
White should probably give up a pawn to
blockade Black's centre: 27 Zlc:t .A.f2 28
�f7+ We7 29 Zlfl .i.xg3 30 'ii?g 2 i.e? 31
ll!gS and Black's pawns are not a threat, so.
White is a bit better.
27 ...h7 28lt:lg5 e5

Now Black has significant counterplay.
29 :n e4 30 J:lf7+ Wei& 31 J.e2
After 31 .i.xe4?! .i.e3! 32 llxg7 .i.xgS
Black escapes with a draw.
31 b5 32 :Xg7 J.e3 33 J:lg6+ Wd7 34
..•
liJf7 J:le8 35 g5 J.d2 36 J:lg7 J.xc3 37
.:117 d4 38 .i.g4+ �7 39 g& d3
40g77
In time trouble White slips up. Instead 40
'ifi!g2 cS 41 li:)d8+! Wb6 42 llb7+ Wa6 43
:c17 gives good chances.
40...J:lg8 41 J.f5 d2 42 J.g4 e37!
Now it is only a draw. 42...1lxg7? 43 :Xg7
.i.xg7 44li:)gS, leaving White clearly better, is
even worse, but Black could have kept de­
cent winning chances after 42....i.f6! 43 ¢'g2
cS.
43 lt)h& lbg7 44 :Xg7+ J.xg7 45 lOts
.i.e5 46 �2 c5 47 lrule3 c4 48 J.e2
'>Pd& 49 g4 .i.f4 50 Wt3 WaS
Of course not 50....i.xc3? 51 Wxe3 c3 52
'1Pd3 b4 53 .i.dt when White wins as all
Black's pawns arc worthless.
51 �c2 J.g5 52 J.d1 % - %
Game 24
Firman-Bick
Stratton Mountain 2003
1 e4 e5 2 l0f3 �f& 3 lL!xe5 d& 4 liJf3
lllxe4 5 d4 d5 6 J.d3 J.d& 7 0-0 0-0 8
c4 c& 9 liJc3l0xc3 10 bxc3 dxc4
Black can also try the immediate 10....i.g4
3 fi)xe5: Black Plays 6 ... .i.d6
and now:
a) 1 1 c5 (stabilising the space advantage)
1 t....i.c7 1 2 :C1 �7 1 3 .i.g5 li:)f6 (or
13 ... f6 1 4 .*.h4 llfl 1 5 1fc2 g5 1 6 .i.g3 .i.xg3
1 7 hxg3 - Stellwagen-S.Ernst, Groningen
2002 - and here White can gain an edge with
17 ...li:)f8 1 8li:)h2 .i.hS 1 9li:)f1) 1 4 111d2!? (14
h3 leads to an unclear position after 1 4...i.h5
15 g4 i.g6 1 6 �5 h6 17 .i.h4 .i.xd3 1 8
'iVxd3 g5 1 9 .i.g3 �4) 14.....ixf3 1 5 gxf3 b6
1 6 lle2 bxc5 17 dxc5 h6 1 8 .i.xh6! �7 (not
1 8...gxh6?? 1 9 1Wxh6 l:tb8 20 �h1 and .l%gt +
will lead to mate) 1 9 .*.gS f6 20 i.f4 is messy,
Glauser-Ractsky, Lugano 2000.
b) 1 1 lib 1 allows 1 1 ...dxc4 1 2 .i.xc4 bS.
For example, 1 3 .*.e2 �7 14 a4 a6 1 5 c4
bxc4 1 6 11fc2 .i.xf3 17 .Lf3 Wh4 18 g3
1Wxd4 1 9 lld1 (in Ivanovic-Khalifman,
Plovdiv 1986 White now lost the thread with
19 .i.xc6?! llad8 20 .i.gS?! WcS 21 .i.xd8?!
'ifxc6 22 .i.a5 �5 23 llfd1 li:)f3+ 24 Wf1
li:)xh2+ 25 Wg1 li:)f3+ 26 Wft :CB! and
Khalifman had a winning attack) 1 9. .'iff6 20
.
.i.xc6 llad8 21 .i.xd7 l:lxd7 22 .i.f4 l:lfd8
with a level position (Khalifman).
1 1 J.xc4 �g4 1 2 h3
12 Wd3 is the subject of the next game.
1 2....i..h5 1 3 J:le1 l0d7
1 4 g4
White can also play Jess aggressive])' with
14 .i.ft. Now 1 4...l:lc8 1 5 llxe8+ Wxe8 1 6
i.e3 b5 1 7 a4 a6 (17 ...b4 1 8 cxb4 .*.xb4 1 9
J:lct li:)b6 is also equal) 18 axb5 axb5 1 9
59
Th e Petroff Defence
llxa8 11fxa8 20 dS cxdS 21 g4 i.g6 22 i.xbS
�7 was completely level in Ehlvest-Anand,
Unares 1 991 .
1 4 .. ..ig6 1 5 .*.g5 ••6 16 .*.e7! .*.xe7 1 7
llxe7clDb6
Or 1 7 ... :ad8 1 8 11t'e1 cS 19 .!DeS c!DxeS 20
:xeS lk8 21 f4 and White's initiative con­
tinues.
1 8 .i.b3 llab8 1 9 1i'e1
Taking control of the open file. lnst(!ad 1 9
c4 l(fe8 20 11'e1 'ifxet+ 2 1 ltaxe1 :xe7 22
:xe7 �fB 23 :C3 l(d8 is only equal.
1 9...•a3 20 1fe5 lDd6
If 20...'it'b2 White plays 21 l(d1 11fxc3 22
c!Dgs, when 23 c!De6 is a nasty threat.
21 .i.xd5 cxd5 22clDg51 1fxc3 23 �61
23...1i'xa1 +1
Taking a rook with check is always tempt­
ing, but now Black is lost. Black has to tty
23... £6, although White is clearly better after
24 :xg7+ Wh8 2S 'ifc7 'ifxc7 26 :xc7 l(fe8
27 .!DeS.
24 Wh2 f6 25 :Xg7+ WhB 26 'il'c7 1 -0
White threatens 27 :xh7+ followed by 28
11t'g7 mate, and 26...)%g8 is met by 27 fle7
after which f6 cannot be defended.
Game 2J
Hamdouchi-Le Roux
B e!fort 20()3
1 e4 e5 2 lDf3 clDf6 3 �xe5 d6 4 .lDf3
clDxe4 5 d4 d6 6 .*.d3 .i.d6 7 0-0 0-0 8
60
c4 c6 9 lDc3 clDxc3 10 bxc3 dxc4 1 1
.i.xc4 .i.g4 1 2 'il'd3 �d7
Sensible development. Black has a solid al­
ternative in 12 ... i.hS 13 c!Dcs bS 1 4 'il'h3
bxc4 1 S 'il'xhS 'WaS 16 i.d2 1t'dS (Black
doesn't win a piece with 16...f6?! because of
17 1i'g4 heading for c6; for example,
17 ...i.xeS 1 8 dxeS 'ifxcS 19 ltfel and
White's lead in development is overwhelm­
ing) 1 7 11'e2 :C8 18 i.e3 i.xeS 1 9 dxcS
"iVxeS 20 'ii'xc4 'ife4 21 1i'xe4 1lxe4 22 llfcll
c!Da6 23 l:ld7 and White had a solid advan­
tage in Abropov-Solozhenkin, Helsinki 2000.
Instead of 13 ...b5, Black should play the sim­
ple 1 3...i.xeS 1 4 dxeS 'iVxd3 1 S i.xd3 Zld8
1 6 i.c2 i.g6 17 i.xg6 hxg6 1 8 ltb1 b6 with
c:quality.
1 3 li)g5
Starting a kingside attack. 1 3 :C:1 allows a
clever trick with 13...bS 14 i.b3 .!DeS! 1S.
jfc2 (the idea is 15 dxcS?! i.xh2+, now 16
.!Dxh2 'iVxd3 1 7 c!Dxg4 :ac8 1 8 .i.e3 'il'xc3 is
clearly in Black's favour, Klingcr-Fuchs,
Obcrwart 1 993) 1 S...c!Dxb3 1 6 c!DgS g6 1 7
axb3 .i. f5 and Black has equnlised. I
1 3 ...�f6 1 4 h3 .i.h5 I
This leads to immense complications. AI
1
safe, reliable option is 14 ....i.d71? 1 5 i.b3 h6 '
1 6 c!Df3 jfas with equality.
1 5 f4 h6 1 6 g4
This is the critical try. Black has no prob- i
lcms after 1 6 c!Df3 .i.xf3 17 ltxf3 (or 17 11'xf31
'flc7 18 i.d2 cS with good counterplay)l
I

17...:e8 18 .i.e3 .i.c7 19 llafl 1Wd6 20 .i.f2
:C4 21 .i.h4 ltae8 22 �xf6 11'xf6 23 �ht
Act 24 a4 hS 25 g3 a6, Kamsky-Bareev,
Linares 1993.
1 6...b5
It is also posl>ible to take the piece imme­
diately: 16...hxg5 17 fxgS lLlxg4 18 hxg4 1Wd7
(18...�xg4? is a mistake: 19 1i'e4 1i'd7 20 g6
..ie6 21 .i.xc6 fxe6- Capablanca-Northrop,
New York 1909 - and here the simple 22
Wh4 is winning) 19 �3!? (the independent
try; 19 gxhS 9g4+ 20 �f2 ltaeB 21 ltg1
'1Vh4+ 22 �g2 b5 23 .i.b3 trans�cs to
16...b5) and now:
a) 19 ...'.xg4+ is safest: 20 1i'xg4 �xg4 21
llxf7 b5 22 .i.b3 �h7 with equality (Byrne,
Mednis).
b) 19 ....i.xg4 is interesting but also risky:
20 Wh4 .i.fS (or 20....i.e6 21 .i.d3 g6 22
l:t£6!? c5 23 .i.e3 and White has a dangerous
attack) 21 .i.£4 .i.xf4 22 llxf4 bS 23 .i.b3 and
White certainly has compensation. Now in
the game Moro7.evich-Ippolito, New York
1997 Dlack erred with 23...ltae8?, after which
White could have won with 24 Wf2! 11fe7 25
l:txfS 1t'e2+ 26 Wg1 :C4 2711fh3 (Ippolito).
1 7 .i.b3 hxg5 1 8 fxg5 li:Jxg4 1 9 hxg4
..d7
Another hugely complicated line is
19....i.xg4 20 g6 (or 20 1We4 .i.h3 21 g6
.i.h2+!- not 21....i.xft? 22 1Wht! and Black
l"lln resign - 22 Wxh2 .i.xfl with a very
complex position) 20....i.e6 21 llxf7!? (Black
3 li:Jxe5: Black Plays 6 ... .i.d6
tnkes over the attack after 21 gxf7+ .i.xf7 22
:Xf7 ltxf7 23 1Wg6 .f6 24 .i.xf7+ 1Wxf7 25
'iPxd6 ltf8) 2t....i.xb3 (21....e8? allows a
neat forced win: 22 'trh3! :Xf7- the trick is
22_..i.xh3? 23 ltfS+I- 23 Wh7+ Wf8 24 .i.g5
l1f6 25 ..ixf6 gxf6 26 g7+ and Black wiU be
mated -Nunn) 22 'iPh3 l:lxf7 231Wh7+ �£8
24 WhB+ We7 25 .i.g5+ J:l£6 26 .xg7+ We6
27 .i.xf6 'iPg8 28 axb3 .xg7 29 .i.xg7 WfS.
After all the tactics White has a slight edge in
the ending.
20 ..1671
Now Black has an edge. The correct way
to hold equality is 20 gxh5 .g4+ 21 Wf2
l:aes 221tg1 'trh4+ 23 wg2 "W'h2+1 24 wn
.i.f4 25 1i'f3 (not 25 .i.x£4? .xf4+ 26 ..t.>g2
llc3 and White must give up his queen)
25..llc1+ 26 �xel 'ii'xgl+ 27 We2 .i.x.cl 28
:Xcl .xcl 29 g6 lle8+ 30 Wd3 'iWb 1 +
(30 .lle7 31 gxf7+ wm 32 �3 'ilrbt+ 33
.
.i.c2 .xa2 34 1i'c8+ on - also draws) 31
-
Wd2 .el+ 32 Wd3 'ilrbt+ 1/z-'12 A.Sokolov­
OU, Odessa 1989.
20 ...hg4 21 •xd7 .i.xd7 22 lbf7 lbf7
23 g& .i.e8 24 .i.g5
White is also slightly worse after 24 �e3
aS 25 .i.e6 Wf8 26 gxf7 .i.xf7 27 llft lla7,
Nunn-Salov, Brussels 1988
24 85 25 �g2
•..
Black also has an edge after 25 l:lfl llaa7
26 llf3 a4 27 .i.e6 llac7 28 ..i.d8 l:lb7 29
.i.gS b4 30 cxb4 �xb4.
25 84 26 .i.e& �8 27 l:lh1 l:lf& 28
...
61
Th e Petroff Defence
hf6 gxf6 29 .*.f5 Wg8?1
Black fails to make the most of his
chances. Better is 29.. 1[b8 30 'itf3 (White
achieves nothing after 30 llh8+ 'itc7 31
llh7+ Wd8) 30 ..b4 31 cxb4 llxb4 with a
. 1i'g6+ 23 �h1 1i'e6 when a draw was a&rreed)
definite edge for Black.
30 l:.h7
Now the threat of 31 .i.e6+ forces Black
to bail out to a drawn ending.
30 hg6 31 ..i.xg6 .*.f4 32 .llb7 .*.d2
•..
33 ..i.e4 l:.c8 34 llb6 ..i.xc3 35 :.Xc6
:.Xc6 36 ..i.xc6 % - %
Game 26
V .Gurevich-Meijers
Germaf!Y 1999
1 e4 e5 2 lbf3 lbf6 3 lbxe5 d6 4 l/)f3
�e4 5 d4 d5 6 ..i.d3 ..i.d6 7 0-0 0-0 8
c4 c6 9 cxd5 cxd5 1 0 lt!c3 ltlxc3
This is the automatic answer to 10 lbc3,
but the rarely played 10...1le8 is also a con­
sideration. Now I I lle1 transposes to note
'c' to White's 10th move in Game 23, 11
.ixc4 dxe4 12 �g5 .ifS 13 f3 c3 14 �ge4
.if4 was unclear in Morozevich-Gelfand,
Cannes 2002, while 11 h3 and 11 ..c2 are
also serious possibilities for White.
1 1 bxc3 ..i.g4 1 2 llb1
Instead 12 h3 .i.h5 13 1rb3 is too ambi­
tious: 13....i.xf3 14 ..xb7 �7 15 gxt1 �b6
16 ltbt 'it'f6 17 �g2 llac8 18 1i'xa7? (18
llb31lc7 19 Wa6 .i.f4 20 .i.xf4 Wxf4 is bet-
62
rer but Black still has an edge) 18 . .1lxc3 19
.
llxb6 Axd3 20 .i.e3 1i'g6+ 21 �h1 'WfSI (in
Capablanca-Marshall, New York 1909 Black
missed his chance with 21...'1Fe6 22 Wg2
22 Wg2 Qf 22 llxd6 Black's idea is
22.. '1Fxh3+ 21 �g1 ltxe3! 24 fxe3 'Wg3+ 25
.
Whl 1i'xd6 with a clear advantage because of
White's exposed king) 22. ..1la3! 23 1rb7
1i'g6+ 24 Ciilh1 llxe3! 25 :Xd6 Wh5 26 fxe3
'irxh3+ 27 Wg1 1i'g3+ 28 Wh1 1i'xd6 and
Black has an obvious advantage.
1 2...b6
12...�7 is studied in Games 27-28.
1 3 .llb5 ..i.c7
The d-pawn was genuinely threatened. For
example, 13...a6? 14 %hd5! wins a clear pawn
because 14...i.xh2+? 15 �xh2 i.xdt 16
llxd8 ltxd8 17 llxdl is winning for White.
1 4 h3
14 c4 dxc4! leads to very sharp play after
15 i.e4 lbc6 16 llg5 (if 16 .i.xc6?! Black's
idea is 16...'1Fd6 17 ..LaB i.xf3 18 i.f4 1i'xf4
19 g3 i.xd1 20 gxf4 i.c2, when the queen­
side pawns will be very dangerous) 16...i.xt1
17 1i'xt1 '1Fd6 18 llg3 (18 g3 eventually pe­
ters out to equality after 18... �d4 19 1i'e3
llfe8 20 :d5 1i'xd5 21 .i.xd5 llxe3 22 fxe3
tbe2+ 23 �g2 :ds 24 i.xt7+ Wh8 25 i.xc4
�xcl 26 llxct) 18...�xd4 19 1i'g4 g6 20
.i.xa8 f5 21 'Wh4 l:lxa8 (the tempting
2L.l&2+? fails to 22 Wht �xg3+ 23 hxg3-
White is clearly better because 23..1l . xa8?

loses to 24 ...xc4+ rtlg7 25 .i.f4) 22l:r.h3 hS,
OU-Akopian, Manila 1992.
1 4 a6 1 5 hxg4
.•.
White has a promising alternative in 15
%lxd5!?. For example, 15...'tWxd5 16 hxg4
lbc6 17 lLlgS h6 18 .i.h 7+!? Wh8 19 .i.e4
1i'd6 20 lLlf3 l:.fe8 21 .i.c2 'ii'd7 (White has
an edge after 2t...b5 22 'ird3 'iVg6 23 'ifxg6
fxg6 24 .i.xg6) 22 g5 Wg4 (Benjamin­
Zamora, Philadelphia 1999). Now Benjamin
assesses 23 'ii'd3 g6 24 gxh6 .i.f4 25 i.b3
.i.xc1 26 llxct as clearly better for White.
1 5...axb5 1 6 •c2 g6
1 7 �h6
This seems the natural try. White has also
played 17 .i.xbS 'ird6 18 g3lLlc6 19 Wg2 f5
20 .if4 'ird7 21 gS i.xf4 22 gxf4 'irc7 23
.ixc6 'irxc6 24 �5 'ird6 (Black should
avoid chasing material with 24...'1'a4 because
;tfter 25 Wd3 Wxa2 26 l:r.ht White has a
promising attack - Nikcevic) 25 a4 %laS 26
l:tbt JibS with an unclear position,
i\.Sokolov-Nikcevic, Vmjacka Banja 1998.
17 ...Ilea 1 8 �xb5 Jle4 1 9 c4 lbg4 20
.
cxd5 liJd71?
Completing development is logical but
20....i.f4 is also worth considering: 21 lLleS
lth4?! (Black has to play 2t ...l%xg2+1 22 Wxg2
..ixh6 with an unclear position) 22 .i.xf4
llxf4 23 'ird2l:[f6 24 cli)g4l:[d6 (Kotronias­
Roi'.entali�, Debrecen 1992) and here White
could have gained a dear advantag.: with 25
llc I lLld7 26 1Wf4 cli)f6 27l:r.c61 (Lepeshkin).
3 liJxe5: Black Plays 6. .. �d6
21 .c6 liJf671
Black overlooks the coming tactic. Afrcr
21...lLlffi 22l:r.e l .i.f4 23lLle5l:r.h4 24 i.xf4
:xf4 the position is unclear.
22 o!tle5 l:lxd47
Now Black is lost. Instead 22...%lc8 23
cli)xt71? .xd5 24 1Wxd5 cli)xd5 25 i.d7l:r.xd4
26 i.xc8 Wxf7 27 lie I is only a little better
for White.
23 liJxf7! llh4
The obvious point is 23 Wxf7 24 11fe6
...
mate!
24 g3 1Vxd5 25 gxh4 1Vxc6 26 �xc6
lba2 27 o!tlg5 �f4 28 l:ld1 ..i.d2 29
liJxh71 .i.xh6 30 liJxf6+ �7 31 liJg4
�d2 32 Wg2 l:lc2 33 �e4 llb2 34 lla1
1 -0
Game 27
Pavlovic-Raetsky
Bie/ 1999
1 e4 e5 2 liJf3 liJf6 3 liJxe5 d6 4 liJf3
liJxe4 5 d4 d5 6 �d3 �d6 7 0-0 0-0 8
c4 c6 9 cxd5 cxd5 10 o!tlc3 liJxc3 1 1
bxc3 �g4 1 2 lib 1 liJd7
1 3 h3
Forcing the bishop to the kingside. In­
stead 13 l:r.b5 lLlb6 14 h3 gives Black the
option of 14.. ..i.d71?. For example, 15 l:r.bt
i.a4 16 'ire2 lieS 17 .i.e3 %lc8 18 cli)gs (the
game is level after 18 �5 .i.xeS 19 dxe5
l:.xe5 20 .i.xb6 :Xc2 21 Ld8 Zld2) 18...g6
63
Th e Petroff Defence
(18 ...h6 is very risky: 19 �xt7!? Wxt7 20
Wh5+ �g8 21 .i.xh61? and White has a
strong attack) 19 �xh7! (mstead 19 11'f3?!
l:lc7 was a bit better for Black in Kruszynski­
Raetsky, Katowice 1990) 19..1lxc3 (an ambi­
tious attempt to play for a win; after
19...�xh7 20 'ifhS+ �gB 21 Lg6 fxg6 22
.xg6+ Wh8 23 .i.gS .i.e7 White has a per­
petual check) 20 �gS 11'f6 with a complex
position.
1 3 ih6
..
1 4 1lb5
This is probably the most promising of
White's options. Alternatives include:
a) 14l:lxb7 �b6 15 .i.d2 .i.g6!? 16 .i.xg6
hxg6 171Ve2 (or 171Vb3 .c8 18l:lxb6 axb6
19 .xdS 11'c7 with an unbalanced position)
17 ...•c8 18 11'a6 l:ld8 is again unclear. Oliv­
ier·Raetsky, Geneva 1999 continued 19 llks
.i.xeS 20 dxeS l:ld7 21 l:lxb6 axb6 22 .xb6
l:lb7 23 11'd4 1Vc4 24 1fxc4 dxc4 and here
Black is a little better.
b) 14 a4 b6 15l:lb5 �£6 16 .tg5 .i.c7 17
g4 a61? 18 l:lb3 .i.g6 19 �eS .id6 20 �xg6
(an aggressive choice is 20 f4 .1xd3 21 1i'xd3
1i'c7 22 .i.x£6 gx£6 23 �B .1xf4 24 �4
l:lfe8 25 �fS. when White has reasonable
compensation) 20...fxg6 21 .i.h4 .1c7 22l:le1
gS!? 23 .i.xgS 1l'd6 24 l:les (White must
avoid the blunder 24 f4? h6 25 .tx£6 11'xf41
26 llb21lxf6) 24...1i'c6 and in this level posi­
tion the players agreed a draw in Semenova­
Chetverik, Balatonbereny 1994.
64
1 4...�6 1 5 c4
This is the normal move but 15 a4 is also
worth considering: 15...a6 (after 15....i.g6 16
.txg6 hxg6 17 1i'b3 Ilea 18 .tgs 11fd7 19 aS
�4 20 llxb7 .1c7 21 a6 11'c6 221Vb5 White
has a pleasant advantage) 161lxb6!? (the less
ambitious 16 lib1 leads to an unclear posi­
tion after 16...1lb8 17 .i.gS Wc7 18 aS �4
19 .i.xc4 Wxc4) 16...1l'xb6 17 .i.xh7+ �h8
(if 17...Wxh7 18lbg5+ Wg6 19 g4 White has
a strong attack) 18 .i.d3 with an unclear posi­
tion (Miralles).
1 5 .txf3
.••
15...�xc4 is less accurate because of 16
llxdS! (16 .txc4 .i.xB '17 1i'xR dxc4 18
l:lxb7 1i'e8 19 .i.e3 .e6 20 llct c3 21 dS
•es 22 g3 llac8, as in Kulaots-Lauk, Tallinn
2000, is only equal) 16....th2+ 17 �xh2
11'xd5 18 i..xc4 1i'xc4 19 •xhS 'iixd4
(19....xa2 is even worse: 20 dS llad8 21 l%d1
lld7 22 d6 and White is much better) 20 �0
.c4 21 a3 b6 22 .i.e3 llad8 23 l%ct 1l'd3 24
a4 and White has an edge, Dvoirys­
Vladimirov, Bamaul 1988.
1 6 •xf3 dxc4 1 7 .tc2
1 7 •..•d7
Black's two other options are probably in,..
ferior:
a) 17...a6 allows a dangerous attack: 18
.i.gS! 1i'c7 (the queen sacrifice 18...axb5 19
.i.xd8 llfxd8 is not sound: 20 11Fhs g6 21
1Wxb5 .i.c7 22 a4 l%xd4 23 •cs lld7 24 g3
and White was clearly better in Beliavsky-

Petursson, Reykjavik 1988; or 18...f6 19 'ii'hS
h6 20 .i.xh6! axbS 21 Lg71 f5 22 .i.xfB
.i.xfB 23 i.xf5 and Black's king is in trouble)
19 .i.f61 g6 (19...axbS? nms into a mating
attack after 20 .i.xh7+! Wxh7 21 'ifh5+ Wg8
22 9gS g6 23 1th6) 20 llb2 and White stiU
has a strong initiative.
b) 17...1i'c7 18 a4 a6 19 llb2 llab8 20
.i.h6! 9c6 (the point is if 20... gxh6 then 21
9f5 Wg7 22 'lfxh7+ Wf6 23 .xh6+ Wc7 24
lie 1+ Wd7 2S .i.f5+ Wc6 26 .i.t->4+ Wd7 27
llcbl and the attack breaks through) 21 9f5
g6 22 9a5 �d7 23 9c3 1fdS 24 .LfB �xfB
2S llet and \Vh.ite had an edge in Yandar­
biev-Skatchkov, St Petersburg 2001.
18 a4
18 llab8
•..
Supporting the b-pawn, but this is not the
only move:
a) 18...g6 is an important option - sec the
next game.
b) Black should avoid 1B ...i.c7?! 19 g3 a6
. 21 llh51?
20llhS g6 21 lies .i.ds 22 aS 9xd4, as in D.
Kaiumov-Chetvcrik, Harkany 1992. Now
White could have built a powerful attack with
23 .i.e3! 1ic3 241lk l llcB (not 24...�7? 2S
.ixg6 1i'xcl + 26 .i.xcl �xeS 27 .i.f5 and
White wins) 2S llhS. c) Worth considering,
however, is 18...llfe8!?, for example 19 .i.fS
ti'c7! 20 i.e4 a6 21 llg5 llxe41 (the careless
21 h6?? is refuted by 22 llxg7+! Wxg7 23
...
.i.xh6+!) 22 'lfxe4 �a4 23 lle l g6 24 h4
li'lc3 (also possible is 24...lld81? 25 h5 b5
3 /i)xe5: Bla ck Plays 6 . • . �d6
with unclear play) 25 9f3 � 26 L3 �b4
27 h5 �3 28 hxg6 hxg6 (Ponomariov-Safin,
Yerevan 2001) and here White should force a
draw with 29 'Wf6! .i.xa3 30 llhSI gxhS 31
'Wg5+.
1 9 �e3
Once again there is a choice:
a) 19 .i.b2 should be equal after 19..ic7
20 J:lcS l:fd8? (this is a fatal slip; 20....i.d6 2 1
l:b5 .i.c7 is level) 21 llh5 g6 22 d511fd6 23
g3 gxh5 (If 23...f6 then 24 llxh71 and the
attack crashes through) 24 9xh5 and White
was winning in Zagrebelny-Chctverik, Gyula
1992.
b) 19 .i.g5!? is a promising alternative to
19 i.e3. For example, 19.. .1lfc8 20 l:lfbt g6
21 .i.e3 Wc6 22 a5 'lfxf3 23 gxf3 �8 24
l:xb7 l:lxb7 25 l:xb7 lle7 26 l:b1 with an
edge for \Vh.ite, De Firmian-Kosebay, Co­
penhagen 1996.
1 9 /i)cS 20 llfb1 b6
...
nus is not \Vh.ite's only attacking method.
21 h4 liJc7 22 h5 h6 23 g4 .i.c7 24 gS a6 (if
24...hxg5 25 l:xg5 f6 26llg2 Wh8 27 1fe4 f5
28 1fh4 White has a promising attack) 25
l:l5b4 bS 26 gxh6 f5 27 axbS axb5 28 1fg2
llf7 with a complex position, Shilov-Ractsky,
Lugansk 1989.
21 ...g6 22 .i.h6 lld8 23 .f6 �fS 24
.i.xfB •xd4
The only move. 24...llxfB? loses to the
clever 2Sllxh71 Wxh7 26llb5. For example,
65
 The Petroff Defen ce
26...Wh6 27 llhS+I Wxh5 28 'iff41 and mate
is unavoidable.
25 •xd4
After 25 Jl..g7 'ifxf6 26 .i.xf6 ltd6 27 ::lh4
llxf6 28 ltxc4 �6 Black has won back the
piece and keeps a clear extta pawn.
25 ...:txd4 26 .i.c5 ltd2 27 :te5 :txc2 28
:tea+ �7 29 .i.d4+
29 .i.fB+ is le::;s clear: 29. ..'itf6 30 :dt c3
3t:dd8 with an ob::;cure position.
29 f6 30 .i.e37
••.
This spoils White's advantage. 30 ltdt c3
3t Jl..e3 Wf7 32 ::ldd8 lta2 33 ltxc8:Xc8 34
llxcS :Xa4 35 llxc3 promises some winning
chances.
30 . :tc3 31 ltd1 ltd3 32 :C1 ltd7 33
..
¢1 1 7
Instead 3 3 Jl..f4 :laS 34 llxc4 tbd6 35
:c7! holds on to equality.
33...g5 34 �e2 ¢17 35 llh8 Wg7 36
lte8 lte7 37 :td8 :C7 38 h4 h6 39 f47
The final mistake. 39 :d4 :bb7 40 hxgS
hxgS 41 ltcxc4 liJe7 still leaves White some
hopes of saving the draw.
39 ...gxh4 40 f5 :tbb7 41 .i.f4 ltc6 42
ltcd1 a6 43 lt1d4 h5 44 .i.c1 fi::Je7 45
:t8d7 ltcc7 46 :Xc7 :Xc7 47 .i.f4il)xf5
48 .i.xc7 il)xd4+ 49 �e3 lLlf5+ 50 ¢14
il)e7 5 1 We4 b5 52 a5 b4 53 � b3
0-1
Game 28
Anand-Shirov
Unarr:s 2000
1 e4 e5 2 il)f3 il)f6 3 lLlxe5 d6 4 il)f3
il)xe4 5 d4 d5 6 .i.d3 .i.d6 7 0.0 0-0 8
c4 c6 9 cxd5 cxd5 1 0 il)c3 fi::Jxc3 1 1
bxc3 .i.g4 1 2 :tb1 il)d7 1 3 h3 .i.h5 1 4
ltb5 fi::Jb6 1 5 c4 .i.xf3 1 6 •xf3 dxc4 1 7
.i.c2 •d7 1 8 a4 g 6 1 9 .i.e3
White's two other options are less threat­
ening:
a) 19 Jl..h6 :res 20 :fbt :ladS 21 i.gS
Jl..e 7 22 aS Jl..xgS! (this i::; more accurate than
22...�5 23:Xb7 'ife6 24 Jl..d2 .i.f6 25 ltet,
66
which gave White an edge in Svidler­
Akopian, Yerevan 1996) 23 axb6 .i.e7! 24
bxa7 b6 25 i.e4! 11'xa7 26 i.dS i.d6 27
:Xb6 :bSI 28 :Xb8 llxb8 29 llxb8+ i.xb8
30 Jl..xc4 tli'at+ 31 .tn tli'xd4 with a drawn
position (Svidler).
b) 19 i.d2 c3 20 i.xc3 :ac8 21 i.e4 :c4
22 :bb1 :xa4 (Black has a promising alter­
native in 22...:fc81? 2'\ .tat :xa4 24 i.xb7
:c7 25 i.e4 :a3 with counterplay - Barlov)
23 i.xb7 :V 24 :ret (or 24 i.c6 11fc7 25
:at :b3 26 :fb1 :xbt+ 27 :xb1 rlcs 28
dS liJxd5 29 :b7 1Wxb7 30 i.xb7 :Xc3 31
'IVdt i.f4! heading for a drawn opposite­
coloured bishop ending - Kaspamv)
24...tli'c7! 25 ltat %lb8 26 i.e4 ltb3 was
equal in Kasparov-Shirov, .Iinares 2000.
1 9...:ac8
20 ltfb1
After 20 i.h6 :res White has an unusual,
way to force a draw: 21 i.fS! gxfS 22 '1Vg4+11
fxg4 23 ltgS+ Wh8 Va-lfz Zaw Win Lay­
'
Pacicncia, Surabaya 2002.
20 c3 21 a5il)c4 22 :txb7
.••
This is promising but 22 .i.h6!? is also in­
teresting: 22.. %lfe8 23 :xb7 'ifc6 (not
.
23 ..11t'xb7? 24 '1Vxb7 %lb8 25 a6! and White
.
has a clear advantage) 24 :lfl! lte7 25 ::lxe7
i.xc7 26 ltdt (Greenfeld mentioned the·
unclear line 26 d51? 'ife5 27 .i.f4 'iff6 28
1llg4:cs 29 .i.h6) 26...�2 27 'ild3 .i.d6 28
a6! 'ifc4! and White was a bit better in
terman-Greenfcld, Haifa 2002.
AI-11

22.......6
More solid than 22...:C7 23 .ih6lle8 24
Axc7 'flxc7 25 'flxc3 when White has a dan­
gerous initiative. For example, 25...loa3? 26
llet! llb8 27 'flxc7 .i.xc7 281le7 ltlxc2 29
:Xc7 ltlxd4 30 llxa7 gave White a winning
ending in Motylev-Ristic, Novi Sad 2000.
23 1la1
Less threatening is 23 .ib3 'fit'S 24 11fxf5
gxfS 25 llal (Grischuk-Shirov, FIDE World
Championship, New Delhi 2000 ) and here
Grischuk suggested the equalising 25...ltld2!?
26 g3ltlxb3 27 :Xb3 c2 281lct llc4.
23...J..b8?1
Now Black is in real trouble. 23...Ac7 24
a6 lDxe3 25 fxe3 'fle7 26 l:.abt is only mar­
ginally better for White.
24 i.b3 .d6
If 24 ... c2 25 Act Wd6 26 :Xc2 9hz+ 27
Wft ltlxe3+ 28 fxe3 lLc2 29 .i.xc2 White is
clearly better (Shirov).
25 g3 f.i)xe3 26 J..xf7 + �h8 27 •xe3
.,6
Pushing the pawn again achieves nothing:
27...c2 28 llct llc7 29 llxc7 .ixc7 30 .ic6
i.xa5 31 d5 (Shirov).
28 .i.e& llceB 29 d5 .ie5
3 f.i)xe5: Bla ck Plays 6... J.. d6
30 lla2?
Now White loses control of the position.
Even worse is 30 'ifxa7? llxe6 31 dxe6 .i.d4
32 l:xh7+ WgS 33 'flb7 Wxf2+ 34 Wht c2!
35 :t? .i.xal 361lxf2 ct'fl+ 37 Wg2 :xl2+
38 Wxl2 .td4+ when Black has decem win­
ning chances. White could have maintained
an edge with 30 :b4!? llb8 31 :c4 c2 32
llct llbl 33 ll4xc2 .td4 34 Wf41lxc1+ 35
llxct .i.xf2+ 36 Wg2 (Rogers). However, the
best option is Shirov's later suggestion: 30
lla41 c2 31 llc4 .td4 32 1i'f4 Wxf4 33 gxf4
.i.xf2+ 34 Wg2 with good winning chances.
30....id4 31 ...,
31 1td3!? .txf2+ 32 Wg2 .i.d4 33 llc2
with an unclear position is better.
31 .....f3 32 Wh2 •xd5 33 .ixd5 llxe1
34 �2?
Now White is lost. 34 f4 :res 35 .to
llct still leaves White with good drawing
chances.
34 ...i.xf21 35 llf7
The point is 35 llxf2 llxf2+ 36 Wxf2 c2
and the pawn queens.
35 1lxf7 36 .ixf7 .ic5 37 .ib3 ¢>g7 38
•..
:C2 .id4 39 a6 ¢>16 40 lla2 'ite5 41 h4
�e4 0-1
67
The Petroff Defence

Summary
The variation 6...J.d6 7 0-0 0-0 8 c4 c6 has become fashionable and is likely to remain popular
because of the huge number of complicated, untypical and insufficiently studied positions.
After 9 .c2, defending the knight with 9... f5, as in the historic Wtll.iams-Staunton game
(London 1 851) isn't had because it's not that easy for White to take advantage of the weakened
a2-g8 diagonal. After 9 .c2 �a6 10 a3 Black has a few effective responses: 1 O J:e8 1 1 llk3
•.

J.f5; t t ...J.g4; or the immediate 1 0....i.g4, responding to 1 1 lDe5 with 1 1 .....tf5, 1 t ..J.hS or
1 1 .....txe5 12 dxe5 �c5. The fashionable 1 o...f5 appc..-ars to give White the better chances after
an accurate attack on the queenside with 1 1 �c3 �c7 1 2 lDe2 � 1 3 b4.
In the event of 9 1let. the pawn sacrifice 9.....tg41? deserves careful consideration. After 10
Le4 dxe4 1 1 llxe4 f5 1 2 lle6 1id7 Black has full compensation, while he has prospects of
equalising after 12 1let ..txf3 13 Lf3 1t'h4. In the case of 9 ... J.f5 White can claim an advan­
tage with 1 0 llk3 �xc3 1 1 bxc3 ..txd3 1 2 Wxd3 dxc4 1 3 11'xc4 �7 14 ..tgS!?, while after
9...1le8 10 ru �xc3 1 1 bxc3 the position appears to be favourable for White.
To us, the move 9 �c3 seems less accurate than 9 cxd5 cxd5 1 0 llk3 because Black can
proceed with 9...�xc3 10 bxc3 dxc4 1 1 Lc4 ..tg4. We can't find any white's advantage in
either the calm variations or the sharp 1 2 1id3 �7 1 3 clOgs �f6 1 4 h3 .i.h5 1 5 f4 h6 1 6 g4.
Recently 9 cxd5 cxd5 1 0 ru �c3 1 1 bxc3 .i.g4 has not been so popular. White's position
seems to be more promising after 1 2 1lb1 b6 but. if Black defends accurately. White will not
gain an advantage after 12 ..�7. Nowadays there are fewer fans of this complex line, which
requires knowledge of variations developed earlier and where t.-very moment is critical.

1 e4 e5 2 �f3 �f6 3 lnxe5 d6 4 �f3 lnxe4 5 d4 d5 6 ..i.d3 .i.d6 7 0-0 0-0 8 c4 c6

(D) 9 cxd5
9 1fc2 �a6
10 .ixe4 Ga111t 19; 1 0 a3 .i.g4 - GtUJJe 20; 1 O... f5 - Game 21
-

9 1le l
9....if5 - Game 22; 9...1le8 - Gan1e 23
9 llk3 �xc3 1 0 bxc3 dxc4 1 1 .lxc4 .i.g4 (D): 12 h3 - GttnJe 24; 1 2 .d3 - G011Je 2;
9 cxd5 10 �c3 lnxc3 1 1 bxc3 .i.g4 1 2 .Rb1 (D) �7
••.

12 ...b6 - Gm11e 26
1 3 h3 .i.h5 14 .Rb5 �6 1 5 c4 .i.xf3 1 6 11xf3 dxc4 1 7 .i.c2 ...d7 1 8 a4
1 8. 1lab8 Game 27; 18...g6 - Gm11e 28
.. -

B. . . c6 1 1 ... i.g4 12 fl.b 1

68
 CHAPTER FOUR I
3 ltJxe5: Deviations
from the Main Line
1 e4 e5 2 t0f3 �f& 3 o!bxe5 d& 4 �f3
o!bxe4 5 d4
A fter 1 e4 cS 2 lLlf3 lL!f6 3 lLlxeS d6 4
ll:)f3 lLlxe4 5 d4 we've already identified two
principal trends: S...dS 6 �d3 �e7 7 0-0 c!bc6
(< :hapters 1 -2) and 6 ....i.d6 (Chapter 3). It is
reasonable to devote a separate chapter to
deviations from these �nes. Black rarely dc­
c�nes to play the main continuation S...dS in
favour of the passive s...i.e7 (Game 32).
After 6 �d3 the knight retreats to f6 rather
than to gS (in view of 7 li:)xgS ..ixg5 8
'ii'c2+). l11en White normaUy proceeds with
7 h3 in order to restrict the c8-bishop.
After 5 d4 dS 6 ..td3 Marshall suggested
6 .....tg4, which failed the test of time in view
of 7 "iVe2+1 'fle7 8 0-0 li:)c6 9 �bS, when it is
difficult for Black to complete his develop­
ment. In Game 31 we discuss MarshaU's idea
of an early .....tg4 improved by the insertion
of the moves 6 ...li:)c6 7 0-0. The two critical
lines arc 6 ...c!bc6 7 IJ-0 ..ig4 8 c4 lLlf6 9 cxdS
..txf3 10 'tixf3 'it'xdS 1 1 'it'xdS (or 11 "iVe2+;
I I l:cl+ �c7 reaches a tabiya from Chapter
2) and 9 li:)c3 i.xf3 10 "iVxf3 li:)xd4 1 1 'it'c3+
(or 1 1 'it'h3; again 1 1 l:e l + ..ic7 rt-aches
Chapter 2). l r should be said, however, that
in modem practice Black usually avoids giv­
ing his opponent a variety of possibilities and
often develops his bishop to e7 early on.
After S ... dS 6 ..id3 i.e7 7 0-0 the idea
7 ...ll:)d6 is played very seldom, with Bilguer
declaring this move a 'motiveless retreat'. In
fact, there arc some motives, i.e. a counter­
measure to c2-c4 and a preparation of .....i£5.
However, a fter 8 i. f4! Black fails to develop
his bishop to f5 and is forced into passive
defence.
If Blacks wishes to move his bishop to f5,
he should do so immediately after 5 ... d5 6
�d3 i.e7 7 0-0 �f5. Then 8 c4 can be met
by 8...dxc4 since the rook is not yet attacking
the knight on c4, and the arising position
with an isolated pawn gives approximately
equal chances. 8 :Ct lL!c6 (Games 29-30) is
more common, after which the position
branches. Game 29 reviews the fashionable 9
lbbd2 as weU as 9 lbc3 and the pin 9 i.bS,
taking control of the eS-squarc; Game 30 is
devoted to the more fundamcnral 9 c4.
Game 29
Kasparov-Karpov
Mo.rco»' 1981
1 e4 e5 2 t0f3 �f& 3 lDxe5 d& 4 �f3
�xe4 5 d4 JJ..e7 6 .i.d3 d5 7 0-0 JJ..f5
Retreating voluntarily with 7...li:)d6 makes
White's task easier: 8 i. f4 0-0 9 :let ..ie6 (or
9.....ig4 10 h3 ..ihS I I l:leS!? i.xf3 12 1i'xf3
69
Th e Petroff Defence
c6 1 3 c3 �7 14 Ae2 ltlf6 1 5 � with a
clear advantage, Lau-Mathe, Munich 1992)
t o c3 �7 1 1 1Vc2 h6 (this is forced as
1 1 ...g6?! allows the standard tactic 1 2 i.h6
lle8 1 3 lLe6! fxe6 1 4 i.xg6, when White is
much better) 12 ltlbd2 :e8 1 3 ltlft ltl£8 14
ltlg3 i.g5 15 1Vd2 f6 16 h4!? Lf4 1 7 1fxf4
'ird7 1 8 ltlh2!? 1ff7 1 9 ltlg4 and White had a
promising initiative in Timman-Skembris,
Corfu 1 993.
8 �1
This is more promising than the alterna­
tives:
a) 8 ltles ltld7 9 1ff3 ltlxe5 10 dxe5 'ird7
1 1 ltlc3 i.g4 1 2 e61 (a neat trick but Black
has a defence) 1 2...i.xe6 (not 12 ...'irxe6?! 1 3
ltlxd5! 1Wxd5 14 1Vxg4 ltlf6 1 5 Wa4+ 1Wd7
16 1Vxd7+ �xd7 1 7 :e1 - Ro. Perez - when
the bishop pair and especially the nasty pin
give White a clear advantage) 1 3 ltlxe4 dxe4
14 i.xe4 c6 1 5 :d1 1Wc8 1 6 :ct 0-0 1 7 Whs
f5! 1 8 i.d3 i.f6 19 c3 :Cs and Black had
equalised in Sutovsky-Ro.Pcrcz, Istanbul
Olympiad 2000.
b) 8 c4 has the obvious drawback of los­
ing time with White's Icing's bishop. 8...dxc4
9 i.xc4 0-0 10 �c3 lL!c6 and now:
b1) I f t 1 i.d3 rhe correct answer is
1 1 ...�c3 12 bxc3 'iVd7 13 :e1 i.d6 14 :b1
b6 1 5 :bs i.xd3 1 6 'irxd3 :res, when Black
has no problems (but not 1 1 ...�d6 12 i.xfS
ltlxf5 13 d5 ltlb4 1 4 a3 lba6 1 5 1rc2 �h4 1 6
�xh4 i.xh4 17 :ld 1 , when White has an
70
edge, Zhang Pengxiang-Delchev, Unares
2002).
b2) 1 1 lle1 �6 1 2 i.d3 i.f6 1 3 dS
i.xd3 14 1Wxd3 ltlb4 1 5 1i'd1 aS (a sharper
option is 1 5...ltlf5 1 6 g4 �6 1 7 a� ltla6 1 8
i.f4 1Vd7 19 h3 with an unclear position
according to Tal) 16 a3 t'ba6 17 .i:e3 :Cs 1 8
'iVa4 �b8!? 1 9 i.d4 (Komeev-Ro.Perez,
Collado Villalba 2000) and here Perc?. now
suggests the correct route to equality is
1 9...ltld7!? 20 :xeS+ �xeS.
8 &6 9 �bd2
...
This is the best of White's alternatives (the
main move 9 c4 is discussed in the next
game). Other options include:
a) 9 i.bS i.f6 t O �bd2 0-0 1 1 �f1 �e7
1 2 c3 ltlg6 1 3 i.d3 �6 14 i.xf5 ltlxf5 1 5
'iVb3 b6 16 1rbs a6 1 7 'iVd3 1Wd7 1 8 ltlg3
�xg3 1 9 hxg.� aS 20 i.g5 i.xg5 21 ltlxg5
:reB and Black has easy equality, Karpov­
Korchnoi, World Championship (Game 4),
Merano 1 981 .
b) 9 ltlc3 ltlxc3 10 bxc3 i.xd3 1 1 cxd3
(1 1 'irxd3 also gives nothlng after 1 1 ...0-0 1 2
c4 dxc4 13 1rxc4 i.f6 1 4 c3 :eS) 1 1 ...0-0 1 2
1i'b3 llbs 1 3 :e2 bs 1 4 i.d2 1Wd7 1 s :net
:res and Black had equalised in A.lvanov­
Pierrot, Buenos Aires 2003.
9 .c!Dxd2 1 0 'irxd2 .i.xd3 1 1 'irxd3 0-0
••
1 2 c3
Kasparov quietly defends d4 before start­
ing active play. 12 i.f4 is also possible:
12 ... i.d6 13 ll)g5 g6 1 4 Wh3 h5 1 5 i.xd6 (1 51
 3 li:lxe5: D e viations from the Msin L ine
i.e3 illustrates the potential weakness of d4:
1 5...lnxd41 1 6 ttlxf7 ttlf3+! 1 7 Wxf3 lLf7 1 8
'lfh3 llf5 and BL1ck has equalised) 1 5...cxd6
1 6 ttlf3 1Vb6 1 7 l:tab 1 ttlxd4 1 8 ttlxd4 Wxd4
1 9 Wd7 and White's activity certainly com­
pensates for the missing pawn but is proba­
bly not enough to force an advantage.
1 2...1t'd7 1 3 .i.f4 a& 1 4 J:le3 J:lae8 1 5
J:lae1
White is more active but Black is very
solid and simplifying exchanges seem likely.
1 5.....td8 1 6 h3
Safeguanling the back rank and keeping
the tension. 1 6 �5 eases Black's task after
1 6...ttlxe5 17 dxe5 'Wbs 1 8 b3 Wxd3 1 9
l:lxd3 c6 20 c4 i.a5 2 1 llbt dxc4 22 bxc4
lld8 23 lla3 (Kavalck-Smyslov, Amsterdam
1 981) and here the simplest path to equality
is 23. ..i.d2 24 i.xd2 llxd2.
.
1 6 J:lxe3 1 7 llxe3
••.
1 7 f6
.•.
Avoiding a nasty back rank trick:
1 7...lte8?! 18 Wf5! l:te6 1 9 h4 g6 20 'irh3
1i'c8 21 ltxe6 'l'xe6 22 'irxd) fxe6 23 tl\g5!
and White has a clear advantage (Kasparov).
18 J:le2
This simply improves coordination while
awaiting developments. In a later game 1 8
ll:\12 was tried and following 1 8...ttle7 1 9
tL'lbJ 'irf5! 20 'l'xf5 ttlxf5 2 1 l:tc2 b6 22 tl\c1
<tlf7 23 ttld3 aS 24 J,>4 ttle7 a draw was
a�-,rreed in Leko-Kramnik, Dorunund 1 999.
18 Af7 1 9 li:ld2 .i.e7
•..
Playing directly for exchanges with
1 9...l:te7 leads to some difficulty: 20 tL\b3
llxe2 21 Wxe2 i.e? 22 'l'g41 Wxg4 23 hxg4
i..d6 24 i..xd6 cxd6 25 f4 and Black's dou­
bled d-pawns ensure White of a slight edge in
the ending.
20 ltlf1 .i.f8 21 1t'f3 J:le7 22 ltle3
The assessment is typical of this variation:
White has a nagging pull.
22...�8 23 .i.xc71?
An interesting way to unbalance the play.
23 ...1t'xc7 24 ltlxd5
24 1t'd6?1
•.•
Black should prefer 24...:Xe21? 25 ttlxc7
l:tet+ 26 ..t>h2 i.d6+ 27 g3 .ixc7 28 ...d5+
�f8 29 Wg2 with unclear play: Black has
plenty of material for the queen but regaining
coonlination without losing more pawns will
be very difficult.
25 li:lxe7+ .be7 26 1t'e4 .i.f8 27 1t'e8
Kasparov su�>ested that 27 c4 b6 28 d5
gains a clear advantage.
27 g6 28 a4
•..
Now Black manages to untangle. 28 h4!?
Wg7 29 h5 keeps an edge.
28 ...Wg7 29 b4 1t'c7 30 J:le3 li:lf7 3 1
1t'e& •d8 32 a5 h5 33 ee4 ed7 34 ••6
1t'd8 35 �1 li:lh6 36 g4 hxg4 37 hxg4
li:lf7 38 We2
Or 38 f4 'ire? 39 llf3 i..d6 40 1i'e4 Wc8
and Black still holds on to CI.Juality.
a8 ltlg5 39 as •d7 40 Wc�3 ..td& 41
.••
�2 % -%
71
The Petroff Defence
Came JO
Herrera-Ro .Perez
Varadero 2000
1 e4 e5 2 �f3 �f6 3 �es d6 4 �f3
�e4 5 d4 d5 6 .id3 .te7 7 0-0 .tf5 8
lle1 �c6 9 c4
9 o!Ob4
•••
Provoking a tactical sequence while fight­
ing for control of d5.
Instead 9 ...0-0 makes it much easier for
White to fight fur the advantage:
a) 1 0 cxd5 ..xd5 1 1 llk3 (the tactics fa­
vour Black, who is clearly better after 1 1
..c2? li)b4 12 .Le4 li)xc2 13 .i.xd5 .llaeB!
14 .i.xb7 .i.b4 15 :XeS llxe8 16 .i.c6 l:b8,
Shirov-Ivanchuk. Monaco (rapid) 2002)
1 t ...li)xc3 12 bxc3 llfe8 (Black can try to
blockade with 1 2... b5 but after 13 L5
•x5 t4 li)e5 li)a5 1 5 a4 b4 16 .td2 White
has a strong initiative) 13 .i.f4 .i.xd3 1 4
..xd3 ..d7 1 5 :C3 .i.f6 16 llae1 lle7 1 7
.i.g5 :Xe3 1 8 fxe3 .i.xg5 1 9 li)xg5 J!f1 20 e4
and White's impres.'iive centre gave him good
chances in Yagupuv-Sorokin, St Petersburg
2001 .
b) Also good is 1 0 llk3 lbxc3 1 1 bxc3
.i.xd3 1 2 ..xd3 dxc4 1 3 ..xc4 with a further
branch:
b1) 13.....d7 1 4 llb1 b6 15 d5 � 1 6
1Wd3 llfe8 (or 1 6...c 6 1 7 c4 cxd5 1 8 cxd5
llad8 19 .i.gS! and White keeps the initiative)
72
17 .tf4 l:ad8 18 llbd1 and White has a slight
edge. Now the game Bruzon-Ro.Perez, Santa
Clara 2000 continued 18 ... .i.f6 19 h3 :Xe1+
20 llxet b5?! 21 .i.g5 llk4 (Nogueiras gives
2t ....i.xg5 22 li)xg5 5 23 d61 1Wxd6 24 Wx5
with a winning advantage for White) 22
.i.xf6 gxf6 23 llc..--41 and White has a clear
plus.
b2) 1 3....td6 14 �5 ..f6 15 llbt (1 5
..xb7?1 grabs a poisoned pawn and Black has
tricks based on a later ....txh2+, i.e. 1 5.. .1lab8
1 6 ..a6 li)xd41 17 ..d3 li)xf3+ 1 8 WxO
1rxt3 19 gxO a6 and Black's stronger pawn
structure gives him an edge) 15 ... a6 16 1Wd3
b5 17 a41? ...g6 (or 17 ...bxa4 1 8 1la1 l:lfe8 19
l:lxe8+ llxe8 20 llxa4 aS 21 .i.d2 llb8 22
l%a 1 and White is a bit better) 1 8 1Wxg6 hxg6
19 axb5 llfb8 20 c4 axb5 21 c5 .i.f8 22 .i.f4
lla4! and despite Black's enterprising play
White still has an edge, Abreu-Ro.Percz,
Cuba 2003.
1 0 cxd517
Now the play becomes very complicated.
The simple 1 0 .i.ft also allows White to try
for an edge, for example:
a) 1 o... dxc4 leads to very sharp play after
1 1 li)c3 (or 1 1 .1xc4?1 0-0 12 a3 ll)d6! 13
.txf7+!? :Xf7 14 axb4 .i.g4 1 5 li)bd2 .i.g51?
and Black has superb compensation for the
pawn) 1 1 ...lbf6 1 2 .i.xc4 0-0 1 3 a3 llk6 (try­
ing to win material with 13 ...lbc2?! fails to 14
lilh4 .i.e4 1 5 lllxe4 lllxa1 16 lilts, when
White has a large advantage - the knight on
at has no chance of escaping) 14 d5 ll)a5 15
.i.a2 c5 16 .i.g5 :Cs 17 1Wa41 .i.d7 18 1Wc2
h6 19 .i.h4 and White had a strong initiative
in Karpov-Portisch, Tilburg 1982.
b) 10 ...0-0 1 1 a3 llk6 12 cxd5 (1 2 ltlc:J
ltlxc3 13 bxc3 dxc4 14 .i.xc4 rransposes to
7...liJc6 8 c4 ltlb4 9 .i.e2 0-0 10 li)c3 .i-5 I 1
a3 lilxc3 1 2 bxc3 liJc6 13 lle1 dxc4 14 .i.xc4
- see Game 2) 12...1l'xd5 13 ltlcJ lllxc3 14
bxc3 .i.f6 1 5 .i.f4 ..d7 16 lla21? llac8 17
l:ae2 :Xe2 18 1fxe2 li)a5 19 lbe5 1Wa4 20
g4!? (White's advantage is his activity rather
than his structure, so it is vital to continue
 3 li:Jxe5: Deviations from the Main L ine
aggressively) 20....i.e6 (or 20....i.xe5 21 .i.xe5
.i.d7 22 .i.xc7 lieS 23 'ifd2 'ifxa3 24 .:.xeS+
.i.xe8 25 •g5! and White keeps an edge) 21
g5 .i.d8 and White was a bit better in Korn­
ec.'V-Ro.Percz, Albacete 2000.
1 0 ...lnxf217
This is the critical try. 10...�xd3 1 1 '1Vxd3
'ifxd5 transposes to 7...�6 8 c4 �b4 9 cxdS
�d3 10 9xd3 'IVxdS 1 1 .:.e1 .i.t5 (sec
Games 9-1 0).
1 1 .84+
The only good move. If 1 1 .tb5+?! c6 1 2
.a4 0-0 1 3 dxc6 bxc6 14 Lc6 �fd3 Black
has a clear advantage. 1 2 dxc6? is even worse:
1 2...�xd1 13 c7+ '1Vd7 14 .i.xd7+ .i.xd7 1 5
�a3 �f2! and Black i s winning.
1 1 ....id7
1 2 i.b5
Again White has no choice. After 12
1i'xb4? �xd3 1 3 .:.xc7+ '1Vxe7 14 '1Vxb7 0-0
1 5 .i.d2 Wc2 Black wins easily.
1 2.../l)fd3
Advancing further into White's position
with 1 2...�bd3?! leads only to trouble: 1 3
.:.e2 a6 (or 1 3...�xc1 14 .:.x.£2 �3 1 5 .:.d2
�f4 16 �5 .i.xb5 1 7 9xb5+ <i'fB 1 8 .:.f2
and White is much better) 1 4 .i.xd7+ '1Vxd7
1 5 'ti'xd7+ <i'xd7 1 6 .:.xf2 �xf2 17 <i'xf2
.i.d6 1 8 � and White's material advantage
is almost decisive.
1 3 1le3 0-0
1 3....i.xb5 is no improvement 14 'IVxbS+
�f8 15 .:.xd3 �2 16 9c4 �xa1 17 b3 and
the wayward knight wiU soon be coUected.
14 i.xd7 /l)xc1 1 5 �c3
1 5... lDcd3
Black can also try 1 S...ltkxa2, after which
White's best is 1 6 .:.xa2!? �xa2 1 7 �a2 a6
1 8 .i.t5 with an edge. Instead after 16 �xa2
�xd5 Black hits c3 and thn:atens ...�b6. 17
:act �xe3 1 8 .:.xe3 .i.d6 leading to an un­
clear position.
1 6 a3 lDc5
Ro.Percz as.'ICsses 16.Jnxb2 17 9b3 lilc4
1 8 llxe7 'flxe7 19 axb4 as a bir better for
White.
1 7 dxc5
This aiJows White to keep some control
lnstc.-ad 17 '1Vxb4 aS 18 'it'bs �xd7 1 9 :act
.i.d6 20 1Wxb7 .:.b8 21 Wa7 .:.xb2 22 '1Vxa5
�b6 is very messy.
1 7 ....hc5 1 8 axb4 i.xe3+ 1 9 �1
White's two minor pieces are preferable to
Black's rook and pawn.
1 9...1tf6
t 9... a6 20 .:.et .i.b6 21 .i.£5 is also mar­
ginaUy in White's favour.
20 i.b5 1lad8 21 11rc2 •f4 22 .i.d3 g6
The other way to cover h7 is 22...h6, but
White keeps the attack going with 23 bS
.:.res 24 .:.n .
23 b5 lltea 24 :n ts 25 a3 •de 26
1i'c4 26 ...1tc57
Overlooking a nasty trick. 26.. Wh8 27
.
.i.c2 :d7 28 �4 limits White to ju<�t an
edge:.
73
The Petroff Defence
27 1i'xc5?1
White misses it as weU: 27 d6+! q;g7 28
dxc7 :d7 29 b6 is effectively winning.
27 i.xc5 28 i.c4 h6?1
••.
This aUows White a second chance for the
same tactic. Instead 28 ....i.d6 29 l0g5 lieS 30
l0e6 lieS 31 g3 is very unpleasant, but at
least Black can play on.
29 d6+ Wg7 30 dxc7 .ZZ.d7 31 b61 a6 32
i.e& 1 -0
Game 3 1
Kupreichik-Vusupov
Minsk 1987
______________..
1 e4 e5 2 lbf3 lbf6 3 lbxe5 d6 4 li)f3
lbxe4 5 d4 d5 6 i.d3 i.g4
Or 6...l0c6 7 0-0 .i.g4.
7 0-0
Also interesting is the simple 7 'lre2 'fie?
(instead 7 ... f5?! allows an instructive trick: 8
h3 .i.hS 9 g4! fxg4 to ltics .i.f7 1 1 hxg4 and
White is clearly better) 8 0-0 l0c6 9 .i.bS and
White has an edge, which will grow consid­
erably if Black plays too direcdy with 9...a6
10 .i.xc6+ bxc6 I 1 lle1 'iie6 1 2 l0bd2.
7 lbc6 8 c4
.•.
Immediately undermining the e4-knight
but 8 :e1 is also possible. Now 8 ....i.e7
transposes to Chapter 2, while 8... f5 is an­
other possibility: 9 c4!? .i.d6?! (Black must
play 9 ....i.b4, which leads to very unclear play
after 10 lle3 0-0 1 1 a3 .i.aS 1 2 cxdS 'lrxdS)
74
10 cxdS .i.xf3 1 1 1ixf3 l0xd4 12 'lre3 (this is
simplest but 12 1id1 1? also works:
1 2 ....i.xh2+ 1 3 �fl 'ifxdS 1 4 ltic3 and White
is much better) 1 2......f6 1 3 .i.xe4 fxe4 14
l0d2! (less ambitious is 1 4 'lrxe4+ Wf7 15
.i.gS 'iVxgS 16 'ifxd4 llhe8 1 7 l0c3 when
White was only a litde better in Capablanca­
Marshall, New York 1910) 1 4...l0f5 1 5
9xe4+ liJ.c7 1 6 l0 f3 h6 17 9a4+ Wd8 1 8
'iVbs Wc8 19 .i.e3 when White's extra pawn
and safer king ensures a huge advantage.
8 ...lbf6
Trying to grab material with S. ...i.x£3 9
.
'lrxf3 l0xd4?1 backfrres after 10 'lre3 l0f5 1 1
'lrf4 'lrd7 1 2 .i.xc4 dxe4 13 'ifxe4+ .i.e7 1 4
'lrxb7 0-0 1 S l0c3 when White is clearly bet-
ter.
9 l003
Developing and increasing the pressun: is
the correct approach. The simple 9 cxdS
leads nowhere after 9 ... .i.xf3 (9...l0xd5 10
ltic3 .i.e? transposes to 6....i.e7 7 0-0 ltic6 8
c4 l0f6 9 ltic3 .i.g4 10 cxdS lClxdS - see
Game 1 1) 10 'iVxO 'iVxdS 1 1 'iVe2+ (1 1
'lrxdS ltixdS 12 �3 0-0-0 1 3 .i.c4 ltice7 1 4
.i.d2 't!f> 1 5 D.fe1 .i.g7 1 6 .i.g5 1ld7 17 .i.xe7
ltixe7 1 8 .i.xf7 Wb8 1 9 lle2 lOeB 20 .i.b3
llxd4 21 g3 aS, as in Adams-Piker, Dort­
mund 2000, is also a1ual) 1 t ..ie7 12 .i.bS
'ii'd6 1 3 ltic3 0-0 1 4 .i.xc6 bxc6 1 5 1ldt llfe8
1 6 1WO lOdS t 7 l0xd5 'lrxdS 1 8 'i'xdS cxdS
19 .i.f4 c6 20 JLct llac8 and White had no
more than ei.Juality in Short-1-I.Oiafsson,
 3 fi:lxe5: De viations from th e Main L ine
Reykjavik 1987.
9 .i.xf3
•.•
9..ie7 ttansposes to 6...i.e7 7 0-0 ltk6 8
c4 lt.)f6 9 lt.)c3 ..tg4 (sec Game 1 1).
10 1i'xf3 li:lxd4 1 1 Wh3
This is probably the most dangerous op­
r:ion but 1 1 'ire3+ is also worth considering:
t l ...c!De6 12 cxdS lOxdS 1 3 lt.)xdS ..xdS 1 4
..ic4 'irbs I S a4 Wa6 1 6 llcJt Jl.c7 (the seem­
ingly more active 1 6....i.c5 simply helps
White to open lines: 1 7 1i'f3 c6 1 8 b4!? i.xb4
19 lib I Jl.cS 20 :Xb7 and White has the
initiative) 17 b4! 0-0 (1 7 ...Jl.xb4 is too greedy:
1 8 .i.b2 0-0 1 9 11h3 �5 20 Wg4 .i.e7 21 h4
and White has a strong attack) 1 8 1Vh3 br6 1 9
'ifc3 (keeping Black'll queen out o f the game;
instead 19 Jl.b2 allows 19 ...Wc41 20 l:d7
J:lae8 21 i.dS 'ifxb4 22 Jl.c3 lt.)f4 23 .i.xb4
lt.)xh3+ 24 gxh3 i.xb4 25 :Xc7 b6 26 llxa7
�g7 when Black has cleverly equalised, Kas­
parov-Karpov, World Championship (Game
6], London/Leningrad 1 986) 19 ...lt.)g5 20
..ixgS (20 i.b2 may seem more natural but
20 ...i.f6 21 1i'xf6 1fxf6 22 ..txf6 lt.)xe4 23
.
.i.e7 J:lfe8 24 l:ld7 lt.)c3! equalises according
to Kasparov) 20...i.xg5 21 Wxc7 llad8 22
ltxd8 llxd8 23 ..txb7 We2 24 i.o -.,2 25
lift Wxb4 26 'irxa7 (Bauer-Koch, Belfort
2002). Clearly White's extra pawn makes his
position preferable, but the oprosite­
coloured bishops will make any winning at­
tempt very difficult
1 1 dxc4 1 2 hc4 .i.e7 1 3 .i.g5 0-0
.•.
This is certainly the natural move but
Black should seriously consider 1 3 ...1i'c8. For
example, 1 4 'ire3!? �6 (14...ltk2 is too
·
greedy: 1 5 We2 'lrg4 - 1 S...lt.)xa1? ill refuted
by 1 6 ..ixf6 gxf6 1 7 l0ds - 1 6 ..xg4 �g4
1 7 Jl.xe7 Wxe7 1 8 lfitS+ Wf8 1 9 :act l0d4
20 lt.)xc7 l:d8 21 J:lfe1 and Black's awkward
king delays his development, giving White a
pleasant edge) 1 5 Jl.xf6 .i.xf6 1 6 lDd5 1Wd8
1 7 1tb3 i.d4 18 •xb7 0-0 1 9 llab1 ltb8 20
1fa6 c6 21 1fxc6 i.xb2 and Black has es­
caped to �::quality, Bologan-Koch, Belfon
2002.
1 4 llad1 c5
If 14 ... h6 White should play calmly with
IS ..tf4 cS 16 lt.)bS when his initiative is very
threatening. 1 5 .i.xh6?! ill premature:
I S...gxh6 1 6 J:ld3 Wh7 17 llg3 �8! and
Black defendll successfully.
1 5 llfe1
White's lead in development and the vari­
ous potential pins constitute good compensa­
tion for the pawn.
1 5 h6 1 6 hh&l
•••
This is even stronger than the tempting 1 6
llxe7!? hxgS (Black can give up the queen
with 1 6...1fxe7 1 7 lOds lt.)xdS 1 8 .i.xe7
lt.)xc7, but after 1 9 b4 b6 20 bxcS bxc5 21
'ifhs White still has an edge) 17 llxb7 llb8
1 8 llxb8 ..xb8 1 9 b3 'ifeS when the position
is unclear according to Yusupov.
1 6...gxh6 1 7 1i'xh6 fi:lh7
1 B lld3
75
The Petroff D e fence
This is the right rook. After 18 lle3?! �g5
19 D.g3 cRh8 20 �5 f5 Yusupov again gives
the verdict 'unclear'.
1 8...J..g5 1 9 Wh5
1 9 D.g3 allowsBlack some additional de­
fensive resources: 19 ./i�fS 20 1Wg6+ li)g7 21
. W'xhS 24 J:l'th5 f5 25 llxh7+! �7 26 li)f6+,
/&4 (not 21 h4? �f6 22 ll:)ds Wh8 and
Black refutes the attack) 21 ...Wh8 22 .ixf7
We? with a position that is very difficult to
assess accurately. For example, 23 ..xh7+
Wxh7 24 li)xg5+ 'irxg5 2S :xgs l:txf7 26
l:txcS :ld8 and Black's extra knight wiU have
an interesting battle against the three pawns.
1 9...1rf6 20 .1Z.g3
A strong alternative is 20 .rlh31? 1i'g7 21 f4
.L£4 22 ttxis .tgS 2.1 li)e7+ i..xe7 24 l:txe7
li)c6 25 1lee3 with a dangerous attack.
20. . ..1Z.ae8
Al'lo possible is 20 ll:)f5 21 llh3 1i'g7 22
...
f4 llaeB 2.1 lle2 when White's attack contin­
ues although there is no clear win. However,
after 22 lieS? (I! - Kupreichik) 22 li)f6 23
...
1Wf3 li)h4 24 'ti'g3 .tf4! Black successfully
defends.
21 l0e4
21 Wh8?
..•
Black cracks under the pressure. 21 ...q.,g7
is essential, when the fight continues. Follow­
ing 22 �xf6 (Black has counterplay after 22
h4 'ti'g(i 23 11fxg6+ fxg6 24 hxgS llf4 - Yu­
supov) 22. .:Xet+ 23 .tft �f6 24 f41 .txf4
. g3 .ie7 1 1 f5 .idS 12 0-0 0-0 13 �3 .ic6
2S llf3 �xf3+ 26 ..xf3 the position is still
unclear.
76
22 h41
By freeing his back rank, White threatens
the queen. Now White has a large advantage.
22 .1Z.xe4
••.
If 22...1i'h6 Kupreichik suggests 23 llxgS!
when White is much better.
23 .IZ.xe4 J..f4 24 Jlg4 J..h6
Or 24. ..td6 25 :Xd4! cxd4 26 .i.d
. . 3 and
Black must resign.
25 1rxc5 �6 26 Wh5 1rd& 27 ,j,xn
1t'd1 + 28 *h2 •d6+ 29 f4 •c7 30 J..b3
1rd6 31 .IZ.e8 hf4+ 32 g3 .txg3+ 33
�h3 1 .()
Mate is inevitable.
Game 32
Tukmakov-Bronstein
Moscow 1971
1 e4 e5 2 l0f3 tnt& 3 tnxe5 d6 4 l0f3
tnxe4 5 d4 J..e7
Of course S...dS is the more popular and
reliable move.
6 J..d3 l0f6
Trying to equalise by seeking exchanges
with 6 ...�g5 just falls short: 7 ��5 i.. xgS 8
1i'c2+ Wffi (or 8....te7 9 0-0 0-11 l O �c3 /0c6
1 l �5 lieS 1 2 .ie3 and White also has a
slight edge, while after 8. .i..e6 9 f4 i.. h4+ t O
.
1 4 dS .id7 1 5 llk4 White has a dangerous
initiative) 9 0-0 .i.xct 1 0 llxct �c6 1 1 c3 g6
 3 l:iJxe5: De viations from th e Main L ine
12 �2 '*g7 13 �1 .i.d7 14 f4 �8 15 'iff3
11fh4 16 g3 'ifg4 17 11ff2 and White was a bit
better in Palac-Mun.1', Pula 2002
7 h3
White plays to dominate his opponent's
queen bishop. 7 0-0 is a sensible alternative,
t(,r example 7.....i.g4 8 �bd2 0-0 9 �1 c5 1 0
h3 ..i.hS 11 ll)ft llk6 and now:
a) Breaking the pin with 12 g4 is weaken­
ing: t 2 ... .i.g6 13 li)g3 d51? (White has a small
advantage after 13...l:le8 14 c3 .c7 15 li)5
i.f8 16 l:lxe8 l:lxc8 17 .i.f4 cxd4 18 li)3xd4
a6 19 .a4 �5 20 .i.g3 li)b6 21 .c2, Ste­
f.'tnsson-Kholmov, Pardubice 2001) 14 dxc5
.i.xc5 15 .i.e3 .i.xe3 16 l:lxe3 d4 17 J:lc1
� with complicated play.
b) 12 li)g3 is better: 1 2....i.xf3 (or
12....i.g6 13 .i.xg6 hxh>6 1 4 dS li)b4 15 c4
and White has a pleasant space advantage) 13
11fxf3 li)xd4 14 11fxb7 �8 15 ll)f5 li)x5 16
.i.x£5 .i.fB 17 l:lxe8 'ifxe8 18 .i.d2 J!ll 1 9
.i.d3 d5 20 c4 and White has a strong initia­
tive, Ljubojevic-Smyslov, London 1984.
7 0-0 8 0-0 lieS
...
Black has alternatives but no path to
l.'tJUaJity:
a) 8...c5 9 llk3 li)c6 10 J:le1 a6 (to.. lbh4
.
doesn't solve Black's problems: White has an
edge following 11 .i.f4 d5 12 dxcS Lc5 13
.i.eS) 1 1 d5 li)a7 (after 11...li)b4 Karpov
suggt.-sts 12 .i.e4!? as a way to keep the initia­
tive) 12 a4 .i.d7 13 aS l:lc8 14 .i.f1 h6 15
.i.f4 .i.f8 16 J:lxe8 •xeS 17 .i.h2 •ds 1 8
li)d2 'flc7 19 �c4 li)xc4 20 li)xe4 and
White was a bit better in Karpov-Smyslov,
Moscow 1972.
b) 8...llk6 9 c3 J:le8 10 J:le I .i.d7 11
lLibd2 .i.ffi 12 llk4 d5 (exchanging with
1 2...li)xe4 still leaves White with an edge
:1fter 13 .i.xc4 h6 14 .c2 'flf6 1 5 .i.e3) 13
ltig3 .i.d6 14 l:lxe8+ 'iVxeS 1 5 'ifc2 h6 16
.i.d2 .f8 17 ltif5 .i.xf5 18 ..i.xf5 l:le8 19
i.d3 li)dg 20 c4 and White had an initiative
in Bnmstein-SmysJov, Leningrad 1971.
9 c4 c6
The alternative is 9..,li)bd7 10 li)c3 and:
a) to. .ll)fB 11 :C1 ll)g6 (White has a clear
.
plus after 11...d5 1 2 �5 c6 13 cxdS li)xdS
14 li)xd5 cxdS 15 .i.b5) 12 11fc2 .i.d7 13
.i.g5 .i.c6 14 d5 i.d7 15 l:lad1 a6 16 �2
li)hS 17 .tel •cs 18 �d4 cS 19 dxc6 bxc6
20 li)5 .i.x£5 21 .i.x5 11fc7 22 1Wa4 li)f6 23
l2XI4 c5 24 li)c6 and White was much better
in Luther-Bellin, Catalan Bay 2003.
b) 10. c6 11 J:le1 ll)f8 12 .i.f4 a6 13
..
'flb3?! (this allows cnunterplay; 13 d5!? li)g6
14 .i.h2 cS 15 'li'd2, with a slight edge, is
better, while also promising is 13 b41? �6
14 .i.h2 a5 15 b5) l 3H,�6 14 .i.h2 .i.f8
(Fischer assessed 14... b5!? 15 a4 bxc4 16
.i..xc4 d5 17 ..i.ft c5 as unclear) 1 5 :C2 b5 16
1ic2 ..i.b7 17 J:lae1 g6 with a complex posi­
tion, Fischer-Petrosian, 5th matchgame,
Buenos Aires 1971.
1 0 l:iJc3 e6 1 1 b4 l:iJbd7 1 2 .i.e3 t'iJf8 1 3
e4
1 3...85
White's initiative continues after 13 ...d5 14
cS li)e6 15 'ilc2 ..i.f8 16 bS.
14 b6 t'iJg6 1 5 lle1 t'iJh5 16 •d2 .i.d7 1 7
lZab1 ecs 1 8 .i.f1
18 dS!? cS 19 �4 with promising play is
also interesting.
1 8....i.f5 1 9 l:lb3
1 9 g4? wouW be a naive blunder:
19 ...i.xg41 20 hxg4 1ixg4+ and Black has a
.
clear plus,
1 9 h6 20 �2 t'iJf6 21 .-r,2 d5 22 bxc6
••.
bxc6 23 cxd5 .i.b4
77
The Petroff Defence
lf23...cxd5 then 24 l:c1 ..d8 25 �b5 and
White keeps an edge.
24 llc1 cxd5 25 tDb6 WbB+ 26 Wg1
lle7?!
This aUows White's advantage to grow.
Black can limit the damage with 26 ..ltc8 27
lDc1 .:Xct 28 1Wxct 'lrd8 29 l2X!3, when
White is only A bit better.
27 tDe5! i.d7
The tactical justification is seen after
27...I.Oxe5 28 dxe5 '1Vxe5 29 i.d4 'lre4 30
�d3 Wf4 3 1 i.xfS 'lrxfS 32 .J:f3 •e6 33
�xf6 gxf6 34 llg3+ �£8 35 1Wc2, when
Black's weakened kingside causes terrible
trouble.
28 tDd3 ..txb5 29 axb5 ..td& 30 �5!
After 30 b6 a4 31 llbS a3 32 •a2 l:b7 33
l:cbt White's advantage, if any, is' very small
indeed.
30...1rb6 31 li:la4 'iFdB 32 b6 llb8 33
�5 Ilea 34 11a1 tDh4 35 llxa5 tDt5 36
lla7
36 �2!? I.Oe4 37 I.Oxe4 l:xe4 38 �c3
with a clear advantage is a good alternative.
36...tDxe3 37 fxe3 i.g3 38 a4
78
This is a good move but White could also
cash in immediately with 38 I.Oa6 I.Oe4 39
I.Oxb8 'l'xb8 40 i.d3.
38 . . .'iFd6 39 e5?1
This allows Black to complicate. 39 I.Ob7?1
J:lxb7 40 .:Xb7 I.Oxc4 with unclear play is also
unconvincing, but the simple solution is to
prepare e4-e5 by supporting the knight with
39 'l'c3 - White wins easily after 39 ...i.h2+
40 �ht 1Wf4 41 e5.
39....be5 40dxe5 'iFxc5+ 41 Wh1?!
Now White's advantage disappears en­
tirely. 41 Wh2 I.Oe4 42 l:c7 'iraS 43 i.b5
maintains a slight edge.
41 ltla4 42 Wh2 1le6 43 1lc7 'iFa5
...
43..ltc6 44 .:Xc6 ...xc6 45 �d3 1Wg6 is
also good enough for equality.
44 ..i.d3
44 . . .llg6
The b-pawn is poisoned due to a back
rank trick: If 44...:Cxb6? then 45 l:xb6 l:xb6
46 'l'f2! wins, while after 44.. ltbxb6? 45
1i'f2! White has a clear advantage.
45 i.xe4 dxe4 46 •d4 e3 47 llxe3
'iFxb6 48 'iFxb6 llbxb6 %-%
 3 li:lxe5: De via tions from the Main L ine

Summary
Firsdy, the passive continuations covered in this chapter (5 ... .i.e7 and a quick ...l0d6 after
5... d5) are not of great interest. The key position of the chapter is the one arising after S...dS 6
.Ll3 .i.e7 7 0-0 .i.fS 8 %let lesc6 9 c4. In response to 9...0-0, White successfuUy develops his
knight to c3 either immediately or after 10 cxdS ..xdS. In the case of 9...lesb4 White maintains
110 opening initiative by retreating with 10 .i.fl, but 1 0 cxdS also appears to be a good response

- the tactical blow 10 ..lC!xf21? is not a refutation. Probably Black should transpose to the line
.

6.. .i.e7 7 0-0 �6 8 c4 l0b4 9 cxdS lC!xd3 10 ..xd3 'tixdS 1 1 :C t -*..fS (see Chapter 2) by
.

means of to lC!xd3 1 1 'tixd3 'il'xd5.

...

1 e4 e5 2 li:lf3 li:lf6 3 lbxe5 d6 4 �f3 .!Llxe4 5 d4 (D) d5

S....i.e7 Gm11e J2
-

6 .i.d3 (D) i.e7

6 .i.g4 Game J 1
... -

7 0-0 .i.f5 8 .:e1 lOc6 (D) 9 c4 Gmne 30

-

9 lesbd2 Ga111t 29
-

5 d4 6 .i.d3 B. . . lilc6

79
 CHAPTER FIVE I
3 ltJxe5: Fourth and Fifth
Move Alternatives

1 e4 e5 2 .!Llf3 �f6 3 �xe5 d6 4 �f3 idea. In the event of 5 ... �5 White retreats
.!Llxe4 his bishop to e2 and makes up for lost time
After 1 e4 eS 2 �0 �f6 3 �xeS d6 4 by attacking the knight with d2-d4. After
�0 �xe4 White sometimes declines to play S. . dS
. White continues the attack on the t..-4-­
5 d4 in favour of the other continuations, knight with 6 1i'e2, while if S...�f6 White
and these fifth move alternatives are the sub­ clears the way for the d-pawn by means of
ject of the fll'St part of this chapter. c2-c3 and .ld3-c2.
5 �c3 (Games 33-34) was played as fur 5 d3 (Game 39) is the most modest of
back as thc 1 9th century, while later on White's 5th move options. White offers
Nimzowitsch focused on it. Black the opportunity to transpose to the
5 1i'c2 (Games 35-36) was introduced by a Exchange Variation of the French Defence
1 3-ycar old Paul Morphy against LOwenthal after S.)t�f6 6 d4 dS where, as it is known,
(Nt:w Orleans, 1 850), and the line found an extra tempo in this symmetrical line
worthy adherents in the form of Emanuel promises White just a microscopic advan­
L'lsker and Jose Raul Capablanca. Owing to tage. However,. Black can equalise after
the success of such celebrities (sec, for ex­ 6....le7 foUowed up by ...c7-c5.
ample, Game 35) 5 'i6'e2 was considered to Moving onto 4th move alternatives, be­
be nearly a refutation of the Petroff Defence, sides 4 lDO White has two interesting op­
but now we know sevcral ways for Black to tions.
equalise. The fmt major game to witness 4 �c4
Kaufmann, a chess player from Vienna, was l hPaulscn-Schallopp, Frankfurt 1 887.
suggested 5 c4 (Game 37) as a way to pre­ The idea to transfer the knight to c3 is not
vent Black from supporting the e4-knight particularly impressive. In particular, Black's
with ... d6-d5. In principle it's still possible to attack on f2 after 4...lbe4 5 d4 dS 6 lDe3 1i'f6
play S ... dS, but an attack on Black's central- (Game 40) deserves serious consideration for
ist.-d forces with 6 �c3 gives White the better Black.
chances. Instead Black should continue his The Cochranc Gambit 4 lbxf7!? (Games
development with either s ...�c6 or s ...i.e7
. 41 -42), originated by John Cochrane in the
to avoid any problems. 1 840s, stands in rotal contrast to the other
5 .i.d3 (Game 38) is a comparatively new lines in the Petroff Defence. Staunton also

80
 3 li::J x e5: Fo urth and Fifth Mo ve A lterna tives
analys1..-d this knigh1-for-two-pawns sacrifice
in his Chess Praxi.r (1 860). Curiously enough,
both Cochrane and the famous maestro con­
centrated on 4...�f7 5 .i.c4+?1, which is
poor in view of 5...d5!. It nlight be appropri­
ate to rename the Cochrane Gambit to the
Bronstein Gambit (he was the one who
bn:athed nc.:w life into the sacrifice by indicat­
ing the possibility of 5 d4!) or the Vitolinsh
Gambit (the Latvian 1 M played a number of
brilliant attacks with 4 lllxf7). After the piece
sacrifice Black's king gets sruck in the centre
and White's initiative can be lasting and dan­
gerous.
Finally, Game 43 srudies a couple of rare
3rd move options for Black. Strictly speak­
ing, 3...l!lxe4 was the move srudied at tl1e
very beginning of the Petroff Defence the­
ory, Danliano considering it as far back as
1 51 2! After 4 We2 'ife7 Black gives up his
knight but immediately wins it back due to
the pin on the e-flle, and he is left a pawn
down with the hope of active play for his
pieces. Black's only other option is 3...'ife7,
but hereWhite has more than one way to
obtain a clear plus.
,...---
.- ----------..
Game JJ
Alekhine-A.Rabinovich
Moscow 1918
1 e4 e5 2 lL!f3 li::Jf& 3 truces d& 4 li::Jf3
li::Jxe4 5 li::Jc3
5 ...d5?!
11us is not a particularly good pawn sacri­
fice, but the alternatives covered here are also
not great. The best move, 5...l!lxc3, is con­
sidered in the next game.
a) 5...i.f5?? 6 lllxe4?? .ixe4 7 d3 i.g6 8
i.g5 i.e? 9 i.xc7 Wxe7+ 10 i.e2 lllc6 1 1
0-0 0 was a comedy of error.; that led to
C(jUal chances in Miles-Christianscn, San
Prancisco 1987. After the infinitely stronger 6
We21 Black immediately resigned in Zapata­
Anand, Bid 1987.
b) 5...lllf6 doesn't promise full equality: 6
d4 i.e7 7 i.d3 0-0 and now:
b1) 8 ltle2 with a further branch:
b1 1) 8 ...1le8 9 0-0 lllc6 10 c3 i.f8 1 1 llJg3
d5 12 Wc2 h6 13 lllh4 �h8 1 4 b4 l!lg8 1 5
l!lhf5 lllce7 16 l!le3 l!lf6 17 a4 l!leg8 1 8 b5
lllg4 19 lllxg4 (1 9 c4 aUows 19 ...:xe3! 20
fxc3 'irh4 21 h3 "irxg3 22 hxg4 i.d6 23 i.a3
-.nz+ 24 �f2 "irg3+ with a draw, Ulibin­
Akopian, Thilisi 1 989) 1 9...i.xg4 20 l!lf5 and
White has a bit of pressure.
b12) 8...c5 9 h3 lllc6 10 c3 l:te8 I 1 ()..() b6
12 'Wc2 i.b7 13 dxc5 bxc5 (probably
1 3...ltle5!? is stronger: 14 lllxe5 dxe5 1 5 :d1
'flc7 1 6 cxb6 'lt'c6 17 £3 Wxb6+ 1 8 �h1
:adS gives Black play for the pawn, while 1 5
cxb6 .ic5 16 �h2 'lt'xb6 also seems to give
BL1ck decent counterplay) 14 lllg51 g6
(1 4...h6? lead to a fabulous end in the follow­
ing game: 15 i.h7+ Wf8 16 lllxf7! �f7 17
'flg6+ �f8 18 i.xh6! gxh6 19 lll f4 '1t'd7 20
..xh6+ Wf7 21 ...g6+ �f8 22 l!lh5 :ebB 23
:act lllxh7 24 •xh7 l!le5 25 f4 i.£6 26
Wh6+ i.g7 27 fxe5+ �g8 28 lll f6+ 1-0
L.Dominguez-Ro.Perez, Cuba 2003) 15 'lt'b3
llle5 16 lllxf7! ..i.d5 17 l!lxd8 ..i.xbJ 18 axb3
lllxd3 1 9 l!lc6 ..i.d8 20 lll6f4 lllxct 21 lllxct
g5 22 l!lfd3 and White retains some advan­
tage (L.Dominguez).
b2) 8 h3 is also good, fur example 8...l!lc6
9 a] lieS 10 0-0 h6 1 1 llet ..i.f8 1 2 llxe8
1Wxe8 1 3 l!lb5 ...d8 14 c4 a6 1 5 lllc3 l!le7
(1 5 ...d5 16 c5 .i.e6 17 b4 also looks better for
White) 1 6 d5 i.f5 1 7 .te2 i.h7 18 b3 'ifd7
81
Th e Petroff De fen ce
1 9 .i.b2 and White's chances were preferable
in Acs-Haba, Bled 2002.
6 ..e2 �e7 7 lillle4 dxe4 8 •xe4 0-0 9
�c4
Black obtains good play after 9 .i.d3 g6 1 0
0-0 lbc:6 1 1 .tbS .tf5 1 2 •e2 :Cs 13 .txc6
bxc6 1 4 d4 l:b8, when the bishops and the
development compensate for the pawn defi­
cit.
9 ...�d6 10 0..()
White is aiming for rapid development.
Alekhine gave the following line: 1 0 d4 l:e8
1 1 liJcS .i.xeS 12 dxeS l'Dc6 1 3 .i.f4 �4 14
0-0-0 l:xeS 1S l:d8+! 'ifxd8 1 6 .i.xeS ..e7 17
l:ct �xeS 1 8 "ffxeS ..xeS 1 9 llxeS �fB,
when the draw is nt:ar.
1 0...lle8 1 1 •d3 1Llc6 1 2 b3!
White needs to get his dark-S(]Uarcd
bishop into play. After something like 12
..c3 .i.g4 1 3 .i.dS .i.xf3 14 .i.xf3 �d41
Black would have good play for the pawn.
1 2...•f6 13 �b211
The master of attack strikes again! Instead
of clinging onto his pawn, White launches a
strong counterattack.
1 3 ...•xb2 1 4 �5 .te6
Mter 14 ..g6 White has 1 S .i.x£7+ ..tg7 1 6
. 22 llxe6+1
.i.xe8 11'eS 1 7 ..c3 � 4 1 8 f4 and a very
nice position (Aickhine); 1 4...�8 1 S 1Wxh7+
�fB 1 6 llaet .i.e6 17 c3 also looks very
promising for White.
1 5 �xe& fxe6 1 6 •xh7+ � 1 7 llae1 ?1
This natural looking move allows Black to
82
bring the queen home. 17 c3! is stronger.
1 7...'ifxd2 1 8 'iih8+ �e7 1 9 1Wxg7+ �d8 20
l:adt 1Wf4 21 g3 ..f8 22 �xe6+ ami White
wins (Aiekhine); or 1 7....i.e5 1 8 Wh5 �g8 1 9
Wf7+ �hs 20 1Wg6! �gS 2 1 d4 .i.f6 22
'ifb7+ �f8 23 �xc6+ and White has a very
large plus.
1 7 ......f6 1 8 �5
1 8 'ifh8+?1 is weaker. after 18...�e7 1 9
l:xe6+ Wd7 20 l:xf6 l:txh8 2 1 l:£7+ �7 22
h3 g6 Black's chances in the endgame are
preferable.
1 8 ...�8 1 9 lle3 �f4?
Now I know we should not talk badly of
the dead, and Rabinovich was truly one of
the great fathers of Russian Chess, but here
his defence was not particularly good. White
is also very happy after 1 9 �4?! 20 llh3 g6
...
21 Wh7+ WfB 22 11i'd7 :e7 (22...�g8? 23
l:h8+1 would be an unpleasant surprise) 23
�h7+ �g7 24 l'Dxf6 llxd7 25 �xd7 l:d8 26
%ld3 l'De2+ 27 �h1 l:lxd7 28 g3 (AJekhine).
However, Black can play 1 9 ...�eSI 20 'ifb7+
WfB 21 f4 �£7 22 'ifhs �g8. probably keep­
ing the position level.
20 Wh7+ Wf8 21 Wh8+ h7
This is probably what Rabinovich missed.
22......xe6
Or 22.. .'�d7 23 l:txf6 :xh8 24 llxf4 with
a winning endgame.
23 •xg7+ �d6 24 �xe6 llxe6 25 d4
llee8 26 c4 ll8e7 27 •f8 lle4 28 •f51
 3 liJxe5: Fourth snd Fifth Move A lterna tives
llxd4
28..lt7e6 29 1Wc5+ �d7 30 dS also wins
for White.
29 c5 mate ( 1 -01
r-------.
Game 34
Khalifman-Atalik
Halkidiki 2002
._______________..
1 e4 e5 2 ctlf3 l0f6 3 �xeS d6 4 ctlf3
ctlxe4 5 �c3 lLlxc3
The best move.
6 dxc3
Also possible is 6 bxc3, when the game
could continue 6 ... .i.e7 7 d4 0-0 8 ..id3 .i.g4
9 0-0 lbc:l7 to lZ.b1 �b6 1 1 c4 .i.f6 (or
1 1 ...c5 1 2 h3 .i.hS 1 3 dxcS dxcS 14 a4 1Wc7
with an unclear position) 1 2 .i.e3 llb8 13 c3
'ii'd7 1 4 .l:lbS a6 1 5 llaS dS 16 cxdS lLlxdS 1 7
.i.d2 c6 1 8 h3 .i.e6 and Black is okay, Up­
ton-Dutreeuw, Batumi 1999.
6 -*.e7
.••
1 -*.f4
White has tried two other means of devel­
opment
a) 7 .i.d3 � 8 .i.e3 .i.g4 9 ..ie4 doesn't
I<K)k particularly appealing, even though this
worked well in the following example:
9 ...'ird7 t O 'ird2 0-0-0 1 t 0-0-0 llhe8 1 2
lLld4 d S (1 2. ...Ldt ? 1 3 ltlxc6 bxc6? 1 4 'ird3
and White wins) 13 lbxc6 Wxc6 1 4 .i.xdS
'iVa6 1 5 'ii'd3 .i.xd1 (this is necessary;
I S...'iraS 16 f3 c6 1 7 'irc4 .l:lxdS was played
in Nimzowitsch-Marshall, San Sebastian
1 9 1 1 , when 1 8 'trxg4+ f5 19 'trc4 looks like a
clear extra pawn to us) 16 'irxa6 bxa6 1 7
:lxd 1 and White has fine compensation for
the exchange.
b) 7 .i.e3 lbc6 8 'ird2 looks more sensible:
8 ... 0-0 9 0-0-0 ltles 10 lbd4 cS (to ...a6 1 1 f4
lbg4 1 2 .i.d3 cS 1 3 �fS lle8 14 1lhe1 .i.£8
1 5 h3 ltlxe3 16 ltlxe3 gives a position where
the control over the light squares favours
White) 1 1 lLlbS 'tWaS (or l l ....i.e6!? 1 2 lLlxd6
'irb6 13 b4 .i.xd6 1 4 'irxd6 1fxd6 1 5 llxd6
cxb4 16 cxb4 .i.xa2 17 .i.f4 f6 when Black
has equalised, Nunn-Z'Iu.Polgar, Brussels
1 985) 1 2 a3 a6 1 3 ltlxd6 lld8 1 4 ltlxc8!
llaxc8 (but not 14 ...:Xd2? 15 ltlxe7+ �f8 1 6
llxd2 �xe7 17 lldS ltld7 1 8 .i.e2! when
Black has too many problems: 1 8...b6 19 .i.f3
ltlf8 - 19 ...1la7 20 lle1 �fB 21 .i.f4 - 20
lieS+ 1-0 Van Der Wiei-Van Der Sterren,
Wijk aan Zee 1 984) 1 5 'ire2 bS 1 6 llxd8+
l:lxd8 1 7 .i.d2 ltlg6 1 8 �b1 b4 1 9 cxb4 cxb4
20 axb4 .i.xb4 21 .i.ct .i.e7 and Black has
fine com[>t:nsation for the pawn, Volokitin­
Kozakov, Lvov 2001 .
7 . . .li1d7
7. ..li1c6 also looks sensible, for example 8
'ifd2 .i.g4 9 .i.e2 'ird7 1 0 h3 .i.fS 1 1 0-0-0
0-0-0, avoiding the danger of opposite side
casding. Now after 1 2 .l:lhe 1 lZ.he8 1 3 g4 .i.g6
14 .i.bS a6 I S ..ta4 bS 16 .i.b3 ltlaS t 7 c4
�xc4 1 8 ..txc4 bxc4 19 WdS Wbs 20 WaS+
'ifb8 21 'ifxa6+ Wb7 22 Wxb7+ �b7 Black
83
The Petroff Defence
has equal chances in the ending. In this line
17...lDxb3+ is weaker: 1 8 axb3 1i'c6 19 lDd4
1i'b7 20 cxbS axbS 21 1i'a5 with an attack for
White, Boricsev-Raetsky, Fribourg 2000.
8 •d2
8 �5
•••
Black needs to get in ....tg4 to have a
chance of equalising. After 8...0-0 9 0-0-0
lDcs 1 0 h3 lieS 1 1 g4 White develops some
pressure: 1 t ...lDe4 12 ..e1 .i.f6 1 3 .i.e3 c6
14 .i.d3 'it'aS 15 .lxe4! llxe4 16 .i.b61 .i.g5+
(Black also suffers after 16 ...1i'a4 17 lld41
..xd4 1 8 .lxd4 llxet+ 19 :txe1 .i.e6 20
.lxf6 brxf6 21 lDd4) 1 7 lDxg5 .xgS+ 18 .i.e3
'if'g6 19 ..d2 dS 20 f3 :CS 21 h4 gave Black
serious problems in the form of an advanc­
ing attack in Tseshkovsky-I.Zaitsev, USSR
1975.
9 0-0-0 .tg4 10 .te2 0-0 1 1 h3 .th5
1 1 ....i.e6 12 lDd4 ..d7 13 b3 looks better
for White.
1 2 g4 .tg6 1 3 h4 J:leB
13 ....i.e4!? 14 hS lDe6 1 5 .ie3 c5, immedi­
ately establishing counterplay in the centre,
also looks good.
14 h5 .te4 1 5 l:hg1 ll:le& 1 6 .te3 c5 1 7
g5 d5 18 g&!
White has no reason to wait.
1 8...hxg6
Forced. After 1 R if6 19 gxh7+ �xh7 20
••
li)gS+ lDxgS 21 .lxgS d4 22 f3 .tf5 23 .i.d3
White has a raging attack.
1 9 hxg& fxg&
84
1 9....txg6? is bad on account of 20 lDeS
.tf5 21 .tbS :t£8 22 lDd7, when White has a
clear advantage. Now Black loses after
22...lle8 23 ..xd5, so he is forced to play
22 ... d4, which is hardly ideal.
20 ll:le5 .tf& 21 ll:lg4
21 ...d4?1
A natural push, but it's tactical suspect.
Black can improve with 21 .....a5!?, when
after 22 � 1 the push is stronger: 22...d4 23
.i.c4 bS (the difference; 23...dxe3? 24 lDxf6+
gxf6 25 ..xe3 still docs not work, though) 24
.ixe6+ llxe6 25 cxd4 'ifxd2 26 :txd2 cxd4
27 .txd4 :dB 28 lDxf6+ gxf6 with a likely
draw.
22 .tc4
22 lDxf6+ is also good for White:
22...1i'xf6 23 cxd4 cxd4 (23...lted8? 24 .i.c4
bS 25 .tgS! and White wins) 24 .txd4 lDxd4
25 ..xd4 .xd4 26 :txd4 llad8 27 llg..H
:txd4 28 llxd4 .tf5 with a slight pull in the
endgame (Atalik).
22 ...dxc37
A very risky move, even if there were no
forced win. 22...dxe3? 23 1i'xe3 .tgs 24 f4
would also not work for Black, but after
22 ... b5 23 .i.xc6+ :Xe6 24 cxd4 •ds 25
lDxf6+ brxf6 26 1i'a5 his position is still de­
fendable.
23 1te21 .b&
Or 23 ...cxb2+ 24 Wb1 1i'c7 (24...1i'c7 25
.i.xcS! wins) 25 lDxf6+ gxf6 26 1Wg4 'ireS 27
:tdS! and White wins (Atalik).
 3 liJxe5: Fourth and Fifth Mo ve Alterna tives
24 lt!xf6+ gxf6 25 bxc3 •c7 26 'ifg4!
11Ve5
27 lld5!1 1bc3
White wins in all lines: 27 ... i.f5 28 'ifxf5
or 27....ixd5 28 1i"xg6+ Wffi 29 i.xc5+!
when Black will get mated.
28 'ifxe4 1fa1 + 29 �2 •xg1 30 lld7l
1 -0
r-----
Gatne 35
Em.Lasker-Marshall
St Petersbutg 1912
.________________.
1 e4 e5 2 lt!f3 etlf6 3 lt!xe5 d6 4 lt!f3
lbxe4 5 'ife2
This line is usually played with the hope of
a draw. However, many White players have
found that achieving the draw from an even
ending is not always so easy.
5 .. -•e7 6 d3
The score from this position in our data­
base is actually 51°/co for White and 49% for
Black, which suggests complete equality.
However, it also shows that people play on
from here...
6 ...etlf6
The most common move, but not the
1 1nly one. For example, 6...lL!c5 7 lL!c3 i.g4 8
.i.e3 (8 ll:lds Lf3!? 9 'ifxe7+ .ixc7 10
ltlxc7+ Wd7 1 1 .!Dxa8 .ic6 is not clear, but
should be fine for Black) 8...c6 9 h3 .i.h5 lO
hr4 .ig6 1 1 .tg2 lL!bd7 12 lL!d4 ltle6 13 f4
li:lxd4 1 4 .i.xd4 f5 1 S 0-0-0 'ii'xe2 16 ltlxe2
gave White the chance to irritate Black in
Hodgson-Barua, l .ondon 1 986.
7 .i.g5
7 �e6
..
For some reason 131ack insists on keeping
the queens on, which is not necessary be­
cause the endgame promises Black guod
chances. 7...1i"xe2+ is studied in the next
game, while 7.../i)bd7 8 /i)c3 1rxe2+ (8...h6?!
9 .ic3 /i)b6 1 0 0.0-0 .id7 1 1 1i"d2 U-0-0 1 2
llc1 is a tad uncomfortable for Black) 9
.i.xe2 transposes to 7...9xe2+ 8 .ixe2 /i)bd7
9 /i)c3 (see the notes to the next game).
8 ltlc3 ltlbd7
8 .../i)c6 is answered by the natural 9 /i)e4
0-0-0 10 /i)xf6 gxf6 1 I .ie3 d5 1 2 d4 .ibr4 13
0-0-0 when White's position is preferable.
8. .. h6 looks dubious after 9 .ixf6 9xf6 1 0
d4 i.e7 1 1 'Wb5+ /i)d7 1 2 .i.d3 (1 2 'iVxb7
0-0 1 3 'ifxc7 i!Db6 appears very risky for
White) 1 2...g5 1 3 h3 (or 13 ll:le4 'ti"g7 14 h3
with a slight edge - Keres) 13 ...0-0 1 4 'iVxb7
llab8 1 5 1re4 'iVg7 16 b3 (we think White is
better here) 1 6 ...cS? (1 6.../i)cS 17 9c3 .if6
1 8 0-0 llfe8 1 9 1rd2 g4 - Tanasch - with a
messy position is necessary) 17 0-0 cxd4 18
/i)JS i.d8 19 .ic4 .!DeS 20 ..xd4 ..xd4 21
/i)xd4 LdS 22 .ixdS .if6 23 J:ladt and
White enjoyed a large plus in Capablanca­
Marshall, St Petersburg 1 914.
9 0-0-0
9 d4 leads to equality after 9...h6 1 0 i.h4
g5 1 l i.g3 �IS 12 /i)xdS .ixdS 13 0-0-0
85
The Petroff Defence
'ifxe2 14 .i.xe2 .i.g7. More interesting is 9
�bS!? .!llb6 1 0 .Lf6 gxf6 1 1 g3 dS 1 2 0-0-0
.i.h6+ 13 �b 1 with an unbalanced position,
something that no one would have thought
possible just a few moves ago.
9 . . .h6 10 .th4 g5
1 0... 0-0-0 seems to be more accurate. Af­
ter 1 1 d4 gS 1 2 .ig3 �b6 1 3 1i'b5 a6 1 4 'it'aS
i.g7 the position is unclear.
1 1 .tg3 .!Dh5
This looks like a loss of time, but Black
does not have it easy. 1 l ..ig7 12 �4 0-0
1 3 h4 g4 1 4 �£5 ..d8 1 5 �xg7 �xg7 1 6 hS
is not a serious alternative.
1 2 d4 �xg3 13 hxg3 g4?!
This move deprives the bishop of the £5-
square, which proves to be a problem in the
later tactics. 1 3... 0-0-0 1 4 We3!? with a slight
edge for White was better.
1 4 �h4 d5 1 5 1rb5! 0-0-0
1 5...Wb4? loses to the pretty 1 6 �xdS!.
1 6 'fra5
White should avoid 1 6 .!llxdS? .ixdS 17
..xdS 'it'gS+ 1 8 ..xgS hxgS, winning a piece.
16 ••• a6 17 .txa6! bxa6
Or 17...Wb4 18 Wxb4 .ixb4 1 9 .i.d3
.ixc3 20 bxc3 when White is simply just a
pawn up (Kasparov).
1 8 •••6+ Wb8 1 9 �b5 o!Db6 20 lld3
20 1i'g5+
••.
After this we cannot find a defence for
Black. The last chance for was given by Kas­
parov in his series of books MY GTMI Prr:deces-
86
sors.20...�c4 is the only move, although after
21 l:tb3! 'ifgS+ 22 �b 1 (also strong is 22 f4!?
gxf3+ 2..� �b1 �2+ 24 'lt>at �xb3+ 25 cxb3
'ifxg3 26 .!l:\x£3 .id6 27 llct .i.f4 28 �eSI
.ixeS 29 Wa7+! �cS 30 dxeS �d7 31 �xc7
with a clear plus for White) 22...�2+ 23
�a1 �xb3+ 24 cxb3 .i.d6 25 Wa7+ �cS 26
�xd6+ l:txd6 (26...cxd6 27 f4! brxf3 28 �xf3
1fe3 29 l:te1 and Black can no longer prevent
the deadly l:tcl+) 27 'ifa8+ �d7 28 'ifxh8
White has every chance of winning even if
Black can put up some resistance.
21 c;h,1 .td6
Black also loses after 21...'ife7 22 l:lb3
l:d6 23 a4 :c6 24 aS, when White's attack is
conclusive.
22 1lb3
Or 22 l:tc3!? .!llc4 23 Wa7+ �cS 24
�xd6+ cxd6 25 �a1 ! .:r.de8 26 llb3 and
Black is busted.
22. . .1lhe8 23 a4! .tf5
24 flJ.7
Or 24 aS .i.xc2+ 25 �xc2 l:le2+ 26 q;,b1
and it is aU over.
24....td7 25 a5 1i'd2 26 axb6 lle1 +
27 �a2 c6 28 �b5 cxb5 29 1ta7+ 1 -0
GaHJe 36
Aronian-Akopian
Ohrid 2001
1 e4 e5 2 �f3 �f6 3 lbxe5 d6 4 lilf3
lbxe4 5 1i'e2 1i'e7 6 d3 lilf6 7 .tg5
 3 liJxe5: Fourth and Fifth Mo ve A lterna tives
'lrxe2+
This is definitely the solid choice.
8 .txe2
8. . ..te7
8...lbbd7 is equaUy good after 9 lbc3 h6
and now:
a) to .i.h4 gS 1 1 .i.g3 lbhS 12 lbd4 (12
�iS �d8 13 d4 lbb6 is totaUy level too)
1 2...lbxg3 1 .3 hxg3 �f6 1 4 0-0-0 .i.d7 15
.l:ldel 0-0-0 16 lbd1 cS 17 lbf3 .i.g7 1 8 �e3
.i.c6 with equal chances, ApiceUa-Nikcevic,
France 2000.
b) to .i.d2 br6 (1 0...lbb6 1 1 lObS!? i!LlbdS -
1 1 ...�d8!? - 12 c4 .i.d7 13 a4 c6 14 i!Llbd4
lbc7 1 S 0-0 .i.e7 1 6 b4 1ead to a white advan­
tnge in Spassky-Yusupov, Moscow 19B1) 1 1
0-0-0 (1 1 lObs �dB 1 2 c4 .i.g7 1 3 0-0 :CB
14 J:lfcl a6 I S lbc3 lbcS 16 b4 .!Lle6 17 d4 aS
18 bS gS is another example of Black obtain­
ing equal play) 1 1 ....i.g7 12 d4 .!Llb6 1 3 h3 0-0
1 4 .i.d3 .i.e6 tS J:lhcl J:lfcB 1 6 b3 a6 1 7 .!Lle4
lt)xc4 1 B .i.xe4 .i.dS 1 9 .i.xdS lt)xdS 20 c4
lt)f6 with complete equality, Westerinen­
Ractsky, Hafnarfjonlur 1 999.
9 liJc3
Or 9 c4 h6 10 .i.f4 lbc6 1 1 lbc3 .i.fS 1 2
0-0-0 0-0-0 1 3 J:lhc1 gS 14 .i.c3 �g4 1 S lbdS
�xe3 16 lt)xe7+ �xe7 1 7 fxe3 J:lde8 with a
level position, Spassky-Karpov, Hamburg
1982.
9 ...h6
Another option is 9...c6 and now:
a) 1 0 0-0-0 lba6 1 1 J:lhc1 lbc7 1 2 llk4
tt:'lxe4 1 3 dxe4 .i.xgS+ 14 tlligS We7 1 S f4
�e6 1 6 lt)tt3 lbcs 17 ll'lf2 (1 7 f5 J:lbB 1 B
J:ld4 b6 1 9 :Cd1 J:ld8 20 c4 aS also looks
leveQ 17 ...J:ldB 1 8 .i.f3 f6 19 h4 .i.e6 20 g4 aS
21 a3 h6 22 hS J:ld7 23 lt)h1 J:leB 24 lt)g3
�dB 1/z-'/z Spassky-Hort, Reykjavik 1 977.
b) 10 0-0 lt)a6 1 1 J:lfe1 lbc7 12 d4 dS 13
.i.d3 �6 14 .i.c3 0-0 1 5 J:lad1 .i.d6 t 6 .!i)eS
.!i)e8!? (White is better after the risky 1 6...cS
t 7 .!i)bs .i.bB 1 8 c3 a6 19 lt)a3 - Yusupov)
1 7 tt:'le2 f6 1 8 �f3 tt:'!Bc7 1 9 b3 .i.d7 with fuU
equality, Spassky-Yusupov, Toluka 1 982
10 ..th4
10 .i.f4 does not look very dangerous ei­
ther: what exactly is the bishop meant to be
doing here? Anyway, one game continued
to...lbc6 1 1 0-0-0 .i.e6 1 2 d4 .!i)b4 1 3 a3
.!i)bds 14 tt:'lxdS .!i)xdS 1 s .i.d2 0-0-0 1 6
J:lhe1 J:lhe8 1 7 h3 .!i)b6 1 8 b3 .i.dS 1 9 i:e3
.i.e4 20 c4 li)d7 21 g3 cS 22 b4 .i.£6 23 lbg1
cxb4 24 axb4 .!i)b6 25 �b2 .i.fS with
chances for both sides, McShane­
Mamedyarov, Lausanne 2004.
1 0...liJbd7
10 ...lbc6 is also fine, for example 1 1 0-0-0
.i.d7 12 h.1 0-0-0 1 3 d4 J:lde8 14 .i.c4 J:lhfB
l S .i.g3 .i.d8 16 dS .!Lle7 17 .i.h2 a6 1B a3
.!i)g6 19 J:lhc1 J:lxe1 20 J:lxcl J:lc8 21 J:lxe8
.!i)xe8 22 Wd2 lt)h4 23 lllih4 1/z-1/z Shott­
Anand, Wijk aan Zee 2000.
1 1 liJd4 lOb& 1 2 a4 a6 13 a5 liJbd5 1 4
o!Oxd5 liJxd5 1 5 .txe7 liJxe7 1 6 .tf3 c5
1 7 .!082 liJc6 1 8 .txc&+ bxc6
87
The Petroff Defen c e
Black has equalised.
1 9 0-0-0 Ag4 20 f3 .i.e& 21 d4 cxd4
21 ...c4 22 ltlf4 �d7 23 dS cxdS 24 ltlxdS
would give White a real reason to play for a
win, even though the disturbance of the bal­
ance is very minor indeed
22 llxd4 �7 23 llhd1 llhd8 24 l£!f4
llab8 25 b4 .llb5 26 tnd3 lldb8 27 c4
llxb417 28 .!i)xb4 llxb4 29 h4
29 llxd6 l:txc4+ 30 �b1 l:tb4+ also leads
to a draw.
29...g5 30 hxg5 hxg5 31 llxd6 llxc4+
32 Wb2 l:tc5 33 l:t6d3 .llb 5+ 34 ¢'c3
llxa5 35 Wd4 �6 36 �3+ We5 37
llh1 c5 38 llh5 WI& 39 l:th&+ Wg7 40
llh1 c4 41 l:tc3 l:ta2 42 llhc1 llxg2 43
l:ta1 g4 44 fxg4 llg3+ 45 Wd2 l:txg4 46
llxa6 WI& 47 lla5 llg6 %-%
r-----
Game 37
Areshchenko-Mista
Cappelle Ia Grande 2003
'--------------•
1 e4 e5 2 .!i)f3 .!i)f& 3 .!i)xe5 d6 4 lllf3
lDxe4 5 c4 .i.e7
Black has a couple of serious alternatives
to this very natural move.
S ...dS seems a bit risk')':. 6 lbc3 .i.e6
(6 ...ltlf6?! leads to an French Defence, Ex­
change Variation with the loss of a tempo!
White should be able to organise a real ad­
vantage here: 7 d4 i.e7 8 cxd5 ltlxd5 9 i.c4
ltlxc3 1 0 bxc3 0-0 1 1 0-0 lbd7 1 2 :et ltlb6
88
1 3 .i.d3 .i.c6 14 'W'c2 g6 1 5 a4 is one exam­
ple) 7 'lfc2!? ltlxc3 8 dxc3 dxc4 9 lbd4 .i.dS
1 0 .i.e3 lbc6 1 1 0-11-0 and White has good
compensation for the pawn in the shape of
rapid development.
However, s.. lbc6 seL1DS to be a strong al­
.
ternative. Now we have:
a) 6 lbc3 will most likely transpose to the
main game (see the note to Black's 7th
move).
b) 6 d4 dS 7 &3 .i.b4 8 1ib3?1 (8 .i.d2
0-0 with equality is preferable) 8...'1fe7! (an
improvement over 8....i.e6 9 cS!? ltlxcS 10
dxcS d4 1 1 .i.c4 dxc3 12 0-0 .i.xc4 1 3 1i'xc4
cxb2 14 i.xb2 0-0 1 5 a3 i.aS 1 6 .1Ld1 11t'e7
1 7 lidS l:tad8 1 8 .J:tgs g6 1 9 h4 when White
had good compensation for the pawn in
Velicka-Raetsky, CappeUe Ia Grande 1 996) 9
.i.e3 lbas 10 11t'c2 dxc4 1 1 i.e2 i.fS and
Black has emerged from the opening with an
advantage.
c) 6 i.e2 i.e7 7 0-0 0-0 8 d4 i.f6 9 dS
ltle7 10 ltla3 l:le8 1 1 ltlc2 with a further
branch:
ct) 1 t ...h6 1 2 l:le l a5 13 l:lb1 i.f5 1 4 i.c3
.i.h7 (an improvement over 14 ...1Wd7 1 5
ltlfd4 i.h7 1 6 .i.g4 'lidS 1 7 11t'e2 c6 1 8 dxc6
bxc6 1 9 o lbcs 20 J:lbdl 'il'b6 21 b3, which
is slighrJy preferable for White, Kholmov­
Raetsky, Vorone7.h 1 988) 1 5 ltlfd4 ltlfS 1 6
lbxfS .i.xfS 1 7 i.d3 1Wd7 and Black i s equal.
c2) Also strong is 1 t ...ltlf5 1 2 ltb1 h6 13
lle1 i.d7 1 4 i.d3 ltlcs 1 5 ..tfl (1 5 :XeH+
 3 lDxe5: Fo urth Bnd Fifth Mo ve A lternB tives
'1Vxe8 t6 lbe3 lbxd3 17 '1Vxd3 with an equal
position is better) 1 S...J:lxe1 1 6 'irxe1 •e8 17
'irdt ?l ..c4 18 b3 aS 19 .i.b2 .i.xb2 20 llxb2
l:le8 21 llket and here, with Black slightly
better, the players agreed a draw, Brodsky­
Ractsky, Cappclle Ia Grande 2000.
6 lDc3
6 d4 is also interesting, for example:
a) 6...0-0 7 .i.d3 lbgS!? leads to wild play
after 8 lbxg5 hgS 9 1ths l:e8+ t o �d1
Let 1 1 'irxh7+ �f8 1 2 1th8+ �e7 13
1!rh4+ �f8 1 4 1Wxd8 J:lxd8 l S Wxc1 dS and
we end up in an endgame where Black has
enough counterplay for the pawn, Slobodjan­
Van Der Sterren, Bundesliga 1 999.
b) 6...dS 7 cxdS [1 llk3 lbxc3 8 bxc3 0-0 9
cxdS 1WxdS transposes to 7 cxdS) 7...1Wxd5 8
llk3 lbxc3 9 bxc3 0-0 10 .i.d3 cS 1 1 0-0
cxd4 1 2 lle1 .tf6 1 3 J:lb1 lbc6 1 4 .i.e4 ..xa2
1 5 9d3 h6 1 6 cxd4 .tg41? (or 1 6...J:ld8 17
.i.a3 .i.e6 18 llb2 ..c4 with an unclear posi­
tion) 1 7 .i.a3 iLxO 1 8 gxf3 tLlxd4 1 9 .txf8
l:lxf8 20 llxb7 1faS 21 1ffl .i.eS 22 IIebl f5
23 .c4+ �h7 24 .i.dS 'ifd2 and Black had
good play for the exchange in Gajewsky­
Mista, Poland 2003.
6 lDxc3
.••
6 ...lbg5 is worse: 7 .i.e2 0-0 8 0-0 lbc6 9
d4 J:le8 to lbds lbxf3+ 1 1 .i.xf3 .i.f6 12
.i.e3 .t£5 1 3 b4 a6 1 4 a4 h6 1 S .d2 and
White was better in Nisipeanu-Motylev, Bu­
charest 2001.
7 dxc3
7 0-0
.•.
We consider ...lDd7-cS to be the most reli­
able way to develop, but there arc other op­
tions: 7 ...llk6 8 .td3 lbeS 9 lbxcS dxeS 1 0
'ifc2 .i.gS 1 1 0-0 .te6 12 J:let .i.xct 1 3
l:axct •gs 14 lle3 0-0-0 1 5 J:lce1 f6 1 6 b4
lld7 and Black has solved his problems.
Maroczy-Marshall, San Sebastian 1 91 1 .
8 -*.d3 l0d7 9 1ic2 g6
9...h6 seems to be less harmonious. After
10 .i.e3 .i.f6 11 0-0-0 (the unambitious 1 1
0-0 J:le8 1 2 l:ad 1 lbeS 13 ltlxe5 .ixeS 14
llfe 1 9h4 1S g3 'iVhs gives equal chances)
1 t ...lbc:S t 2 lbxeS .i.xeS 13 f4 .i.f6 1 4 Wf2
lle8 1 S h3 .i.d7 16 :he1 a6 1 7 J:le2 b5 1 8 c5
White was a bit better in Velimirovic­
Motylcv, Novi Sad 2000.
10 -*.e3
1 0 .i.h6 J:le8 1 1 0-0 looks less aggressive:
1 1 ...lbcs 1 2 l:fc1 .tg4 1 3 lDd4 .tf8 14 .i.e3
(14 hf8 :xet+ 1 S l:xe1 'irxf8 1 6 lbbS
'ird8 is simply levd) 14 ...tLlxd3 1 5 'irxd3
.tg7 16 h3 .i.d7 11 9d2 'irh4 1 8 .tgs 9hs
19 lbe2 h6 20 .te3 .ic6 21 lbf4 1ff5 22
.td4 a6 23 .ixg7 Wxg7 24 ltlds .ixdS 2S
cxdS and a draw was agreed in G.Kuzmin­
Munl7.ia, Cappelle Ia Grande 2003.
10 ...l0c5 1 1 0-0-0
1 1 .i.xcS dxcS 12 0-0-0 is interesting: fol­
lowing 12...i.d6 13 h4 .tg4 1 4 bS!? .ixhS 15
l:xhS!? gxhS 16 .i.xh7+ �g7 the position is
highly unclear.
1 1 ...lDxd3+ 1 2 •xd3 -*.f5 1 3 •d2 c5
1 3....i.e6!? 14 b3 aS with counterplay is
also Logical.
14 -*.g6 -*.xg5 1 5 lDxg5 1fa5 1 6 b3 b51?
Energetically played. White has an edge af­
ter l6 ...'1fa3+ t 7 1ib2 .xb2+ 18 �b2 llae8
19 l:d2.
1 7 cxb5 c4
The most precise. After l7 ...a6 1 8 b6
llab8 19 Wb2 :txb6 20 1Wf4 White enjoys a
slight plus because Black has no c..-asy way to
weaken White's Icing's position.
1 8 �2 lZab8 1 9 b4
19 1Wxd6?! is very risky. After cxb3 20
89
The Petro ff Defence
axb3 WxbS 21 b4 aS Black's attack has every
chance of being successful.
1 9 .Axb5
.•
1 9 ...'i'xb5 is another attractive option.
Following 20 WdS 1rxd5 (20...Wa4!?) 21
.ILdS l:lfe8 22 lld2 the endgame is probably
equal, though not necessarily drawn.
20 �f3
After 20 ..xd6 ..a4 21 c;tlat 9a3 22 l:lc1
l:lb6 23 ..e7 Da6 24 ..e2 �d3 25 'i'b2 ..a4
Black has good play for the pawn.
20 .:fbs 21 �4 Lb4+ ! 22 cxb4
•..
Lb4+ 23 ltlb3
The only move. If 23 ci>a 1 ? then
23...l:lb I +I wins.
23 '1rb51?
••.
Black can keep the game alive with
23 ..cxb3 24 a3 1i'e5+ 25 1i'c3, when it's still
. 2001 .
very complicated.
24 'lrxd6 cxb3 25 .:d4 Ld4 %-%
90
26 Wxd4 bxa2+ 27 �xa2 'lla6+ is equal.
Game 38
Naiditsch-Timman
Dortmund 2002
1 e4 e5 2 �f3 llJf& 3 ofu:e5 d6 4 �f3
�xe4 5 .i.d3
This somewhat anti-positional move has
found some popularity over the last few
years. However, it shouldn't seriously
threaten Black.
5 ltlf6
...
Black has also tried:
a) S...l&s 6 �e2 and now:
at) 6 .. ie7 7 0-0 0-0 8 d4 �e4 (8 ...ll:k6
seems inferior: 9 l:let l:le8 10 .i.d3 �d7 1 1
h3 �f6 1 2 c4 gives White a slight edge, and
now Black should play 12 ... d5 instead of
1 2... c5?! 1 3 dS �fB 1 4 llk..'"-\ lllg6 1 5 ..c2
�d7 1 6 �d2 a6 17 a4 after which White is
very happy, Safin-S.Ernst, Dicrcn 2002) 9 c4
c6 10 'irc2 dS 1 1 .id3 �c6 with a complex
struggle ahead.
a2) 6... d5 7 d4 �e4 8 IJ-0 ..ltd6 9 c4 c6 1 0
�3 (1 0 l:le1 0-0 1 1 .td3 transposes to 5 d4
dS 6 .id3 .id6 7 0-0 0-0 8 c4 c6 9 ltet)
1 0...0-0 11 W'b3 �xc3 1 2 bxc3 dxc4 1 3
.ixc4 h6 14 l:let �7 1 5 ..c2 �b6 1 6 .id3
cS was level in Constantini-Raetsky, Bid
b) S d S 6 'i'e2 'lle7 7 0-0 with another
.•.
branch:
 3 liJxe5: Fourth a nd Fifth Mo ve A lterna tives
b1) 7...ll\d6 8 1i'd1 1? (8 llc1 'ifxe2 9
llxe2+ .i.e7 10 lbc3 .i.e6 1 1 tl'kl4 Wd7 1 2
lLlxe6 fxe6 13 lL!b5 lL!xb5 1 4 .i.xb5+ c6 1 5
.i.d3 .i.f6 1 6 c3 c 5 lead to a drawish ending
in Lcko-Kramnik, linares 1 999) 8...'ifd8 (this
seems best; 8 ...g6 9 l:le1 .i.e6 10 .i.fl .i.g7 1 1
d4 0-0 1 2 i.f4 lbc6 1 3 c3 'ifd7 1 4 lDbd2
�-,rave White a slight plus in Morozevich­
Shirov, Astana 2001) 9 lDcl c6 10 llct+ .i.e7
I I 'l'e2 .i.g4 (1 1 ...lLld7 1 2 b41? lL!f8 13 b5
llk6, as suggested by Romero, looks a bit
better for White) 1 2 b3 tlkl7 1 3 i.b2 lbc5 1 4
lL!a4 i.xf3 1 5 'l'xO lDxd3 16 'l'xd3 Yz-Yz
Morozevich-Kramnik, Wijk aan Zce 2000.
Black has no problems at all in the final posi­
rion.
b2) 7 ...lbc5 is less reliable: 8 llc1 i.e6 9
i.bS+ c6 1 0 d4 cxb5 1 1 dxc5 lbc6
(1 1 ...'1'xc5?! 1 2 lLlg5 'l'c4 1 3 lDxe6 'l'xe2 1 4
llxe2 fxe6 1 5 lixe6+ � fl 16 l:le5 is awful for
Black) 1 2 .i.e3 a6 1 3 a4 b4 14 c3 a5 1 5 cxb4
axb4 1 6 li)bd2 and White enjoys a plus,
Morozevich-Shirov, Sarajevo 2000.
6 h3 �e7 7 0-0 0-0 8 c3 liJc6
Alternatively:
a) 8 ... d5 9 .ic2 lieS to d4 .id6 1 1 i.g5
lLlbd7 1 2 li)1x12 h6 (12... lLlfB t 3 1lc1 l:lxel+
1 4 1fxel h6 1 5 .i.h4 lDg6 1 6 .i.g3 i.xg3 17
fxg3 lL!fB 1 8 g4 wa s better for White in Shi­
ruv-Anand, Leon 2001 - Black's pieces can­
not really get into the game) 1 3 .ih4 li)b6 1 4
a4 a S and Black has nothing to fear.
b) 8...c5 9 i.c2 lbc6 to d4 i.e6 1 1 l:le1 h6
12 i.e3 l:le8 1 3 lL!bd2 i.fB 14 li)e4 lL!d5 1 5
dxc5 li)xe3 1 6 lixe3 dxc5 1 7 ..e2 Wc7 1 8
llc1 l:lad8 with complete equality, Moro­
zevich-Topalov, Dorttnund 2001 . 14 .i.b3
ll\d5 1 5 .i.xd5 .i.xd5 1 6 c4 .i.xf3 1 7 lLlxf3
1i'f6 provides more imbalance, but it's not
clear who if anyone gains from this.
9 �c2 liJe5
White's pos1bon looks preferable after
9 ... d5 1 0 d4 lL!e4 1 1 l:let f5 1 2 .i.f4.
10 d4 lLlxf3+ 1 1 'tlfxf3 c6 1 2 �f4
White gained nothing after 1 2 l:let .i.e6
1 3 llkl2 1fd7 14 li)f1 l:lae8 1 5 lDg3 .i.d8 1 6
i.g5 lL!dS 1 7 Wd3 f5 1 8 i.d2 g6 1 9 .i.h6
l:lfl 20 11ff3, Lal.ic-Nikcevic, Paris 2000. In
fact, the players agreed a draw in this posi­
tion.
1 2 �86
•..
Here we recommend that Black should
deviate with 1 2 ...1fb6!? 1 3 b3 d5 1 4 lLld2
i.e6 leading to a position with equal chances.
1 3 lLld2 'tlfd7 14 J:lfe1 J:lfe8 1 5 �h2
We suggest 1 5 l:le2, for example 1 5.. ..i.f5
1 6 i.xf5 Wxf5 17 l:lae1 .tf8 1 8 lDe4 lL!xe4
1 9 l:lxe4 and White's control of the e-tile
annoys Black; or 1 5...c5 1 6 dxc5 dxc5 1 7
l:laet and White has a slight plus.
1 5 g6 1 6 c4 d5 1 7 c5 liJh5
•.•
1 7 ...lL!e4 leads to trouble after 18 li)xe4
dxe4 19 :Xe4 .i.d5 20 i.b3!, when the attack
against fl leads to a position with an extra
pawn for White.
1 8 'irb3 f6 1 9 liJf1 �f8 20 liJe3 �h6
91
The Petroff Defence
Alternatively 20.. f5 21 ll:lft JLg7 22 'ifc3
. 38 b6 axb6 39 axb6 :C7! 40 c6 lle1+ 41
.i.h6 23 f3 iLf7 with even chances.
21 .!Dg4?!
lbis spells trouble. White should play 21
.td6 iLf4 22 hf4 ll:lxf4 23 ll:lft b6, when
nothing is decided yet.
21 .txg4
•••
21 ....i.g5 is weaker, as after 22 '1Vf3 f5 23
ll:leS '1Vg7 24 llc2 lle7 25 llael White's po.'ii­
tion looks a good deal preferable.
22 hxg4 'irxg4
22...ll:lf4 only leads to equality after 23
'1Vf3 ll:le6 24 .c3 ll:lf4.
23 1fxb7 .tf4! 24 'irxc6
24 g3 would be strongly met by 24...ll:lxg3!
25 fxg3 .i.xg3 26 .i.xg3 '1Vxg3+ 27 Wft Wf4+
with an advantage for Black.
24....txh2+
25 �1
After 25 �xh2 ll:lf4! White cannot defend
92
himself. The main line runs 26 l:lxe8+ llxe8
27 '1Vxe8+ Wg7 28 ..e7+ Wh6 29 .i.e4 dxe4
30 '1Vxe4 1Wh4+ 31 Wgt ll:lh3+!, winning.
25. .:ed8 26 .td1 'irf5
.
Also possible is 26 •xd4 27 Lh5 hrxh5
...
28 g3 h4 and Black should be winning.
27 .txh5 gxh5 28 'ire6+ 'irxe6 29 l:lxe6
� 30 :as .tt4?1
A much bcner choice is 30...l:ldb8 31 b3
lt4, when Black has every chance to win.
31 g3 .th6 32 b4 .lld7 33 b5 .llb8 34 a4
.tg7 35 .llb 1 f5 36 .llb4 f4 37 a5 fxg3 38
fxg3
'iPg2 gxf2 42 'iPxf2 llbe8 43 :a?+ Wg6 44
llxg7+ Wxg7 45 b7 llcl 46 bS'IV llxb8 47
llxb8 llxc6 48 llds llf6+ 49 Wg3 l:lf5 gives
Black winning chances. In this line 43 lla2
.i.. f6 is probably bad for White, although
these complications are really difficult to
aSSCS.'i.
38 •• -Ac7 39 l:ld6
39 b6 also leads to a draw after 39 ... axb6
40 axb6 1lc6 41 l:la7+ Wg6 42 l:td7 .i.xd4 43
l:lxd4 llxc5.
39 .txd4 40 .llxd4 .llxb5 41 ll4xd5 % -
••.
%
Game 39
Dolmatov-Raetsky
Podo/sk Voronezh 1992
1 e4 e5 2 .!Df3 .!Df6 3 .!Dxe5 d6 4 .!Df3
.!Dxe4 5 d3
lbis can lead to symmetrical play and is
rather harmless.
5 ....!Df6 6 d4 .te7
Or:
a) 6...d5 transposes to the Exchange
French, which is outside the scope of this
book
b) 6...g6 7 .i..d.3 1!i'e7+ (the most precise;
7 .. ..i.g7 8 0-0 0-0 9 h3 ll:lc6 to a3 ll:le7 1 1 c4
.i.f5 12 ll:lc3 :Cs 1 3 llct .i.xd3 14 'ifxd3
'1Vd7 15 i.g5 gave White a small plus in Ni-
kolenko-Donchcnko, Moscow 2000) 8 '1Ve2
 3 ltlxe5: Fourth and Fifth Mo ve Alternatives
1i'xc2+ 9 Wxe2 i.g7 10 ltk3 ltlc6 1 1 ltlb5
Wd8 and the endgame should of course end
in a draw.
7 ..i.d3 ..i.g4
7 ...0-0 allows 8 h3 when White may be
able to gain a slight pull: 8...c51? 9 0-0 ltk6 1 0
%let h6 t t ltlc3 i.c6 1 2 i.b5 lbd5 t3 ltlxd5
i.xd5 1 4 dxc5 (this is better than 1 4 c3 a6 1 5
i.xc6 i.xc6 1 6 dxc5 dxc5 17 1i'xd8 i.xd8 1 8
i.c3, Kveinys-Schandorff, Germany 2001 ;
here Black can play 1 R...i.xf3 1 9 gxf3 b6
with a slightly more comfortable position,
even though it is probably not enough for an
advantage) 14...dxc5 1 5 i.f4 i.f6 1 6 ltlc5
and as so often in the Petroff the position is
more or less level but Black is a bit passive.
8 0-0
8 ltlbd2 isn't intimidating: 8...0-0 9 ltln
cS!? 1 0 ltle3 i.h5 1 1 0-0 �c6 12 ltlf5 l:le8 1 3
Jxc5 (1 3 c3 cxd4 1 4 cxd4 i. ffi 1 5 i.g5 h6 1 6
i.c3 ltld5 17 ltlg3 i.g4 1 8 h3 i.e6 equalised
fully in Strzelecki-Raetsky, Katowice 1 990)
1 3...dxc5 14 l:le1 i.ffi 1 5 l:lxe8 1i'xe8 1 6 i.g5
ttlc5 and Black has enough countcrplay.
8 ...0-0 9 Ae1
9 i.f4 should be met by 9...ltld5 (9...ltlc6
I l l c3 l:le8 1 1 ltlbd2 and White is a bit better
- Keres) 1 0 i.d2 i.f6 1 1 h3 i.hs 1 2 c3 c6
wilh level chances.
9...c517 1 0 h3 ..i.h5 1 1 dxc5
This is not particularly ambitious. After 1 1
d5 ltlbd7 (li ...ltlxdS?? loses to 1 2 g4) 12
ltlc3 tOes 1 3 i.e2 i.xf3 1 4 i.xf3 ltle81? 1 5
i.e2 f5 1 6 f4 ltld7 1 7 i.d3 ltlc7 1 8 1i'f3 i.f6
19 ltld1 b5 20 c3 g6 21 i.d2 ltlb6 22 lbe3
'ii'd7 both sides have chances, Shtyrenkov­
Raersky, Voronezh 1989. lnstead of 1 2 ltk3,
1 2 g4 i.g6 1 3 i.xg6 hxg6 14 ltlc3 ltlh7!? 1 5
a4 ltlg5 is interesting - maybe White i s a bit
better here
1 1 ...dxc5 1 2 g4 ..i.g6 13 ..ixg6 hxg6 1 4
•xd8 ..ixd8 1 5 ..if4 ..ia5!
Accurate play. After 15 ...ltlbd7? 1 6 gS
ltlh5 17 i.d6 i.aS 18 l:le7! White has gained
a serious advantage.
1 6 c3
Or 16 lld1 ltla6 1 7 ltla3 i.e? 1 8 i.xc7
ltlxc7 1 9 lld2 l:lfd8 20 :lad1 llxd2 21 l:lxd2
(Panchenko-Tolstikh, Cheliabinsk 1993).
Now Black is no worse after 21 .-ltle6 22
lObS b6.
1 6...Ad8
1 6...ltlbd7?! is imprecise. After 1 7 ltla3
ltlds 1 8 i.d6 l:lfe8 1 9 ltk4 i.b6 20 l:lad 1 the
problems arc mounting up for Black.
1 7 lllbd2 lllc6 1 8 ltle4
White also has nothing after 1 8 lbc4 lLlds
1 9 ltlxaS (1 9 i.d6 i.xc3! 20 bxc3 b5 21
i.xc5 bxc4 and Black is no worse) 1 9...ltlxa5
20 i.g3 f6 21 l:lad 1 �f7.
1 8...lbxe4 1 9 Axe4 AdS
Black has neutralised whatever pull White
ever had.
20 Aae1 f& 21 h4 Aad8 22 �g2 �7
23 A1e2 A8d7 24 .tea .tb& 25 a3 Ad3
26 ltld2 %-%
93
Th e Petroff Defence
r------
Game 40
Smagin-Makarychev
Moscow 1987
1 e4 e5 2 li:)f3 ll:)f& 3 li:)xe5 d& 4 li:)c4
Fortunately for our understanding of the
game, the eccentric 4 tZki3?! has not lead
great practical results. 4 ...c!Dxe4 5 'ife2 1i'e7 6
b3 c!Dc6 7 .tb2 .i.f5 s lik3 0-0-0 9 o-o-o ds
10 f3 c!Dc5 11 'ifxe7 .i.xe7 12 c!Dxc5 .i.xc5 1 3
d4 .i.d6 14 Wb 1 c!Db4 1 5 .i.c3 a6 gave Black
a healthier position in Yandemirov-Raetsky,
Voronezh 1988.
4 li:)xe4 5 d4
•••
Or:
a) 5 'ife2 'ife7 6 c!De3 c!Df6 7 b3 c!Dc6 8
.i.b2 .J.e6 9 g3 0-0-0 10 .i.g2 d5 1 1 d4 (1 1
0-0 h5 12 h4 'ifd7 and Black has good coun­
terplay on the kingside) t t ...1Wd7 1 2 tZkl2 g6
1 3 h3 .i.h6 14 0-0-0 l:r.he8 with even chances,
Murey-Harikrishna, Pardubice 2002.
b) 5 c!Dc3 shouldn't lead to an advantage:
bl) 5 ...c!Dxc3 6 bxc3 (6 dxc3 is not very
dangerous: 6 ....J.e7 7 .J.f4 0-0 8 1i'd2 lDc6 9
0-0-0 .i.e6 1 0 .i.d3 .i.f6 and Black can't be
worse) 6...g6 (also fme is 6...lDd7 7 d4 d5 8
c!De3 �£6 9 .i.d3 .i.d6 1 0 0-0 0-0 1 1 1Vf3
Aes 1 2 l:r.b1 c6 1 3 c4 dxc4 14 .J.xc4 1Vc7 1 5
h3 b5 1 6 .i.d3 .i.e6 1 7 c4 bxc4 1 8 .i.xc4
llad8 and Black has equalised, Gallagher­
Delchev, Batumi 1999) 7 d4 .i.g7 8 .id3 0-0
9 0-0 tZkl7 10 f4 �b6 1 1 �b6 axb6 1 2 f5
lla51 and the inclusion of the queenside rook
into the struggle guarantees Black decent
chances in this complicated position. 1 1
c!De3?! i'> weaker after l l ...c!Da41 1 2 .i.d2 c5!
(12...lt)b2?! 1 3 9£3 �d3 14 cxd3 f5 and
White had a slight edge in Ovetchkin­
Motylev, Russia 2004) 13 dxc5 lDxc3 14 'iff3
dxc5, when we rhink Black has the better
chances.
b2) Even 5...lt)f6 is possible: 6 d4 .J.e7 7
.i.d3 c!Dc6 8 d5 lDe5 9 lDxe5 dxe5 10 0-0 0-0
1 1 llet lt)d7 12 .t£S lDb6 1 3 .i.xc8 llxc8 1 4
llxe5 .i.f6 1 5 lle3 l:r.e8 1 6 l:r.d3 .i.xc3 17
94
bxc3 l:r.e5 with equal chances, Khainallin-
Bezgodov, Cheliabinsk 2004. In rhis line
1 1 .. ic5 is not so good after 1 2 .J.g5 .J.g4 13
'ifd2 (1 3 .i.x£6?! .i.xf2+! 1 4 �fl .i.xd1 -
14...1i'xf6 1 5 'ifxg4 .i.xel+ 16 Wxe1 1ib6
with unclear play is also possible - 1 5 .J.xd8
.J.xe1 16 �e1 .J.xc2 17 .i.xc2 :Ud8 with
an endgame that is hard to assess) 1 3... .i.d6
and Black is only a bit worse, but still, why
choose to be so?
5 d5 6 li:)e3 'Itt&!
•..
lbis move seems to be the strongest -
Black develops quickly and annoys White in
the process. Alternatively:
a) After 6... c5?! White gains a plus with 7
.i.bS+ llk6 8 0-0 a6 (8....te7 was met by 9
c4! dxc4 10 d5 lDJ6 1 1 .i.a4 bS 12 dxc6 bxa4
13 llM15 with some advantage in Jansa­
Miralles, Paris 1 989) 9 .i.xc6+ bxc6 10 c4!
.i.e6 1 1 cxd5 cxd5 12 lDc3 lDf6 (1 2...lDxc3
13 bxc3 .i.d6 1 4 .ta3 leads to a rigid struc­
ture with better chances for White) 13 f4
cxd4 1 4 'ifxd4 l:r.c8 1 5 �a4 .i.e7 1 6 b3 0-0
17 .i.b2 'ifa5 18 l:r.ad1, Sanakocv-Raetsky,
Voronezh 1 987.
b) 6...c6 7 tZk12 .id6 should be playable
too, but Black was unlucky in the following
game: 8 .i.d3 f5 9 0-0 0-0 1 0 f4 (10 c4 f4 1 1
lDc2 lt)xd2 1 2 .txd2 n 1 3 g3 dxc4 14
.i.xc4+ Wh8 is complete!)' unclear, but
probably more unpleasant for White than for
Black) 10....i.e6 1 1 g3 c5 12 c4!? lDxd2 (a
reasonable alternative is 1 2...cxd4 13 lDxd5
 3 li)xe5: Fourth and Fifth Move A lternatives
llX5 14 �f3 lilid3 1 5 ..xd3 �c6 with a
complex position offering level chances) 1 3
.ixd2 dxc4 14 .ixc4 .ixc4 1 5 �xc4 .i.e7?1
(1 5...cxd4 was necessary, although 16 1Wf3
:t7 17 1i'd5 would have been a little uncom­
fortable) 16 1i'b3 1i'd5 (maybe this is also not
the best, but Black completely missed
White's next move) 17 �b6!! ..xb3 18 axb3
with a clear edge, Lukjanenko-Raetsky, Vo­
ronezh 2000.
7 .ibS+I?
A very ambitious, but not nece.'ISarily fool·
ish. piece sacrifice. 7 ..e2 is less adventurous;
following 7....i:e6 8 c3 llX6 9 �d2 0-0-0 1 0
g3 (10 �xd5 .i.xd5 1 1 �xc4 ..g6 t 2 lDg3
.id6 gives Black excellent compensation for
his pawn) 1 0.. 1Wg6 (an improvement over
. problems.
10 ... �xd2 1 1 .ixd2 h5 1 2 .i.g2 1Wg6 1 3 0-0-0
�7 14 .i.f3 with a slight edge for White,
Smagin-Makarychev, Moscow 1 990) 1 1 .i.g2
f5 12 �f3 .id6 13 0-0 1Wf7 Black has good
counterchances.
7 c6 8 0-0!?
•••
8 . ..cxb5
Accepting the sacrifice is not the only pos­
sibility:
a) 8 ... .id6 9 �xd51? 1Wg6 leads to a very
complex position, though 1 0 f3 �5 1 1 �f4
.ixf4 1 2 .i.xf4 1i'xb5 1 3 ..ell? (13 fxe4
1i'xb2 1 4 ..d2 1Wxat 1 5 c4 0-0 16 �3
1i'xft+ 17 Wxft is not that clear) 1 3...0-0 1 4
fxe4 1Vxb2 1 5 eno seems to favour White.
b) 8....ie6 is perfecdy possible. After 9 f3
ll\d6 10 .ia4 ll\d7 Black's chances are no
worse.
9 l0xd5 •dB 1 0 Jle1 1rxd5 1 1 &3 •dB
1 2 lhe4+
1 2 .ig5 is inferior. White's main idea is to
reach a murky position after 1 2...1ixg5 13
�xe4 1Vg6 14 �f6+ �d8 1 5 :C8+ �c7 16
ll\dS+ �d7 17 'lre2 �6 18 :C1 (here Black
is mated after 1 8...b6?? 19 lld8+!). Instead
Black can play 12 ... .ie7 1 3 .ixe7 �xe7 14
.D.xe4+ WfB when, according to Makarychev
and Smagin, the compensation is insufficient
and Black has slighdy better chances.
1 2....ie7 1 3 11re2
13 .tg5 llX6 14 •e2 f6 1 5 d5 �5
doesn't seem to cause Black the same kind of
1 3 ...�c6
An interesting alternative is 13....if51? 14
'it'xbS+ 'lrd7 1 5 l:te2 1ixb5 16 �xb5 �d7 17
�7!? Wxc7 18 llxe7+ �6 when the end­
game is very difficult to assess. However, it's
hard to believe that White should be worse
here.
1 4 .ig5 .i.f5
Black is also all right after 14....ie6, for
example 15 .i.xe7 �xe 7 16 d5 (otherwise the
initiative disappears) 1 6...lilid5 1 7 lldt 0-0
18 lilid5 Ld5 1 9 :C5! .i.c41 20 :Xd8 .i.xe2
21 l:txa8 l:txa8 22 l:txe2 with a drawn posi­
tion. On the other hand, 14 ... b4 1 5 Jl..xe7
�xe7 1 6 1Wb5+ Jl..d7 17 1Wxb4 promises
White plenty of compensation.
95
The Petroff D efen ce

15 .i.xe7 lnxe7 16 .:.e5 0-0 a) 5 .i.c4+?! was the original idea, but it is
1 6...�£8 1 7 'ifxb5 'ifd7 18 ltac:1 would considered more or less refuted these days:
ensure White an initiative. S ... dSI (this is stronger than 5...�e6 6 .i.xe6+
1 7 llxe7 % -% �e6 7 0-0 Wd7 8 d4 WeB 9 c4 /l}bd7 1 0
/l}c3 - White has compensation here as
Black has no easy way to coordinate his
forces - 10...Wb8 1 1 f4 a6 12 a4 Wa7 1 3 b4
lieS 1 4 .i.eJ with a continued attack, Coch­
rane-Mohishundcr, Calcutta 1 848) 6 �b3 (6
exd5 .i.d6 7 0-0 l:lfB 8 d4 �g8 and Black is
close to winning - Bisguier) 6...�e6 7 e5
/l}e4 8 d4 c5 9 'iVf3+ WeB and Black has a
large advantage, as stated in JrhadJ�IIIng as
far back in 1 8611
b) 5 �3 is more inventive: 5...c51?
(5...1We8 6 d4 transposes to 5 d4 '1Ve8 6 /l}c3)
6 .i.c4+ .i.e6 7 .i.xe6+ Wxe6 8 d4 Wfl 9
After 1 7...b4 1 8 /l}d1 ..xd4 1 9 c3 bxc3 20 dxcS �6 10 'iVe2 1Wd7 1 1 .i.c3 dxcS 1 2 f4
/l}xc3 .i.g4! (Makarychev/Smagin) the posi­ lle8 1 3 e5 lDg4 (1 3.../l}d4 14 1Wc4+ 'ifdi 1 5
tion is level. 'iVxe6+ llxe6 1 6 0-0-0 is unclear) 1 4 lldt
r-------. 'irES 1 5 0-0 h5 1 6 .i.ct /l}d4 17 'ifc4+ Wg6
Game 4 1 and the advantage was probably with Black
Vitolinsh-Anikaev in Topalov-Kramnik, unares 1 999.
Riga 1982 5 . . .g6
The most important alternative, 5...c5, can
1 e4 e5 2 lLlf3 lnf& 3 lnxe5 d& 4 lL!xf7 be seen in the next game. Other tries include:
The Cochrane Gambit is probably not en- a) 5.../l}xe4? is bad because of 6 'ifh5+ g6
tirely correct, but it is also not that easy to 7 'IVdS+ Wg7 8 ..xe4 �6 9 dS .i.fS 10 'ifa4
refute. Vitolinsh was a great expert on this /l}eS 1 1 .i.e2 .i.e7 1 2 0-0 when White's posi-
� � � ���
4 �xf7
.•. b) 5 ... �e7 6 /l}c3 l:te8 is a very natural
way to develop, although after 7 .i.c4+ .i.e6
8 .i.xe6+ Wxe6 9 f4 Wfl 10 e5 dxe5 1 1 fxe5
White has a genuine attack. Now Black
should play 1 t .../l}c6!, for example 12 0-0 (12
exf6?! .i.b4+ 13 Wf2 ..xf6+ does not work,
while the slow 1 2 �c3?! �g8 1 3 .i.f2 /l}d5
14 /l}e4 .i.h4 15 0-0 �xf2+ 1 6 llxf2 .J:lfB
gave Black a clear edge for Black in Petrik­
Kujovic, Slovakia 2000) 1 2...'ifxd4+ 1 3 '1Vxd4
/l}xd4 1 4 exf6 .i.xf6 1 5 �g5 /l}xc2 1 6 l:lad 1
�3 17 .i.xe3 l:lxe3 1 8 lld7+ lle7 1 9 ltxe7+
Wxe7 20 /l}dS+ Wfl 21 /l}xc7 ltd8 and
White has some practical problems holding
5 d4 this endgame.
Or. c) 5...'ife8, taking prophylactic measures

96
 3 IC.xe5: Fourth and Fifth Move A lterna tives
aWtinst .i.c4+, is a very recent idea: 6 liX:3 (6
.i.d3 is punished by 6... c5! 7 dxcS dSI 8 liX3
dxt.-4 9 ltlxe4 .i.f5 10 0 .ixcS 1 1 1i'e2 .ixe4
1 2 .i.xe4 1i'a4, when Black is clt.-arly better)
6 ... d5 (6 ... c5 also looks fine) 7 eS .i.b4
and now:
c 1) 8 .i.e2 ltle4 9 0-0 .i.xc3 1 0 bxc3 ltlxc3
(J O_.'tlc6 1 1 c4!? also gives White play) 1 1
.ihS+ g6 1 2 'IVO+ �g8 1 3 1i'xc3 gxhS 14
:1b1 and the position was a complete mess in
11crnandez Romero-Andres, Seville 1999.
c2) 8 11'0 �g8 9 .id3 ltle4!? 1 0 0-0 (this
is stronger than 1 0 .ixe4 dxe4 1 1 'lfxe4
'iVc6! with an advantage for Black - Re­
inderrnan) 1 0....i.xc3 1 1 bxc3 ltlc6 12 .ixe4
dxc4 1 3 1i'xe4 1i'g6 with a complex position,
Ucrgcz-Grimbcrg, Clichy 1 998.
6 lbc3 11t'e8
Alternatively:
a) 6 ...'j;g7 7 f4 We8 8 cS looks very dan­
�crous for Black, for example 8...dxe5 9 fxeS
.li.b4 10 .i.c4!? .i.xc3+ 1 1 bxc3 with the idea
l l ....i.g4? 12 0-011 .i.xd1 13 exf6+ �fB 1 4
.li.h6 mate!
b) 6....i.g7 7 .i.c4+ .i.e6 8 .i.xell+ �e6 9
f4 cllfl 1 0 e5 lieS 1 1 0-0 was played in Vito­
linsh-Anikaev, Frunze 1 979. Now after
l l ...ltlfd7 12 f5 dxeS 13 fxg6+ �g8 1 4 Wo
White would have a strong attack.
7 .i.c4 +
7 .td3!? is another possibility, for example
7 ... .i.g7 8 0-0 llfB 9 eS ltlg4 (the pawn­
w-:tbbing 9 ... dxe5? leads to a disaster after 10
dxeS 1i'xe5 1 1 Jle1 'IreS 1 2 .i.e3 1Wb4 1 3 a3
'1Vg4 14 0) 10 h3 ltlh6 1 1 exd6 �g8 (White
wins after 1 1 ...cxd6? t 2 llet '1Vc6 13 .ie4) 12
dxc7 (1 2 ltld5!? cxd6 1 3 ltlc7 1lc6 1 4 ltlxa8
b6 1 5 ltlxb6 axb6 1 6 c3 with a slight edge is
also pos!>ible) 12...ltlc6 1 3 d5 ltle5 (Vitolinsh­
Domuls, Riga 1983). Now after 1 4 .ixh6!?
.i.xh6 1 5 d6 Wf7 1 6 .i.e4 we're leaning to­
wards giving White a clear edge. This of
course is not all due to 7 .i.d3, but it does
confirm that it's a viable option for White.
7 . . .i.e& 8 d5 i.cB
Here we assess 8 ... .td71? 9 0-0 bS 1 0 .i.d3
.i.g7 as a bit better for Black.
9 0-0 i.g7 1 0 Ae1 llfB
1 1 e5
1 1 f4 regains the piece but then the attack
dissipates after 1 l ...�g8 1 2 eS dxeS 1 3 fxeS
ltlg4 14 d6+ �h8 1 5 e6 ltles 16 d7 (1 6 e7
ltlxc4 17 cxfB1i'+ 1i'xf8 18 dxc7 ltlc6 is very
unclear play but we do not think Black
should fear this) 1 6 ...ltlbxd7 1 7 exd7 Wxd7
1 8 1i'xd7 .i.xd7 with a drawish endgame,
Popov-Grodzensky, correspondence 1 983.
1 1 ...dxe5 1 2 d6+ i.e& 1 3 llxe5 .bc4
1 4 1lxe8 llxeB 1 5 dxc7 lila6
After 15 ...ltlc6 16 .i.gS llac8 17 1i'O
llxc7 18 W£4 li:)eS 19 li:)e4 llc6 20 l:ld1
White manages to generate threats.
1 6 .*.f4 llecB?I
1\ sad decision - Anikaev was probably
afraid of ghosts such as ltlb5-d6. It would
have been better to leave a rook on e8, e.g.
97
Th e Petroff Defence

l6...:ac8 17 1Wd4 bS 1 8 1Wxa7 I?Jxc7 19 29 lDf5 lDef6 30 lDe7 + �8 31 lDxcS

i.xc7 ltt:7 20 .VaS l:lcxc7 with an unclear :XeS 32 •d6+ �g7 33 .id4 .J:le8 34
t:ndgame. �1 g5 35 .ba7 �6 36 11fc7 ..te& 37
1 7 b3 11fxb7 .i.f5 38 .i.e3 g3 39 ¢>g1
I 7 1Vd4! is simple and strong. AftL-r After 39 'W'b6 gxf2 40 .i.xf2 .i.d3+ 41
17 ... bS? 1 B b3 �15 1 9 .i.eS Black's position �gl l:lc2 Black has exccllem attacking
coll.:Jpst:s. chances.
1 7 .i.e& 1 8 clflb5 �
••• 39 t;},g4 40a4 lDhf6 41 ••a o-1
•••

1 8...1?Jd5 is a weak move: 19 I?Jd6+ �g8

20 lbxc8 J:lxc8 21 .i.e5! LeS 22 ..e1 and Game 42
White has a huge advantage. Novozhilov-Raetsky
Comspondence 1983
1 e4 e5 2 �f3 lDf6 3 lDxe5 d6 4 lDxf7
¢>xf7 5 d4 c5
Time has proven this to be the most test­
ing response to the Cochrane Gambit.
6 dxc5
This is the most commonly played move,
but 6 .i.c4+ is worth considering: 6 ...d5 7
exd5 .i.d6 8 0-0 lieS 9 lL!c3 .i.g4 (maybe
9 ... cxd4 10 1txd4 i.eS 1 1 1Wd3 Wg8, when
we prefer Black - we think!) 1 0 f3 .i.f5 1 1
dxc5 .i.xcS+ 1 2 �h 1 .i.d6 I 3 lL!b5 'ili'g8 1 4
19 11ff3?! �5 a6 1 5 lbd4 .i.d7 1 6 llk6 .i.xe6 1 7 dxe6
After the stronger 19 llc1 �g8 20 1Wf3 'ifc7 with a typically messy position, Stcll­
lbaxc7 21 lbxc7 lbxc7 22 c4 Black's uncoor­ wagt:n-I.Smimov, Heraklio 2002.
dinated forces promise White some kind of
edge.
1 9 ....i.xa1 20 .i.e5+ �8 21 .ixa1
�xc7 22 �
Vitolinsh could have offered better resis­
tance with 22 1Wc31? lbxb5 23 1Vh8+ �f7 24
1Wxh7+ �ffi 25 1Vxg6. Ht:re Kaemer and
Petkevich give the line 25....i.f7 26 1Vh6+
�e7 27 1Wg5+ with a draw, but after the
stronger 25 ...lbbc7 26 .i.b2 lidS we still be­
lieve Black can play for a win.
22 .i.d5 23 11fg3 t;},g7 24 c4 il.n 25 h4
.•.

ltX:eS 26 •e5 h5
Also strong is 26...b6!? 27 h5 :c5 28 1Ve4 6 11fa5+
.••

.1Lc8 29 h6 lDf6 30 1Vf3 lbges with clear After 6...c!L!c6 7 i.c4+ we have two possi­
advantage for Black. bilities:
27 g4!? hxg4 28 h5 lDxh5 a) 7...d5 8 .i.xdS+ i.d) (not the only op-
28...gxh5?? 29 I?Jf5 would revt:rsc the tion; 8...1?Jxd5!? 9 cxd5 lL!b4 looks very
trend! strong for Black) 9 i.xt:6+ ci>xc6 10 'ife2

98
 3 l0xe5: Fourth a nd Fifth Move A lterna tives

1Wa5+ 1 1 ltlc3 .i.xcS 12 0-0 (12 1rc4+ 9ile7 was close to losing in Short-Shirov, Dubai
U 0-0 llad8 is unclear) 1 2...�f7 1 3 11'c4+ (rapid) 2002.
b2) 9 ... d5 is also interesting:
b21) 10 e5?! � 1 1 1i'g4+ �f7 12 'l'f5+
�e8 and we do not think there is enough
compensation, for example 13 �c3 llXI4 1 4
Wh3 ..d7 1 5 e6 'iltxe6 1 6 ..d3 �xc3 1 7
bxc3 ltle2+ 1 8 �h 1 �ct 1 9 l:axct � f7 20
llcct 'irc6 with a clear edge for Black, Vi[O­
linsh-Dautov, Minsk 1 988.
b22) 10 ltlc3!? d4 1 1 ltla4 (1 1 1re2?1 is
awarded a '!' by Forintos and Haag, but after
1 t ...dxc3! 1 2 Wc4+ �e7 n es 1Wd4! 1 4 cxf6+
*xf6 White has no compensation) 1 1 ...�f7
with chances for both sides.
and now:
at) 1 3...�g6?1 t4 ltld5 .i.d6 1 5 .i.f4 llad8
1 6 llad1 11'c5 1 7 11'b3 �xe4? (a mistake,
although Black was already in ttnuble) 1 8
.tcJ Was 1 9 •xb7 :Cs 20 b41? (20 1i'd7!
nlso win.<�) 20...�xb4 21 �f4+ .i.xf4 22
11fxc4+ 1Wf5 23 9xb4 and White was win­
ning. Vitolinsh-Raetsky, Naberczhnye Chelny
1 9R8.
a2) 1 3 ..'�f8 14 ltlclS ltleS 1 5 Wb3 ltlxc4
.

I 6 9xb7 lieS and nothing has been decided.

b) 7 ....i.e6 8 .i.xe6+ �xe6 9 0-0 with:

7 �c3 1txc5 8 .i.e3 ••5

8...•c7!? 9 ltlb5 WaS+ 1 0 c3 �6 is a
possible recommendation- White seems to
have difficulties generating an initiative.
9 .i.c4+ .i.e6 1 0 .i.xe6+ Wxe6

b1) 9...�f7 1 0 We2 1re8 1 1 llc1 dS

( I t ...Wt!6!?) 12 eS �c4 1 3 �c3 ltlxc3 1 4
hxc3?! (1 4 ..f3+ �g8 1 5 WxcJ Wc6 is
str onger but Black still looks pretty solid
,

here) 1 4.....e6 1 5 :b1 :bs 16 .i.c3 .i.e? 17

f4 :hfB 1 8 .i.d4 �g8 1 9 :n gS and White

99
The Petroff Defence
11 0-0
It 1i'e2 has also been played: t t....i.e7 1 2
g4 :CB (this and Black's next move are not
srricdy necessary) 13 0-0 l:lxc31? 14 bxc3 oltX6
with a very unclear position, Zelinsky­
Volchok, correspondence 1 988.
1 1 ....!tlc6 1 2 f4 AdS
1 2...J..e7 1 3 11t'e2 ..a6 t4 lbb5 rlif1 is also
pretty messy. If push came to shove, we
would probably choose to be Black.
1 3 g471
'This is too risky and is met by a strong
counter in the centre. 1 3 a3 Wf7 1 4 b4 Wc7
1 5 lDdS gives White the initiative, even
though the position is probably still okay for
Black. However, perhaps 13...'ifh5!? gives
Black the better chances.
1 3...d5
14 f5+?1
This eliminates White's chances to invade
100
with his queen. The stronger option is 14
g5!? and now:
a) 14...lbxe4 1 5 9g4+ Wf7 (15...�e71? 1 6
llae1 i s also unclear) 16 'iff5+ WeB 17lDxc4
dxe4 1 8 1txe4+ .i.e7 19 f5 and the position
continues to be very difficult to as.c;css.
b) 1 4...dxe4 is not so good: 15 'ife2lDd5
16 f5+ �f7 (16...We71 may well be an im­
provement) 1 7 'ifhS+ Wg8 1 8 f6 gxf6 19 llf5
(Bielak-Borys, Poland 1 992). Now after
19...lbce7 (what else?) 20 gxf6 lbxf5 21 f7+
Wg? 22 •g5+ c;P,c.n 23 'il'xf5+ �e8 24 'ife5+
Wd7 25 lbxd5 White has serious attacking
chances.
14 ...¢17 1 5 g5 J.c5 1 6 gxf6 .i.xe3+ 1 7
�1 d4
Stronger than 1 7... dxe4 1 8 fxg7 Wxg7 19
Wg4+ wm, when Black's king is extremely
insecure.
18 fxg7 llhgB!
18 ...Wxg7?1 is inferior. After 1 9 f6+ Wf7
20 llf5 1i'xf5 21 exf5 dxc3 22 ..g4 �xf6 23
bxc3 Black's superiority is not as clear as it
should be.
1 9 f6
'This leads nowhere. White can struggle o�
a while with 1 9 lbds llxd5 20 exdS ..xdS+;
21 'iff3 Wxf3+ 22 llxf3 llxg7, or t 91ih �
�g7 20 f6+ �h8. Although Black is dearl�
better, the g:1mc is not totally decided. j
19...dxc3 20 1i'e2 1i'c5 I
Also strong is 20...cxb2 21 llae1 'ifc5
:;j
Wxd ..xeJ 23 llxe3 lld61 24 eS lle6 25j
 3 lllx e5: Fourth and Fifth Mo ve Alternatives
tlb3 b6, when Black is a knight up for noth­
ing.
21 l:tf5 cxb2 22 l:taf1 � 23 1ih5+
'tte6 24 1ih3 Wi'xe4+ 25 l:t5f3+ Wd6 26
*g2 �d4 27 l:txe3 1Wxe3 28 1Wxb7 l:tbS
29 f7 1Wf3+ 0-1
White resigned because of 30l:lf3 btW' 31
�1l:lbt32Wg2l:lg7.
Game 43
Kholmov-Belousov
Gorky 1974
1 e4 e5 2 �3 �6 3 lllxe5 lll xe4?1
This gambit/beginner's move is simply
bad.
3...1Vc7?! is also poor: 4 d4 (a decent alter­
native is 4 l£lf3!? I£Jxe4 5 �e2 dS 6 0-0 �e6
1 .l:let •f6 8 d3 I£Jcs 9 I£Jgs �d6 to .i.o
when White has a clear advanta�} 4...d6
(after 4 ...1£lxe4 5 �3 d5 6 0-0 ..f6 7 �xe4
dxe4 8 1£Jc3 .i.fS 9 l:le1 White ends up a
pawn ahead) 5 1£Jf3 'irxe4+ (S...I£Jxe4 6 �e2
i.g4 7 0-0 dS 8lte1 1£Jc6 9 c3 0-0-0 tO I£Jg5
i.xe2 t 1 l:lxe2 also gives Black a serious
headache) 6 .i.e2 �e7 (or 6...�f5 7 c4 1i'c2 8
1i'xc2 .i.xc2 9 i£1a3 �e4 tO i£1bS I£Ja6 11
i.gS with an advantage for White - all
Black's pieces are passive) 7 0-0 0-0 8 c4l:le8
9 l£lc3 W'fS to .i.d3 'ti'hs t I h3 c6 12 l:let
lba6 13 .tgS (White has emerged from the
opening with a decent plus) 13...1£Jb4 14
.ixf6 gxf6 1S ll'le4 .i.e6 16 lbg3 WaS? 17
'ii'd2 'ifb6 181i'h6 1-0 'Iimofuev-Phoobalan,
Goa 2002.
4 1We2 1We7 5 1fxe4 d6 6 d4 dxe5 7 dxe5
7 1i'xeS?! is a weaker choice: 7.....xe5+ 8
dxe5 .i.fS 9 c3 (91£lc31? .txc2 10 lDdS �d7
I 1 .i.f4 l£lc6 12 .tbS l:ld8 13 llct is still
probably better for White) 9...1£Jd7 10 f4 (1 0
.tf4 0-0-0 11 1£Jd2 I£Jxe5 12 �xeS l:le8 and
Black is at least equal - Yusupov) t0...0-0-0
1 1 .i.e3 f61? t2 I£Jd2 (12 .i.xa 7 fxe5 13 fxe5
I£Jxe5 14 .i.d4 :C8 15 �d2 g6 gives Black
serious threats) 12... fxe5 13 fxe5 I£Jxe5 14
0-0-0 I£Jct3+ 15 .i.xd3 :Xd3 and Black's posi­
tion is preferable, Kos-Potapov, Ceske Bude­
jovice 19tJS.
7 ...lllc6
8 J.b&
We think this is a good decision. Alterna­
tively:
a) 8 .tf4 g5!? 9 .tbS (9 .i.g3?? fS! 1 o 'lre2
f4 and Black winsQ 9....i.d7 10 .i.d2 .tg7 1 1
I£Jc3 0-0-0 12 0-0-0 .i.xeS and Black has
certainly survived the opening.
b) 8 i£1dl? is interesting: 8...'ii'xe5 9
1Wxe5+ I£Jxc5 10 .i.f4 (10 ll'lbs .tb4+ 1 1 c3
.taS 12 .i.f4 f6 and BL1ck holds his own)
lO....i.d6 11 i.g3 .i.d7 12 IJ-0-0 0-0-0 13
ll'le4 .i.c6 t 4 1£Jxd6+ cxd6 15 f3 llhe8 1 6
l:ld41? rl;c7 17 a4 and White had the advan­
fabFC in the endgame, Vasiukov-Chckhov,
Moscow l97S.
8 J.d7 9 �c3 0-0-0
...
After 9...1i'b4?! 10 .i.c4 �£laS It .i.d3
101
 The Petroff Defence
Whi et is a pawn up for nothin g.
10 0-0
A sim pler way to play is to .i.f4 gS 1 1
i. g3 ( 1 1 e 6 fxe 6 1 2 i.e 3 i. g7 is unc el ar)
1 1 ...i. g7 1 20-00 - h5 1 3 h4! with ac el ar ad­
vantage.
10.. .lt)xe5
10...a6 1 1 i.c 4 lilie S 1 2 /lhls 11Ve 6 1 3
.i.gS 6 f 1 4 .i. f4 give s White good a ta t ckin g
c hance s, while 1 1 i. xc 6i. xc 6 12 Wf5+ '1Vd7
1 3 Wf4 should be be tter for White too.
11 .te3
1 1 .i.xd7+ �xd7 12 '1Va4 a6 13 �S We S
14 WaS also pro vide s White wi th an edge.
1 1 ....txb5
Or 1 1 ...a6 12 /lhls ( 12 .i.xa6 .i.c 6 13
Wf5+ .i.d7 14 .e4 i.c 6 with a re pe titi on­
Kh olm ov) 12...We 6 1 3 �f4 'lieS 14 .i.xd7+
�xd7 1S W£5 and White retain s the i nitia­
ti ve .
1 2lilllb5 &6?1
Missing ac hance . A re f r 12...a6 13 �4 g6
1 41l ef 1 i. g7 1 Sila d1 llhe8 B lac k would be
aU ri ght.
1 3lillla7+ !7
Th is et m ptin g sacrifice is not str ci tly nee-
102
es sary . Afte r 1 3 Wf4 <l.>b 8 1 4 lta dt White has
a b ig advanta ge. StiU , it is very diffic ult to
defe nd again st suc h a sac rifice . Th is, alo ng
wi th the ae sthe tic be auty of the idea,e nc our­
age d K holm ov take the r isk.
1 3...�7 14 1ta4 l0c6 15 •as+ �7
1 6 :ad 1 + We& 17 1txb7 1te6 1 8 Ade1
.te7
1 8..1lb
. 8 lose s to 19 .xc 7 llc 8 20 .g3
.g6 21 i.c S+ .i.e? 22 11 a3 1lg5 23 lle 3
f oU owe dby lif e 1.
1 9 .tc5 Ab8
19...1i d7? is me t wi thbye el gant 20 .xc 6!
11V xc 6 211l xe 7+ �f8 221l xc 7+, winnin g.
20 1ta6 ..d5?1
Th is os
l esby force. A fter 20.. .Wf6 21 11V a4
�f8 221lxe 7 �xe 7 23 :C 1 White hase very
c hance of winnin g,b ut the game is stiU on.
21 1ta3 1tg5 22 .:e3 Wf8 23 .txe7+
rtJxe7 24 f4 1tf6 25 lZ.fe1 :ea
26 :Xe71
F orcin g a winn inge ndgame .
26 ....xe7 27 :Xe7 Axe7 28 g4 f5 29
•c5 g6 30 gxf5 gxf5 31 1txf5+ Af7 32
1tc5+ �8 33 Wg2 h5 34 �3 Ah6 35
1tg5+ Wh7 36 a4 1-0
 3 l0xe 5 : Fourth and Fifth Mo ve Alterna tives

Summary
S ll:lc3 has been underestimated by theory, and Black should act carefully here. In Game 33
Alckhinc played fantastically to defeat Rabinovich, but the pawn sacrifice with 5...d5 is dubi­
ous. 5...ll:lf6 is quite passive, but Black appears to be on the verge of equalLo;ing here. The more
popular and sharper S...ll:lxc3 6 dxc3 �c7 normally results in opposite side castling and a fight
with mutual chances.
The easiest way for Black to cqualL<ie against 5 ...e2 is with S.....e7 6 d3ll:lf6 7 �g5 ..xc2+
8 J..xe2 ll:lbd7 and 9...h6, preparing the simple ...g7-g6 or the extended fianchetto (...g7-g5).
The continuation 5 c4 is more challenging than 5 d3, but Black's play in Game 37 demon­
Mrates that it's also fairly harmless. The modern 5 J..d3 is pretty safe for Black. However, the
pc 1sitions are quite untypical and have not been studied very much.
The sacrifice 3ll:lxc5 d6 4ll:lxf7 deserves consideration because it has been shown dmt it is
not easy for Black despite the extra piece. 5 d4 ..e8, in conjunction with either ... g7-g6 or ...d6-
d5, seems to be promising for Black, while 5...c5 and in particular 6 dxc5 ltk6 7 J.. c4 ..tc6 8
J..xe6+ Wxe6 has not lost its favourable reputation. Of course, the presence of the black king
in the centre puts higher demands on Black but, objectively speaking, Black's chances are pref­
erable in the Cochrane Gambit.
After 3...ll:lxe4 4 '1Ve2 '1Ve7 5 ..xc4 d6 6 d4 dxcS 7 dxe5llk6 White has better prospects.
lie has a pleasant choice between 8llk3 and 8 ..ib5.

1 e4 e5 2 liJf3 ttlf6 3 ttlxe5 d6

3 lDxc4- Game4J
...

4 ttlt3
4llk4 - Game 40
4ll:lxf7 Wxf7 5 d4 (D)
5... c5- Gatne42; 5... g6- Game41
6 ttlxe4 5 tDc3 (D)
.••

5 d3- Game J9; 5 .i.d3- Game 38; 5 c4- Game 37

5 ..e2 'fle7 6 d3ll:lf6 1 J..g5 (D)
7 ... i.e6- Game 35; 7.. ..xe2+- Gm11e36
.

5 ...ttlxc3 - Game34
5...d5- Ga111e 33

5 d4 5 ttlc3 7 �g 5

103
 CHAPTER SIX I
3 d4: The Main Line
1 e4 e5 2 �f3 �f& 3 d4 lL!xe4 4 ..td3 d5
5 �xe5 tLl d7
In this chapter we begin our study of 3 d4,
the Stcinir.t system. 1 e4 eS 2 .!bf3 .!bf6 3 d4
.!bxe4 4 ..ld3 dS 5 .!bxeS is one of the key
lines in the Petroff Defence. Having placed
his knight on eS, White obtains a symmetrical
position with an extrn tempo. In this chapter
we look at Black's main choice, S...lbd7
The authors have special feelings towards
the line with 5...lbd7. In 1990 in Voronezh a
small book (40 pages) was published in Rus­
sian under the title A Fashiolltlble Vllliafioll i11
the Petroff Difelice. It was our first serious joint
work...
In his first correspondence theme tour­
nament devoted to the Petroff Defence
(1982-1983), Alexander Raetsky faced the
foUowing problem: how to initiate a compli­
cated fight after 3 d4 .!bxe4 4 ..td3 dS 5
.!bxeS. At that time the main continuations
were S.....td6 and S.....te7. However, in both
lines Black only dreamed of 'beating off
White's attack and attaining a draw. But
Alexander had a burning desire for some­
thing 'sharp' and 'interesting'.
In search of the active weapon let's refer
to Bilguer. The omniscient f-ltmdbnch informs
us: 'after s. ..lbd7, which was played in the
Barry-Showalter match (1896), White "-asily
104
d'-"Vclops by means of 6 'ire2 'We7 7 .i.f4
foUowed by 0-0 and llet.' The Black Army
leader could hardly expect that many years
later his idea would rise like a phoenix from
its ashes.
Up to the 1980s, S...lbd7 had been a dead
branch on the spreading tree of the Petroff
Defence. However, the year 1980 proved to
be a crucial moment in the hi10tory of the
variation. In the Tilburg super-GM tourna­
ment, Larsen sacrificed a pawn for counter­
play and the bishop pair against Karpov, tl1e
world champion at the time, with 6 'ifc2
.!bxeS! 7 .i.xe4 dxe4 8 1Wxe4 .i.e6 9 'irxeS
'ifd7 (following the game Unzicker-Rogoff
from Amsterdam earlier in the year). The
Danish GM could hardly have hopt:d for
better publicity than he received after this
impressive win. Players started to employ
s...lbd7 at major tournaments and now iliis
is Black's main weapon against tile Steinitz
system - 3 d4. At present almost all the chess
elite, players such as Kramnik, Anand,
Morozevich, lvanchuk, Ponomariov and
Kasimdzhanov, arc advocates of 5...lbd7.
The main continuation is 6 .!bxd7 .i.xd7 7
00- (Games 44-55). Black has a wide choice
between the modest 7.....te7 (Game 55), tl1e
queen lunbrcs 7...11f'f6 (Game 54) or 7...-.,4
(Games 49-53), and the fa10hionable 7....i.d6

(l1ames 44-48).
At the end of the 20th century, the sharp­
rst line 7...9114 8 c4 0-0-0 9 c5 g5!? (Games
49-52) had been hotly debated for almost
two decades - amazing for such a quiet
upcning! Even the representatives of the
chess bt:au monde like Kasparov, Anand,
lvanchuk, Judit Polgar participated in these
interesting discussions. Despite the statistics
lx:ing in White's favour, we believe that there
nrc possibilities to improve Black's play in all
the main lines (10 .ie3, 1 0 f3, 10 li)c3).
At the lx.-ginning of the 21st century
7...i.d6!? is in the limelight. (It is surprising
how the theory is developing: in the above­
mentioned book of 1990 we gave only one
reference out of 78 to the 'third-rate' move
7....i.d6) After the pretty typical 8 c4 c6 9
cxd5 cxd5 White is at the crossroads: he can
play either 10 �3 li)xc3 11 bxc3 0-0 12
'i!Vhs (Games 44-45) or the immediate 10
1i'h5 (Games 46-47). In both cases Black has
reasonable play.
Nowadays Black is not afraid of the tacti­
cal move 6 1Ve2 (Game 56), intending to use
the pin along the e-ftle; in fact, Black even
looks forward to it. Extensive practice has
proven tltat the pawn sacrifice after
6...li)xe5!? 7 .i.xe4 dxe4 8 1i'xe4 .i.c6 9 1i'xc5
1i'd7 is correct; easy development, two bish­
ops in the open, and awJ..:ward position of the
white queen create enough compensation for
Black.
The practical continuation 6 li)c3 (Game
57) opposes the knight on e4. But once all
ti1ur knights have been exchanged (6...li)xc3
7 bxc3 li)xe5 8 dxe5) White is unlikely to
have an advantage, due to his broken pawn
structure. On the other hand, grabbing mate­
rial with 6..ib4 7 0-0 li)xc3 8 bxc3 .i.xc3 9
llb1 provides White with a dangerous initia­
tive. Also in Game 57 is the unimpressive 6
0-U, which was often played in the earliest
days of 5...�7. Black has easy play after
6...li)xe5 7 dxeS �5.
Finally, the unexpected sacrifice 6 li)xf7!?
3 d4: The Main L ine
(Game 58) is one of the numerous discover­
ies by Igor 7...aitsev, one of the most creative
grandmasters in tlte world. This novelty led
to a quick and amusing draw by perpetual
check in Zaitscv-Karpov, Leningrad 1966
after 6...1i'e7 7 li)xh8 �3 8 ..td2 li)xd1 9
:Cl li)xf2 10 .i.xh7 li)e4 11 J:lxe4 dxe4 12
.i.g6 ..td8 13 li)f7. Six years later, Gur­
gcnidze routed Bellin by employing 7 1i'e21?
..txm 8 'iVhs �f6 9 o-o 'fkf7 10 'ikh4 gs 1 1
.ixgS! li)xg5 12 f4 with a mating atmck. But
this is too good to be true, and Game 58
demonstrates Black's superior options.
Game44
Svidler-Ponomariov
Moscon' 2001
1 e4 e5 2 ll:lf3 ll:lf& 3 d4 o!Oxe4 4 �d3 d5
5 li:lxe5 ll:ld7 6 ltlxd7
Sixth move alternatives for White will be
considered in Games 56-58.
6 -*.xd7 7 0-0 �d&
...
This is the solid alternative to 7 ...9114,
which is discussed in Games 50-53. T .ess
common options for Black are studied in
Games 54-55.
8 c4
Bli)c3 features in Game: 48.
s c&
...
8...9114 transposes to 7 ...'ifh4 8 c4 .i.d6-
see the notes to Game 49, while tL0-0 9
cxdS %le8 10 li)c3 9114 11 g3 li)xc3 12 bxc3
105
The Petroff Defence
11'h3 13l%b1 favours White - Black has no
easy way to get his pawn back.
9 cxd5
Another option is 9 ltX3 and nuw:
a) 9...0-0 10 'iihs lC!f6 11 -..,4 dxc4 12
.i.xc4 :Cs 13 .i.gS h6 14 .i..xf6 1Wxf6 t 5
1Wxf6 gxf6 is ftnc for Black. After 16 llfel
.i.e6 17 lbc4 ll..c7 HI .L.c6 fxe6 19 lbcs
.ll.xc5 20 dxc5 l%adR 21l%ad1 Wfl the play­
ers agreed a draw in Sveshnikov­
Mikhalchishin, Kuibyshev 1986.
b) 9...lbxc3 10 bxc3 dxc4?! (Black could
still transpose to the main game with 10 ...0-0)
11 l%et+ ll..c7 12 .i.:u .ie6 13 .L.e7 1Vxe7
14 .ixc4 0-0 15 .lieS 1Wa3 16 .i.xe6 fxt.-6 17
11Ve1 and White had some pressure, Shott­
Hubner, Wijk aan Zce 1986.
9. .cxd5 10�3
.
The main alternative is 10 Whs - sec
Games 46-47.
1 O . .l2�xc3 1 1 bxc3 0-0 1 2 Wh5 f5
. b) 15. g6 16 Wf3 l%e4!? (this compelling
Ths is the most solid move. The pawn
sacrifice with 12..g6 is considered in the next
game.
1 3lZe1
Alternatively:
a) 13 11V f3 Wh8! is fine for Black, for ex­
ample 14 ll..f4 ..c7 15 ll..xd6 1Wxd6 16 i.c2
:CeS 17 :ret l%ac8 18 h4 g6 19 h5 Wg7 20
g3 ..c6 21l%xeS :XeS 22 Wg2 1Wd6 23l%h1
i.bS 24 i.b3 l%e4 with equal chances,
Galkin-Macicja, Istanbul 2003.
b) More interesting is 13 c41Wf6 (13...dxc4
106
14 i.xc4+�hS 15 i.gs •es t61Wh41ooks a
bit better for White) 14 l%bl!? 1Wxd4 15
llxb7 dxc4 16 .i.xc4+ 9xc4. Here the play­
ers agreed a draw in Velicka-Rabiega, Berlin,
1999. After 17l%xd7 1t'c6 White is forced to
play the adventurous 1S llxg7+1 Wxg7 19
i.b2+ i.eS (19...WgS 20 1Wg5+ Wfl 21 1Wf6+
WeB 22 We6+ W<.l8 23l%dt looks very dan­
gerous for Black) 20 i.xe5+ WgS 21 l%<11
when he has substantial compensation for
the exchange.
1 3.. :ec7 1 4 i.d2 :aea
1 5 -*.c2
Maybe White has a brighter future after 15
g3, for instance:
a) lS .. .:Xet+ 16 :Xet WaS 17 1r f3 'it'xa2
1Sll..f4 i.xf4 19l%e7l%f7 20l%xf7 Wal+ 21
Wg2 Wxfl 22 1Vxf4 when White has consid­
erable activity for his pawn, and Black needs
to defend precisely to survive.
..
idea crops up more than once) 17 :Cct bS
18 a4 a6 19 .ih6 :Cs 20 .i.xe4 dxe4
(20... fxe4?! 21 1Wf6 .i.fB 22 .i.xfB .:.XfB 23
1Wxa6 bxa4 24 c4 and there is no compensa­
tion for the exchange) 21 'lfe3 1Vc4 22 axbS
ll..xbS 23 i.f4 .ifB 24 i.eS and it looks like
White has a slight edge, Ponomariov­
Moro?.evich, Moscow (rapid) 2002.
1 5 ... g6 1 6 1rf3lZe4! 1 7 g3
17 ll..xe4? fxe4 1S 1We2 .i.xh2+ 19 �h1
l%5 (Golubev) gives Black a strong attack.
1 7 ...-*.b5 1 8 i.f4
 3 d4: The Main Line

1 H .i.xe4 fxe4 19 'iVg4 .i.d7 offers Black
compensation, while 18 .i.h6 l:lf7 1 9 l:lnb I
'iVc6 is equally unclear.
18 ....txf4 19 gxf41Vd6
19 ...'itxf4 20 .i.xe4 fxe4 21 'itxf4 l:lxf4 is
also fine for Black, but probably just a dmw.
20 .txe4
36 l:l.4b3 'iVxu4 37 Jle2 WgS is probably
better, though Black has all the chances.
36 ...Wg5 37 lle2 Wf4 38 llb31Vxd4+ 39
Wg2 1l'c4 40 llf2+
Or 40 llbe3 'iVct 4 1 Wf2 1Wht and Black
wins.
40 �g5 41 llf7 d4 42 h4+ Wxh4 43
•••

White no longer has a choice. Even llxh7+ Wxg4 441lg3+ Wf5 0-1
though objectively the position is probably
equal, White's game is harder to play. Game 45
20...fxe4 21 1Vg3 llxf4 22 :Sb1 .td7 23 Rublevsky-Vallejo
f3 Ohrid 2001
23 h3 b6 24 c4 'iff6 25 cxd5 l:lf5 26 Wb8+
c.tr>g7 27 llb3 'tWxd4 28 llbe3 (Baklan) is an 1 e4 e5 2 ll'lf3 ll'!f& 3 d4 ll'lxe4 4 .td3 d5
interesting possibility for White, who has 5 ll'lxe5 ll'ld7 6 ll'lxd7 .i.xd7 7 0-0 .td& 8
some countcrplay of his own. c4 c& 9 cxd5 cxd5 10 ll'lc3 lL!xc3 1 1
23 ... b6 bxc3 0-0 1 21Vh5 g6!7 13 Wxd51Vc7
A weaker option is 23...exf3?1 24 'it>f2 b6
25 lZ.e3 .i.f5 26 l:te5 Wa3 27 'iVxf4 'ilfxa2+,
when Black has not completely e'lualised.
24 lle31Vf6 25 Af17
25 l:lb2 'irfB 26 llf2 1Wh6 is equal accord­
ing to Golubev.
25 . . ..tg4!
This 'sacrifice' destroys White's position.
261Vxg4
26 h3 .i.xf3 27 �h2 llf5 28 a4 llgs 29
1i'h8+ Wg7 30 l:tf2 Wh6 is also discouraging
for White.
26...1lxg4+ 27 fxg4 1Vg5 28 llg3 b5 29
Af2 Wg7 30 Wg2 a5 31 llb2 b4 32 cxb4 141Vf3
axb4 33 h3 1Vc1 34 1lgb3 �h& 35 1lxb4 It seems that the text move is the most
1Vd1 36 Wf2?1 dangerous for Black, although it's not the

107
The Petroff Defence
only attractive option. In fact, White ha.., tried
a few other moves here:
a) 14 h3 does not appear to be dangerous
for Black: 14..ie6 1S Wf3 ...xc3 16 ll.bt
(after 16 .i.h6 Black plays 16..llfd8 17 .i.gs
ll.e8 18 .i.f6 .i.fS 19 .i.xfS ..xf3 20 gxf3 gxfS
with equal chances) 16.....xd4 (16....i.xa2 17
llxb7 ..xd4 18 .i.b2 ..f4 19 •xf4 .i.xf4 20
ll.a1 .l:lfd8 21 .i.ft .i.e6 22 ll.axa7 lL'lc8 is
level according to Belikov) 17 .i.e4 We5 18
g3 .i.xh3 19 .i.b2 'fle7 20 ll.fe11lae8 21 .i.c2
(or 21 bb7 1ixet+ 221lxe1l:l.xet+ 23 ..th2
.i.fS 24 •c3 :Cs 25 f4 ll.e2+ 26 .i.g2 .i.eS
27 fxeS .i.e4 28 e6 ll.xg2+ 29 Wh3 llxb2 30
exf7+ ll.xf7 31 ..xb2 with a draw- Belikov)
21.....gS 22 J..b3 (G.Guseinov-Belikov,
Alushta 2001). Now Bclikov gives the
following line as the most natural way to end
the game: 22...• 51 23 1t'c3 .i.cS 24 :XeS
..xbt+ 2s Wh2 W'fl 26 Lf7+1 Wxf7 27
:fS+! �e611 28 ll.e5+ Wf71 29 .l:lfS+! with
perpetual check - a very attractive variation.
b) 14 h4 .i.e6 1S 9f3 ..xc3 16 .i.h6
ll.fd8!? 17 .i.gS (17 9f6?! .i.fB 18 .i.xfB
..xd4! and Black is better) 17...:d7 18 .l:lfd1
.i.f8 (18.....xd4 19 .i.e4 •es 20 g3 ll.c8 also
seems line) 19 ll.act •as 20 .i.f6 a6 21 .i.e4
.i.d5 22 "lfg4 .i.e6 23 •f4 J..d5 24 ll.c5
Wxa2 25 .i.eS with compensation for the
pawn, Nedev-Urban, Elista 1998.
c) 14 .i.h6 ll.fe8
with a further branch:
c1) 15 llact .i.xh2+ 16 Wh1 and now
108
strongest is 16....i.eS! 17 f4 (17 dxc5? %lxc5
18 .i.f4l:l.xd5 19 .i.xc7 ll.xd3 favours Black
slighdy) 17....i.c6 18 ..c4 b5 19 Wb3 ..d71?
20 51? .i.g3 21 Wg1 with complex play,
LepeUetier-Marciano, Auxcrre 1996.
c2) 15 h4 .i.e6 16 •o •xc3 17 •f6 .i.f8
18 .i.xfB llxfB with a fmal split:
c21) 19 .i.e4 19....1:lad8 20 %lfd1 ..td5 21
ll.act 1fa3 22 hS ..d6 23 ..xd6 %lxd6 24
.i.xd5 llxd5 25 ll.c7 ll.xh5 26 d5 ILls 27
ll.xb7 llhxd5 with a draw in Sutovsky­
foridman, Medellin 1996.
c22) 19 llfdt WaS (19..llad8!? should be
fine; maybe White continues with 20 .i.e4, as
20 .i.xg6!? hxg6 21 hS leads to a draw after
21...gxh5 22 'lfg5+ �h7 23 ..xh5+ Wg7 24
1t'g5+ Wh7 25 ll.act �) 20 .i.e4 Wds 21
..xd8 ll.axd8 22 d5 (White should have a
puU here) 22....i.f5 23 .i.xfS gxfS 24 d6 ll.d7
zs Ads ll.rds 26 :tad1 Wg7 21 ll1d3 f6 2s
%lg3+ 1/z-1/z liviakov-Yusupov, Groningcn
1994.
14 . .ixh2+
..
14.....xc3 15 .i.h6 ll.fe8 16 .f6 J..fB 17
.i.xf8 %lxf8 18 .i.e4 is preferable for White as
Black cannor free himself easily. Black has
compensation for the exchange after
l5...•xd41? 16 J..xfB .i.c6 17 'ffh3 .l:lxf8 but
he still has much to prove.
1 5 Wh1 .*. d& 16 c4
16 .i.h6 Wd8!? 17 g3l:l.e8 18 %lfe1 :c8 is
less dangerous for Black.
1 6 .Afe8
••

Maybe 16.....d81? 17 g3 .i.c6 18 d5 .i.d7
19 .i.b2 .l:lc8 is better. Most players would
prefer White, but is there reaDy an advantage?
17 c5 .i.fB 1 8 .i.e4 .i.e& 19 d5 .*.b5 20
d6! 'irxc5 21 .*.d5 .*.g7 22 .i.a3 1rc3
After 22.....d4 23 .i.xf7+ Wh8 24 .i.xe8
.l:lxeB 25 .l:lad1 �4+ 26 Wgl .i.xfl 27 l:lxft
White's strong passed pawn promises him
w-cat prospects.
23 .i.xf7 + �hB 24 'irxc3 .i.xc3 2& .i.xeB
.i.xf1 26 llc1 llxeB 27 llxc3
This endgame is deeply unpleasant for
Black.
27 ...lldB 28 llc7 �gB 29 llxb7 .*.c4 30
lte7 a6
30....i.xa2? 31 d7 .i.f7 32 .i.b4 wins in an
instant.
31 .i.b4 .i.b5 32 �2 lld7 33 lle5 �7
34 �3 .i.c4 35 a3 llb7 36 �4 .i.e& 37
J:la5 ll.b6 38 ¢>e4 llc6 39 .i.c5 �eB 40
'it'e5 ¢>d7 41 �6 .i.f5 42 f3 h5 43 .i.b4
J:lc2 44 llxa6 llxg2 45 lla7+ ¢>c8 46
J:lc7+ ¢>b8 47 .*.c5 llg3 48 f4 llf3 49
�e5 lld3 50 a4 h4 51 aS h3 52 llh7
.i.cB 53 .i.e7 +
53 l:lh8 h2 54 a6 is a simple win.
53 ...�a8 54 .i.d4 lla3 55 .i.b6 lld3 56
�c5?
56 a6, intending 56 ...ixa6 57 d7, would
. (13 ..c4? .ixh2+! 14 Wxh2llkt6 gives Black
have won.
56 ...llc3 57 .i.b4 llc4 58 .i.d2 llc5+ 59
ot>f&?l
59 We4 .l:lhS 60 l:lxh5 gxh5 61 Wf3 would
3 d4: The Main L ine
stiU give real chances of winning.
59 ... llh5 60 d7 .i.xd7 61 llxd7 h2 62
lldB+ ¢>b7 63 a6+ ci>xa6 64 llaB+ ¢>b7
65 lla1 h 1 'ir 66llxh1 llxh1 %-%
GanJe46
Svidler-Yusupov
Bad Hombutg1998
1 e4 e5 2 tl:lf3 lbf6 3 d4 lbxe4 4 .*.d3 d5
5 lbxe5 lbd7 6 lbxd7 .bd7 7 0-0 .i.d6 8
c4 c6 9 cxd5 cxd5 10 'irh5
This alternative to 10 li:)c3 is perhaps
more dangerous for Black.
1 0...0-0
Black has alternatives, but they aU appear
to give White the edge.
a) 10...lbf6 11 l:lci+ �f8 (11....ie6? 12
l:lxe6+ Wf8 13 l:lxf6 and White wins) 12
�4 gives White a slight pull. Now a fantasy
line runs 12. � 13 lbc3 lbg4 14 .i.f4!?
..
.i.xf4 15 "ire?+ Wg8 16 ..xd7 .i.xh2+ 17
�h1 lDf6 18 lbxd5!lbxd7 19 lbxb6 lbxb6
20 Wxh2 and the endgame is wonderful for
White.
b) 10.....f6 was suggt:sted by Igor Zaitscv.
Now White has the following options:
b1) 11 1i'xd5 .i.c6 12 .ibS 0-0-0 13 �3
the advantage) t3 .Lb5 14 'li'xb5 'li'xd4
..
and it is White who has to prove C(JUality.
b2) 11 .i.e3 ..g6 12 'li'xd5 .ic6 looks like
enough compensation for the pawn to us.
109
The Petroff Defence
After 13 11t'a5 b6 14 'lra6 Black is of course
fine, but now the simplest solution is
14 ...llk51 15 .ixg6 hxg6 16 .if4!? (otherwise
it's perpetual check) 16...t'tlxa6 17 i.xd6
0-0-0 18 .tg3 ltxd4.
c) 11 ltlc3 (the most challenging)
t t ...11t'xd4 12 1Wxd5 ..xdS 13 lll xdS ltlcs
(after 13... 5 14 .lf4 .lxf4 15 t'tlxf4 0-0-0 16
ltfe1 .lc6 17 lladl g6 18 f3 1l!f6 Black was
close to being level in Konguvel-Barua, San­
gli 2000) 14 :C1+ llle6 15 .tgS 0-0 (Black is
not much worse after 15 ...h6 16 i.e3 .i.c6 17
lllf4 We7) 16 i.e7 .i.xe7 171llxe7+ Wh8 18
llad1 lL.dS 19 .le4 with a plus for White,
Svcshnikov-1. Zaitsev, Moscow 1991.
1 1-.xdS
11 W k-ads to a draw after 11...g6 12
'irxdS .i.c6 13 �3 �xc3 14 bxc3 Lh2+ 15
Wxh2 1Wh4+ 16 Wg1 .i.xg2 (lvanchuk).
11 -*.c& 1 2 �5 g6
•••
12.-1if6?1 13 llk3 ..xd4 14 .i.xe4 .t.xe4
15 Ad1 .td3 16 'tth3 :t'd8 17 ltxd3 1Wxd3
18 'lfxd3 .i.xh2+ 19 Wxh2 llxd3 20 i.e3
gives White good winning chances, while
13...g6 141th6 ltfe8 I S dS! (Yusupov) is also
good for White. However, 12. ..i.b4 is play­
able, 13 .te3 :CS 14 a3 g6 t 51th3 transpos­
ing to the main game.
1 3 trh3
This move reaches the most critical posi­
tion in the 7 ....td6 1ine.
t 31th6 Des 14 .i.e3 .tfB is an improve­
ment for Black: 15 1th3 .i.d7 (1 S...h5!? 16
1 10
..i.c4 �6 17llxt2ibxc4 18 �xc4 •c8 with
some compensation) 16 ..f3 ..i.c6 17 .i.c4
(here White should consider 17 i.xe4!? ll'te4
18 ..d1 .i.d6 19 lbc3 Wh4 20 h3 ltxe3 21
fxc3 .i.xg21 22 ..g4 1Wxg4 23 hxg4 .t.xft 24
Jlxfl with an edge - Y.Gonza.lez) 17...�6
18 .tdS .txdS 19 'ifxdS � 20 �2 'irxb2
21 llab1 11t'c2 22 llfct 1Wd3 23 ll!f3 ltac8
and a draw was agreed in De Ia Paz-Y. Gon­
zalez, Santa Clara 2004.
1 3...J.b4
Apart from 13...lbg5 (see the next game)
Black has a couple of enticing alternatives:
a) 13...ltc8?! is punished by 14 .txe41 Qess
strong is 14 ..i.M :Cs 15 li:kl2 .td7 16 1i'f3
1i'h4 17 g3 ibxd2 18 .i.xd2 'lrxd4 19 i.c3
.g4 20 .xg4?1 120 1Wf6 ltxc3 21 .xc3 .i.c6
22 f31 could give White a slight L-<ige; it is not
easy to tcU, but not too relevant either]
20....i.xg4 with a draw, lvanchuk-Yusupov,
Linares 1993) 14....i.xe4 15 llk3 lte8 (or
t5.....i.f5 16 •o Wh4 17 g3 •xd4 18 Ad1
•c5 19 .i.e3 1Wc6 20 �5 with some advan­
tage) 16 .i.e3 (also strong is 16 .i.g51? ltxc3
17 Wh4! llh3 18 .i.xd8 .t.xh2+ 19 Wh1
l:lxh4 20 .i.xh4 .td6 21 .i.g5 with good win­
ning chances) 16....tf5 17 1Wf3 .i.b8 18 ltfe1
bS 19 g4 (Iimman-Yusupov, Unarcs 1993)
and now following 19...b4 20 gxf5 bxc3 21
bxc3 ltxc3 22 ltact llxcl 23 ltxc1 White
simply has an extra pawn.
b) 13...1Wh6?! is probably not particularly
good: 14W 1Wxd4 15 .i.xe4 ..i.xe4 16 11h4

f5 17 Jlh6 llfe8?! (17...ltt7, with a slightly
worse position, is a better option) 18 %lad1
(IK tbb5!? might offt:t a slight edge) 18...'.c5
IIJ tbxe4 llxe4 20 1t'f6 .i.e5. Now the game
Jl;llac-Marciano, Cannes 1998 continued 21
'1Vc6+? �h8 22 'lrd7 and here Black is fine
after 22..Wc6 23 b31le8. However, 21 lld8+
l:lxJK 22 'irxd8+ Wt7 23 'IVJ7+ We7 24
'IVliS+ ..e6 25 'ifxb7+ We7 26 ..d5+ 'ire6 27
'IVnH would have given Palac a laq,re advan­
I;�Jt:.
14 ..te3 .lleB 1 5 a3 .tas
15...Wa5?? is punished by 16 axb4 11'xa1
17 b5 .i.d5 18 tbc3, when Black must resign.
16 .llc 1 .Ilea
16...h5?! is weaker: 17 tbc3 ..i.d7 18 11'0
i.g4 19 'irf4 .i.xc3 (Svidler-Anand, Dort­
mund 1998) and here White can play 20 bxc3
g5 21 'irxe41lxe4 22 i.xe4 f5 23 i.xb7 llb8
241lab 1 f4 25 i.d2 with an edge (Anand).
3 d4: The Main Line
17�
17 .i.xe4 i.xc4 18 tbc3 .i.xc3 19 llxc3
:Xc3 20 bxc3 WeB 21 'irxc8 :XeS 22 i.d2
i.d5 was nor particularly testing in Sad­
vakasov-Kasimdzhanov, Lausanne 1999.
1 7 ._.i.b6 18 tOOl lDxf21
•
This seems to be the best option. Yusu­
pov docs not believe in Black's position after
18...i.xd4?! 19 i.xe4 i.xe4 20 lldl i.xe3 21
'irxe3 Wb6 22 'irxb6 axb6 23 lld4, and we
follow our bruru all the way here.
1 9 .bf2 ..txd4 20 lld1
The strongest move. Aftt:t 20 et)d 1??
:Ct+ 21 i.ft .i.b5 22 tbcJ .i.xf2+ 23 �xf2
1f'f6+1 Black has a decisive attack, while 20
Whl?! ..i.xf2+ 21 �f2 ..d4+ 22 Wft .i.d7
23 11'f31le3 is also unappealing.
20 .i.xf2+ 21 Wxf2 1tb6+ 22 Wf1
.•.
22...Wxb2
22 .1le61? is another attractive possibility:
.
231rh4 l:lf6+ 24 'irxf6 (24 �e1 is met by the
convincing 24.....ta4! when White is in trou­
ble, e.g. 25lbxa4? 'irg1+ 26 �d2 'irxg2+ 27
i.e2 lld6+ with mate to follow) 24....i.xg2+
25 �xg2 'iVxf6 26 .i.e4 with a murky posi­
tion.
23 11rd2
Possible is 23 lbe2 llcd81? (23 ... .i.xg2+ 24
Wxg2 ltxe2+ 25 �h1 lie7 26 i.c4 favours
White) 24 %lab1 'iVe5 25 Wh3 'fff6+ 26 �g1
'1Ve5 when it's unclear whether Black is
worse here - for one thing ...llxd3 is a threat.
23._.11rxc3 %-%
111
The Petroff Defence
Game 47
Macieja-Rozentalis
Reno 1999
1 e4 e5 2 l0f3 �f6 3 d4 l0xe4 4 �d3 d5
5 �xe5 �7 6 �xd7 �xd7 7 0-0 .*.d& 8
c4 c6 9 cxd5 cxd5 10 11h5 0-0 1 1 •xd5
�c6 1 2 1ih5 g6 1 3 11h3 �51
Time has shown that this is Black's best
path, after which he seems to be able to
count on equality.
14 .ixg5
Alternatively:
a) 14 'li'h6 �e7 15 .i.e3ltk6 16ltlc3 �f6
17 llad1 �xd4 18 �xd4 �xd4 19 l&4
�xe4 20 .i.xe4 � 21 Dd2llad8 22llfd1
ltd7 is equal, Mikhalchi.<�hin-Pavasovic, Celje
2003. More critical is 18 .i.e4!? �xe3 19
11'xe3 'Wb6 20 .i.xc6 bxc6; this could be a
tiny advantage for White.
b) 14 1tg4 is White's main alternative. Af.
ter 14...� there are three options:
b1) 15 .i.e3 h5 16 'li'h3 .i.d7 17 •e .i.c6
(Black has equalised but now White goes
crazy) 18 'lf'd1? 1fh4 19 h3? (although 19 g3
1th3 20 f3 h4 is great for Black anyway)
t9. ..lng5 20 d5 �h3+! 21 gxh3 •xh3 22
lte1 .i.h2+ 2.1 Wh1 .i.xdS+ 24 f3 .i.f4+ 25
Wg1 �xf3 0-1 Lautier-Gelfand, fo'JDE
World Championship, Las Vegas 1999.
b2) 15 �3 h5 16 'ii'h3 �xd4 looks fine
for Black, for example 17 �4 �2+ 18
1 12
�xc2 �xe4 19 .i.h6 lle8 20 .*.b5 llt:S 2'1
.*.d3 .i.c6 22llae1 llxe1 23llxe1 'ilff6 with
equal chances, Blehm-Ribshrein, Pattas 1999.
b3) 15 �h6
with a further branch:
b31) 15....i.f4!? is an interesting option,
Black's point is that after 16 .i.xf8 'ilfxd4 17
1ih3lbg5 (17...11'xb2 18 �a3llxf8 19 .i.c4
l'l)gs 20 1Vg4 WeS is unclear) 18 Wh4 WxdJ
19 11'xf4 .i.xg2 20 Wxg2 Wh3+ 21 �gt
�f3+ 22 ...xf3 1fxf3 White has some prob·
lems. Instead Holzke-Schandorff, Bundesliga
2001 continued 16 .txf4 1i'xd4 17 c&:3
'lf'xf4 18 11'xf4 �xf4 1/a-1/a.
b32) 15...lle8 16 �3 with a final split
b321) 16...�d4 17 ltad1 .i.eS 18 f4 f5 19
..,3 (19 hf5?! Wb6 20 .i.xg6 �f5+ 21lltz
�xh6 22 'IVgS .tf61 23 '1Wxh6 hxg6 24 1Vxg6+
.tg7 and White has nowhere near enough for
the piece - Van dcr Sterren) 19....i.f6 20 .tg5
�xg5 21 .tc4+ Wh8 22 fxgS Wb6 23 �ht
'illxb2 24llxd4 lle3! zs ll)dt 11'xd4 26 �e3
lies 27 1ih6! 11'xc3 28 llxf5! with an equal
position, Nijboer-Van der Srem:n, Rotterdam
1999. The main point is 28 ..1fc1 + 29 lln.
.
'1Vxc4 30 JlfB+ llxf8 31 'ilfxf8+ ..g8 32
1Wf6+ ifg7 33 1l'd8+ with a draw.
b322) Black can also try 16....i.f4 17 .Lf4
11'xd4 18 .i.e4 f5 (or 18...�xf4 19 'ilfxf4 .i.xe4
20llfel .i.b1! 21 'IVct �f5 22 11'd1 1Vxd1 23
llexd1 LiB 24 f3 �g7 1/z-112 Bologan·
Onischuk, Biel 1999) 19 'iVd1 (19 .txf5 leads
to a draw after 19...�xf4 20 llad1 WeS 21 g3
 3 d4: The Main Lin e

··� 22 Wl'xf5 gxf5 23 g:xf4 Wf7- Kaspamv)

UI,�·JCd I 20 llfxd1 fxe4 21 .i.e3 Wf7 22 b4
hi\ 11 n4 and White has a bit of pressure, Kas­
f'II'UV·Pikcr, Wijk aan Zee 1999.
lll., ••xg& 1 5lbc3 Aae8

This sacrifice leads t o a draw. Or a t least it

Ulnck has a reasonable alternative in
should...
21 �xh2 AM+ 22 �g1 -.6 23 g3
23 f3 .e3+ 24 llf2 �6 25 �fl .c3 26
*gt is quite a cute draw.
23 Ah1 + 24 �g2 Ah2+ 25 Wf31?
•••

I � lfcK, after which 16 lad1 is probably

...
hu�1. J(, tiS!? has been played a few times, for
uurnplc 16 ....i.xd5 17 f4 •ds (17...•f6!? 18
f'tKIIS Wl'M+ 19 �h1 .xdS 20 lad1 .xa2
�1 (51? is very unclear- Van der Wiel) 18
f)KtiS .i.cS+ 19 'iPht 'ilxdS 20 llact (Shirov-
1\l•nmnik, 4th matchgame, Cazorla 1998) and
lll'rc Black equalises after 20...Ile3 21 i.c4
lcl4. The point is that following 22 .i.xf7+?!
*117 23 ll.f3?! Wxf7 24 llxe3 9xe3 25
1Kh7+�f6 Black should win.
11 lad1 Ae7 1 7 d5 -*.d7 18 � Axe4
1 I •xd7 AdS
Another plan is 19..1Wc7
. 20 11xe7 llxe7
21 g3 lidS with an ending that should be
,Jmwn, l.uther-Delchev, Nova Gorica 2000.
ao •xb7
Black is fine after 20 11xd8+ 11xd8 21
1xc4 Wh4! 22 f4 .i.xf4 23 llxf4 ..xf4 24 d6
tfw11 25 .i.f3 1Wd7 26 �ht �fB 27 Lb7
�ticK the king acts as a blockader of the
-
White wants to win, but maybe this is not
in the cards. Safer is 25 �gl llht + with a
draw by perpetual check.
25...J:r.e8!
Cutting off the king.
26 ird7?
After this mistake Black wins with a sensa­
tional attack. White should play 26 .1e41 f5
27 We21 W'h5+ 28 �d2 fxe4 29 �ct 1i'f51 30
g4! (30 d6? llc8+ 31 �b l e3+ 32 �a1 e2 is
winning for Black) 30.....f4� 31 Wbl llb8
32 'tie7 e31 with a very complicated position
(Macieja/Rozentalis).
26.....,5+ 27 1Wg4 Axf2+ !

l'"���ed d-pawn. Instead of 21...1i'h4, the game

I ull tcr-Aiterman, Recklinghausen 1998 con­
.

llnucJ 2t...1Wf6 22 g3 .xb2 23 llb'l •d4 24

An h6 25 llbct i.cS when we prefer White,
rvcn though lhe players agreed a draw.
IO .i.xh2+
...

1 13
The Petroff Defence
Macicja must have either missed or mis­
evaluatl-d this sacrifice.
28 l:txf2 Wxd5+ 29 Wf4 1td4+ 30 ¢>95
30 �11 :e3+ is winning for Black.
30 1txf21
•••
No perpetual- Black is winning!
31 1tf4
White cannot escape, for example 31 �h4
:e5 32 WeB+ �g7 33 'lfc3 f6 34 'lfc7+ �h6
and there is no defence; or 31 �h6 'iVf6 32
1i'f4 :e5l!, winning instantly.
31 1tc5+ 32 .tf5 1te7+ 33 �h6
•.•
Or 33 �g4 1i'e2+ and Black wins.
33...f6 34 ..i.xg6 1tg7+ 0-1
Game 48
Shirov-Kramnik
Be lgrade 1997
1 e4 e5 2 .!Df3 .!Df6 3 d4 .!Dxe4 4 ..i.d3 d5
5 .!Dxe5 .!Dd7 6 .!Dxd7 .txd7 7 0-0 ..i.d6 8
�c317
This move has been popularised by Shi­
rov. 811'h5 has also been tested:
a) 8...c6 9 :e 1 0-0 1 0 .i.xe4 dxe4 1 1 D.xe4
:C8 121lxe8+ (not 1 2 tlkl2?! 'W'f6 13 'iVe2 c5
14 c3 cxd4 15 cxd4 :ac8 when Black's de­
velopment ensures him of an initiative)
12.. .'iVxe813 .i.e3'it'e4 1 4'it'd1 .i.g4 1 5 Wet
.i.f5 16 c4 1Wh4 1 7 g3 11'h5 and Dlack has
good compensation for the pawn, :>Ge Jun­
Karpov, Buenos Aires 2001 .
b) 8...lnf6 and now:
1 14
b1 ) 9 11'h4 .i.e6 (or 9 ...ll'lc4 10 'iVxd8+
llxd8 1 1 n- 1 1 llcl 0-0 1 2 .ixe4 dxe4 1 3
llxe4 llfe8 1 4 llxc8+ llxe8 1 5 .i.e3 .if4 is
equal- 1 t ...ll'lf6 12 .i.g5 c6 1 3ll'lc3 �f8 1 4
llae1 h6 1 5 .i.xf6 gxf6 16ltld1 fj;}g7 and it is
not clear whether White has any advantage at
all, 1iviakov-P.Nielsen, Bergen 2000) 10 ltlc3
.i.c7 1 1 lle1 (1 1 'irg3 0-0 1 2 ll'lb5 lC!cs 1 3
.i.f4 c6 1 4ll'lc7lllxc7 1 5 .i.xc7 Wd7 is level
- 1iviakov) 1 t.. 'it'd7
. 12 .i.ft a6 13 n ll'le4
1 4 11h5 yfJ 1 5 'it'h6 .iffi 16 11'h4 .i.e7 17
11'h6 .if8 1/z-1/z Tiviakov-Lcko, Wijk aan Zee
2001 .
b2) 91le1 + is a worthy alternative.
9 ...�f8 10 'iVe2 (10 11'h4 is harmless due
to 1 o ..ll'lg4! 1 1 'ifxd8+ llxd8 with equality:
.
1 2 h3lllf6 nll'ld2 g6 1 4ltln �g7 1 5ltle5
.i.c8 16 .igS h6 1 7 .i.f4 g5 18 .i.h21lhe8 1 9
lte2 lllgB 20 .Jt.et f 6 2 1 ll'ln Yz-Yz Movse­
sian-Akopian, FIDE World Championship,
Las Vegas 1 999) 10 ... c6 lt ll'ld2 tlc7 1 2ll'lf1
.i.g4 13 flc3 .ixn 14 'iVxn .ixh2+ 15 �ft
.id6 16 .i.g5 '1t'd8 1 7 lte2 (an improvement
over 1 7 :C3?! i!be41 1 8 .i.xc4 'irxg5 1 9 .i.d3
g6 20 llael !J;g7 21 'it'h3 llad8 22 c3 h5 and
White didn't have quite enough for his pawn,
Sutovsky-David, Istanbul 2003) 17 ... h6 1 8
.ih4 g5 1 9 .ig3 with chances for both sides.
However, we cannot sec how Black can ever
be worse here.
8 .!Dxc3
•••
8. ..'it'h4 has also been tried, e.g. 9 g3 i!bxc3
to bxc3 tlg4 (10 ...11'h3 1 1 llc 1+ �fB 12 'iVn

l:6 13 .i.f4 Jl.e7 14 J:lab1 b6 15 c4 favours
White) 1 t J:lcl+ *d8 (11...*£8 12 .i.e2 9£5
1.� J:lb1 J:lb8 14 c4 dxc4 15 .i.xc4 h5 16 h4 is
also a bit better for White) 12 .i.e2 95 13
Ab1 b6 14 c4 dxc4 15 .i.xc4 J:le8 16 .i.e3
.i.c6 17 d5 .i.d7 18 .i.ft h6 t 9 c4 with a
sli�ht plus for White, Kamsky-Karpov,
FIDE World Championship (Game 6), Elista
1996.
9 bxc3 0-0
A recent game in this line continued
IJ ...Jl.c6 10 llb1 :bs 11 9h5 c6 (1t...g6 12
1irh6 *d7 13 .i.gS 9£8 14 1Vh4 h6 15 .i.f6
.i.c7 16 llfe1 looks slightly better for White)
12 .i.g5 .i.c7 13 .i.xe7 9xc7 14 f4 'iff6 15
l:.bc1 *d8 16 g4 g6 17 9h6
and now:
a) 17...*c7?! 18 f5 ..i.d7 19 fxg6 'ifd6 20
l:txf7 hxg6 21 'iff4 'if'xf4 22 l:txf4 and White
cvcntunUy converted his pawn in Shirov­
Bologan, Sarajevo 2004.
b) Shirov described the position after 17
'irh6 as the most fascinating he had ever
analysed. The foUowing variation is a sum­
mery of several page of his analysh.-:
17 .. .i.xg4 18 51 *d7 19 h3 .i.h5 20 fxg6
. 1 2...c6 13 ..tf4 •c7 14 ..txd6 'irxd6 1 5
.if3! 21 J:le5! hxg6 22 J:.£5! 9xf5 23 .i.x£5+
�-,rxfS 24 1i'f6 J:.bg8+ 25 *f2 .i.e4 26 1fxf7+
�c8 27 9e6+ *c7 28 9e7+ �c8 29 h4 and
White is a bit better.
10 Wh5 f5 1 1 lZb1
11 J:lc1 is interesting: 11...1i'f6 12 •o
WhK 13 :b1 b6 14 ..if4 (or 14 'ifxdS .i.c6 15
3 d4: The Main Lin e
1Wb3 J:lad8 1 6 .i.d2 'if g6 with compensation
- Mikhalchishin) 14....J:lae8 15 J:lxe8 .i.xe8 16
lle1 .i.h5 17 9g.l (17 'ifxh5 .i.xf4 18 9xf5
9xf5 19 .i.x£5 .id2 is drawing) 17..Jle8 18
J:.xe8+ .i.xc8 19 Jl.xd6 cxd6 20 11'0 g6 21
9xd5 9e7 and Black has play for the pawn.
Pavlovic-Mikhalchishin., l..enk 1999.
11 .•• b6
1 2 'irf3
Again White has other options:
a) 12 c4 dxc4 13 Jl.xc4+ *h8 14 ..igS
9e8 15 'iVh4 1i'g6 16 .i.d3 J:lae8 17 c4 h6 18
.i.e3 9f7 was level in Palac-Fridman, Pula
1997.
b) 12 :C1 c6 (this is even better than
12...1i'f6 13 •o c6 14 Jl.f4 b5 15 11fg3 .i.xf4
16 1fxf4 J:lf7 17 *rt l:ta£8 18 J:lc2 ..d8 19
J:lbet � when Black is close to equality,
Shirov-Anand, Groningen 1997) 13 .i.gS
..c7 14 c4 .i.e8 15 9h3 dxc4 16 .i.xc4+ (16
.i.x£5 .i.xh2+ 17 *h1 J:lx£5 18 9xf5 .i.d6
gives Black sufficient compensation - Salov)
16....i.f7 17 .i.e6 Le6 18 l:txe6 J:lae8 with
level prospecrs, Shirov-Kramnik, 2nd match­
game, Cazorla 1998.
Afe1 l:lae8
Black has completely equalised. In fact, it
is White who should be careful. For example,
16 c4 J:.xel+ 17 :Xe1 9b4 and White is
alrt."lldy starting to encounter problems.
1 6 AxeS :XeS 1 7 c4 •a31 18 h4?1
After this move Black is probably slightly
1 15
The Petroff Defence
better. Possibly stronger is 18 1ff4 1fxa2 19
1fc1 1Wa5 20 c3 bS 21 lla1 flc7 22 cxdS
cxdS 23 11'c2 g6 24 ..-.,3 .i.c6 25 1tb4 when
White has good compensation for the pawn.
18 11'xa2 19 llf1 dxc4 20 -*.xfS -*.xfS
•••
21 Wxf5 c3
21...a..'i 22 1Vd7 llf8 23 :C1 gives White
enough compensation. One point is that
after 23 ..flxc2 24 1We6+ Black must acqui­
. 9 bxd 0-0-0 10 'itfJ i&.e6 1 1 .i.f4 (what
esce to a draw as 24...�h8?? 25 "lrf71 :gB 26
:CB wins for White.
22 Wd3 WdS 23 •xc3 lle4 24 Wg3
llxd4 25 Wb8+ Wd8 26 Wxa7 h&
26 .:Xh4 27 11t'b7 'Wd6 28 "IreS+ *f7 ztJ
..
1ff5+ is also a draw, but Kramnik could have
continued playing for a win with 26...:C4 27
1Vb7 h5 28 lla1 �h7. Note, however, that
27...:Xc2?? is out of the question due to 28
llat.
27 Wa2+ �8 28 g3 bS 29 We& 1t'd5
30 1t'xd5 llxdS 31 lla1 llc6 32 lla2 �7
33 �1 %-%
Game 49
Anand-lvanchuk
Linares 1993
1 e4 eS 2 .!Df3 .!Df& 3 d4 tt.'\xe4 4 ..i.d3 d5
5 lLixe5 lLid7 6 lLixd7 ..i.xd7 7 0-0 1i'h4
This is the 'active' main line.
8 c4
This is the principal move. The alterna­
tives fail to impress:
1 16
a) . 8 �2 lllxd2 9 1Wxd2 .i.d6 (not
9...1Wxd4?! 10 l:le1+ .i.e7 11 WaSt? b6 12
'Wa6 with compensation - Yusupov) 10
11'e2+ (after 10 :C1+? Wf8! Black wins a
pawn, 11 f4? losing to I t....i.xf4Q 10....i.e6
11 f4 0-0 12 g3 1Vf6 13 f5 .i.d7 14 .i.e3 l:lfe8
and Black is no worse.
b) Blllc3 is also not very strong: 8...lllxc3
else?) 11....i.g4 12 'ite3 .i.d6 13 .i.xd6 llxd6
14 llaet .i.d7 15 'itfJ 'Wf6 16 'Wg3 (or 16
1Wxf6 llxf6 1 7 c4 dxc4 18 .i.xc4 .i.fS 19 .i.b3
l:ld8 20 dS c6 with a likely draw) 1 6..Jlc6 17
:eS!? l:r.xc3 18 .J:hdS .i.e6 19 llg5 g6 20 dS
.i.d7 21 'ife3 lla3 with unck·ar play,
Klovans-Raetsky, Apolda, 1994.
c) 8 g3
and now:
cl) 8...'ifh3? looks natural but White can
play 9 f3 llld6 (9...lllf6?! is even worse after
10 llet+ .i.e6 11 :CS! gS 12 .i.f5 g4 13 .tgs
.i.e7 14lllc3!, when Black is in trouble, Pav­
lovic-Konguvel, Bcnasque 1998) 10 :CI+
i&.c7 11 lleS .i.c6 12 g4 ..h4 13 i&.gS, win­
ning the queen fur insufficient compensa­
tion.
c2) R flf6! 9 .i.xe4 dxe4 10 lllc3 1fg6 11
..
lllxe4 0-0-0 12 :C:1 hS and the weakening of
the light squares guarantees Black compensa­
tion for the pawn.
8 0-0-0
•••
Other moves arc weaker: �
a) 8...i&.d6 9 g3 ..h3 10 lllc3 lllxc3 11

hxd dxc4 12 ltc1+ Wf8 13 i.xc4 i.g4 14
'it'll3 should favour White.
b) 8 ...lt�f6 can be met by 9 llkl21?, for
example 9.....xd4 10 .!DO 'irb6 11 cxdS il.e7
(11....!Dxd5 12 o!Dgsl h6 13 .!Dxf7 Wxf7 14
'it'h5+ WgH 15 ..xd5+ i.e6 16 ..e4 is great
for White) 12 l:lcl 'i'd6 (12....!Dxd5 13 .ic4
c6 14 .i.g5 gives White a very strong attack)
13 i.gS h6 14 i.h4 'ifxd5 15 l:lct i.c6 16
'it'c2 'ild8? (Z.Almasi-Cs.Horvath, Hungary
1997) and now 17 l:lxc6! bxc6 18 .i.xf6 gxf6
19 o!Dd4 is a nice win. 16.. .'1Ve6 is stronger,
hut Black's position is still unenviable.
9 c5
9 cxdS is less critical. Black continues
IJ...i.d6 10 g3 (10 h3? .i.xh3! is simply too
dnngerous!) 10... .!Dxg3! 11 fxg3 i.xg3 12 'iVc2
'it'xd4+ 13 Wg2 i.xh2 14 Wxh2 9h4+ 15
'wt>gl 'iVg3+ 16 'ilg2 Wxd3 with great coun­
tcrplay for the piece.
9 o!Dc3 is also not to be feared.
Black can pL1y the following:
a) 9....!Df6 10 cS .i.e7 11 .ie3 .!Dg4 12 .if4
looks good for White.
b) 9...i.d6 10 g3 .!Dxg3!? 11 fxg3 i.xg3 12
'irc.J2 'iVxd4+ 13 ..e3 i.xh2+ 14 Wxh2 'iVh4+
15 c;t>g1 ..g4+ 16 W£2 d4 17 'lfg3 dxc3 18
bxc3 i.e6 is equal according to Yusupov and
lliibner, but to f4 may be better for White.
c) 9....!Dxc3 (the easiest) to bxc3 dxc4 1 1
�xc4 i.d6 12 f4 f 6 13 .i.dS c6 14 g3 Wg415
.if3 ..g61? (but not 15...'iVe6 16 'iVa4 a6 17
J:tbt hS 18 .i.d2 1t'f5 19 l:lb2 and White had
3 d4: The Main L in e
the attack i n Rublcvsky-Ngucn Anh Dung,
Moscow 2001) 16 l:l£2 hS 17 l:lb2 h4 with
reasonable countcrplay.
9...g5 10�3
Other options will be discussed in Games
51-52.
1 0....ig7
The best reply. The alternatives have not
been successful:
a) 10....!Df6?1 11 g3!? 'iVh3 (1t ...'iVxd4?1 12
c61 is a well-known trick) 12 'ilf3 ltlg4 13
Wg2 'i'xg2+ 14 Wxg2 h6 15 f41 .i.g7 (or
15...l:e8 '1 6 fxg5 hxg5 17 h3 o!De3+ 18 .i.xe3
:Xe3 19 i.f5 with a clear edge) 16 h3 .!Df6
17 fxg5 hxgS 18 g4 and White has a cll:ar
advantage, Tseshkovsky-Gagloshvili, Kras­
nodar 1997.
b) 10...f5?! 11 .!Dxd5 and now:
b1) 11....!Dxf2l? 12 l:lx£2 'lfxd4 13 .i.xg5
(we think that White should play 13 c6!?
..xdS 14 cxd7+ l:lxd7 15 l:ld2 .i.b4 16 .i.c4
..c5+ 17 Wh1 i.xd2 18 i.e6 with a clear
edge) 13 ...i.xc5 14 �3 (14 WOI? i.c6 15
i.xd8 l:lxd8 16 i.xfS+ Wb8 17 lltll 'it'x£2+
18 'ifx£2 i.x£2+ 19 Wx£2 i.xdS 20 .i.xh7
looks great for White) 14...f4 15 i.xf4 l:lhf8
16 Wft ..xe3 17 i.xc3 .i.xe3 18 ..e2 i.x£2+
19 �ht ltde8 20 ..c2 and White had decent
chances to win the endgame, Wetlberg­
Schneider, Torshavn, 1987. Even so, White
should punish Black further with deviations
on move 13 or 14.
b2) tt...i.c6 also does not work after 12
117
The Petro ff Defence
ltk3l:lxd4 (or 12...i.g7 13 g31i'h6 14 olllxe4
fxe4 15 .i.xe4 l:lxd4 16 1i'f.3 and Black is in
trouble) 13 .i.c3l:ld8 14 'Vc2 olllxcS 15 Lf5
.i.d6 16 g3 1fc4 17 .i.xgS l:ldfB 18 .i.xe6+
o!ilxe6 19 .i.c3 and Black is simply a pawn
down, Glck-Raetsky, ZeD am Ziller 1993.
1 1 g3
This is the most critical line for Black.
Other options arc featured in the next game.
11 1rh3
.•.
Another possibility is 11...'ifh6!? 12 olllxe4
(12 o!ilxd5?! .llhc8 13 o!ilc3 o!ilxf21 14 l:lxf2
.i.xd4 with a clear edge for Black- Yusupov)
12...dxc4 13 .i.xe4
and here Black has two choices:
a) 13.....th3 14 'iff.3!? .i.xfl 15 ..txb7+
�b8 16 .i.xgS!? (16 Wxft -'.xd4 17 h4 'ife6
18 .i.xgS f6 19 .i.f4 is very messy) 16 ...Wxg5
17 Wb3 .i.c4 18 Wb4 aS 19 'ifxaS c6
(19...�b7 20 c6+ �b8 21 1lxg5 .i.xd4 also
1 18
unclear) 20 1ib6 9e7 21 -'.&6+ Wa8 22'
9xc6+ Wb8 23 'ifh6+ Wa8 and now taking;
the draw by perpetual check with 24 1i'c6+
might make sense as after 24 Lc4 24...-.,7
the game could go either way.
b) 13... 15 14 ..tg2 with a further branch:
b1) 14....i.c6 15 .i.xc6 (15 d5 .i.xd5 16
..txdS c6 17 'ifa4l:lxdS 18 11xa7 f4 is unclear
according to Vladimirov - White has prob- ·
terns developing) 15...1i'xc6 16 .i.xgS l:lxd4 •
17 •c2 l:ld5 18 .ie3 gave White some ad­
vantage in Macieja-Stefanova, Krynica 1998.
b2) 14...f4 15 d5 (after 15 gxf4 gxf4 16,
•o ..tc6 17 1fg4+ �b8 18 dS- 18 ..txf4?
l:lxd4! would be a nasty surprise- 18....i.xd5i
19 .i.xf4 9f6 Black has exceUent compensa­
tion for the pawn) IS ...llhfB 16 l:le1 �b8 17
d6 cxd6 (Arencibia-Vladimirov, Leon 1991)
and now after 18 11fxd6+ •xd6 19 cxd6 .i.d4
20 gxf4 gxf4 21 .lle4 l:lde8 Black can be·
happy with his counterplay.
1 2 �xe4
The older line is 12 olllxdS l:lhe8
(12....i.g4? 13 .ixe41 .i.xd1 14 .i.f5+! and
White wins- 1-J.Olafsson) 13 Wf3 (13 .i.xe4?.
l:lxc4 14 .i.xgS ..tc61 1 5 .ixd8 .i.xdS 16 f3
l:lxd4 17 'lfct l:ld3 gives Black a decisive
attack) 13.....tl5 14 �3 (14 o!ilc3?! l:lxd4 15
lle1 -or 15 .ixe4?Il:ldxc4 16 ..txgS ..txc3
17 bxc3 l:le2 and Black has aU the threats-
15....i.g4! 16 11fxt7 l:lffi 17 WxfB+ .i.xffi 18
..tn 11'hs 19 olllxe4 .i.l5 and Black's material,
gains give him the advantage) 14....i.g6

and now:
a) 1 5 c6 is risky: 1 5 .. .:xd4 1 6 .i.c2 hS 1 7
c:xh7+ 'itlb8 1 8 Act '1Vd7 1 9 lLlfl g4 20 1i'g2
ilkS 21 i.e3 .i.t.-4 22 f3 .i.xb7 and it was
Ulack who had the attack, Smagin­
H.Oiafsson, Sochi 1988.
b) 15 d5 lLld2 (also good is 15...lLlxc51? 16
.i.fS+ .i.xf5 17 lL!xf5 .i.f6, witl1 chances for
both sides) 16 .i.xd2 .i.xd3 17 '1Vxf7 .i.xb2
(17 ... i.xf1 ? 18 ..xg7 Axc3 19 :Xfl Ad3 20
.i.xg5 is better for White) 1 8 d6 1i'd7 1 9
1Wb3 'ilbs with unclear play, Joscliani­
l lowell, Spijknisse 1989.
12 ...dxe4 13 .ixe4 .ib5 14 .ig2!
Anand's improvement over a previous
).,'llme against lvanchuk, albeit with colours
n:versetl. With 1 4 .i.g2 White obvious does
not worry about losing the exchange; the
attack is much more important, and Black
hsls no way to force exchanges.
The alternative is 1 4 .i.xg5 :Xd4 and
now:
a) 1 5 'Wb3 Axc4 1 6 1i'xb5 h6 17 .i.e3
l:h4! 18 Afd1 1txh2+ 19 'itlfl 'irh3+ 20 �el
l:e4 (20...l:th5 21 1Wa4 a6 22 c6 Wb8 is un­
clear) 21 c6 Axe3+ 22 fxe3 'irxg3+ 23 �e2
'irg2+ 24 We I ...g3+ 1/z-'12 Dolmatov­
Akopian 1988.
b) 1 5 i.g2 'irf5 (or 15 ...'irc6 16 1Wf3 .i.c6
1 7 1Wb3 .i.xg2 18 Wxg2 ...dS+ 1 9 1Wxd5
l:xd5 20 .i.d .i.xb2 21 Aab1 .i.g7 and Black
i� only very slightly worse) 1 6 1i'b3 c6 1 7
i.c3 .i.xf1 1 8 llxfl (1 8 .i.xd4 .i.xd4 1 9 Lfl
3 d4: The Main L in e
LcS i s equal - Howell) and here Black must
make a choice:
b 1) The dubious 1 8 ...Ad7?1 is met by 1 9
•a4:
bt l) 19 ...'itlb8 20 b4 W'd3 21 ..tf4+ WaS
22 .i.d6 (this looks great for White so Black's
reaction Lo; understandable) 22....ZZ.xd6 23 cxd6
1i'xd6 24 bS cxbS? (24... c5 25 ...c4 flc7 26
Act with a clear edge - Howell) 25 '8'xb5
llb8 26 ltb1 'irc7 27 a4 i.c3 28 ltct flc5 29
1ib3 .i.d4 30 .i.xb7+! 1 -0 Gcller-J.Howcll,
Reykjavik 1 990.
b12) 19 ... a6 20 .i.xc61 bxc6 2 1 1Wxc6+
wds (21 ...�8?! 22 Wb6+ l:tb7 23 1i'd6+
�c8 24 1i'xa6 �b8 25 c6 and White wins;
21 ...ltc7?! 22 1i'xa6+ �b8 23 c6 ltc8 24 ltd1
1ic2 25 lld3 ltxc6 26 'iVa7+ �c8 27 1i'd7+
1-0 Hracek-Haba, Czech Republic 1998) 22
lte 1 'irf6 23 'ira4 1i'xb2 24 c6 .i.c3 25 cxd7
.i.xe1 26 1Wc6 .i.xf2+ 27 .i.xf2 W'b1+ 28
Wg2 'Wb2 29 �f3 with a clear plus for White
- Hracek.
b2) 18 ...:hd81 1 9 h4 (or 19 'ira3 Ddt 20
1ixa7 ltxft+ 21 .i.xfl .i.c5 22 1i'a8+ �c7 23
1Wa5+ with a draw - Yusupov) 1 9... .i.e5 20
.i.xd4 .i.xd4 21 �h2 �c7 22 flc4 .i.xcS 23
b4 .i.b6 24 b5 ltd4 25 Wb3 ltd3 26 'iVc4
lld4 and Black escaped with a draw, lvan­
chuk-Anand, Roquebrune 1992.
14 ...1rf5 1 5 .ie3
1 5 'irb3 c6 1 6 ltd l l:txd4 17 ltxd4 .i.xd4
18 .i.c3 is equal - Akopian.
1 5 ....ixf1 1 6 .ixf1 Zlhe8
1 19
The Petroff Defence

1 6...c6 1 7 Wa4 looks good for White: game via 22...1lxJ4! 23 .i.xd4 1Wxd4 24 Wb3
1 7... h5?! 1 8 1i'xa7 llxd4 19 llel l lld7 :C7 25 .i.fl 1i'xc5, with chances fur both
(19...1ldd8 20 .i.a6! and winsQ 20 b4 1Wg4 21 sides.
b5 and White has a very strong attack. 22 .hd4?
...

1 7 1ra4 ¢'b8 18 l:ld1 c6 More resilient is 22...1le7 23 llb4!? f5 24

1 8. .1lc6 19 d5!? llxd5 20 .i.h3! 1i'e4 21
. d5 f4 25 d6 lled7 26 .i.xa6, when Black is
:Xd5 1i'xd5 22 .i.xe6 1Wxe6 23 c6 also a..__ still just about breathing.
surcs White of an edge. 23 l:lxb7+1 �xb7 24 •xa6+ �8 25·
1 9 l:ld3 1re4 20 l:la3 a6 21 .i.d3 1tb6+ �8 26 •xc6+ �b8 27 'Wb6+
�a8 28 .i.b5 1 -0

Game 50
J.Howeii-Makarychev
Fnmze 1989

1 e4 e5 2 li)f3 ll:)f6 3 d4 .!Oxe4 4 .i.d3 d5

5 .!Oxe5 .!Od7 6 li)xd7 .txd7 7 0-0 1ih4 8
c4 0-0-0 9 c5 g5 10 lilc3 .i.g7 1 1 .!082

21 ...1rg4?1
Soon after this game Black's play was re­
paired to some extent with 21 ...1Wd5 22
.i.xa6 .i.xd4 23 .i.xb7 .i.xc5! 24 .i.xc6 .dl +
25 Wg2 Wxa4 26 llxa4 llxe3 27 fxe3 lld2+
28 �f3 llxb2 (Hemandez-J.HoweU, Matan·
zas 1 993). This position should lx: within the

'drawing range', although it will take Black

some time and technique.
22 l:lb3! 1 1 �xd5 .i.xd4 12 i.e3 .i.xc3 1 3 fxe3
docs not seem frightening for Black after
1 3 ..ie6 (but not 1 3...�xc5?! 1 4 .c2 .i.e6
1 5 �xc7 Wxc7 1 6 11Vxc5+ �b8 17 llac1 with
attacking chances for White) 1 4 �7+ �b8
1 5 �f5 .i.xf5 16 llxf5 llhc8 when Black has
good counterplay. 1 1 i..c3 allows the trick
1 1 ...�xc5!?, when 12 �xd5 �xd3 1 3 1i'xd3
Wb8 looks equal.
1 1 1lhe8
...

1 1 ... f5 is an attractive alternative. Follow­

ing 1 2 f3 Black has a choice:
a) 12...�f6 1 3 i.e3 f4 14 i.. f2 11Vh6 1 5
1rd2 l:lhe8 1 6 llacl �>8?! 17 b4 �7 1 8 b5
22 i..xa6?! aUows Black back into the (Black has clearly wasted his chance for an

120
 3 d4; Th e Main L ine

1111�\ck) 18 ...�b8 19 ru ll!f5 20 c6 .i.c8 21 13 'ii'e t 'ii'xet 14 l:lxe1 h6 1 5 .i.d2 lllg8 is

cxh7 �xb7 22 .i.xf5 .i.xf5 23 lba4 with a equal. Now after 16 l:ladt ?l liJc7 17 .i.e3
clc:tr plus, Tiviakov-Raetsky, Makhachkala lbf5 18 ..tf2 �3 1 9 l:lct l:te7 20 b4 l%de8
IIJK7. Instead of 16 ...lbg8?!, Black needs to Black was already better in Vidarsson­
RCI quickly with 1 6 ...g41? 17 1rxf4 gxf3 1 8 Ractsky, l-lafnarfjonlur 19%.
gx £3 .i.h3 1 9 l:lfd1 lbh5 20 1rxh6 .i.xh6,
IIKsuring him of reasonable counterplay.
b) 1 2...l:lhf8

1 3 .llxe21
•.

An idea of Akopian's.
14 'it'xe2
and now it's White who has to make a de­ 14 .i.xe2 is strongly met by 14...�g41
ci:;ion:
b1) 1 3 fxe4?! (this reckless move hasn't
been tried) 13 ... fxe4 14 l:lxf8 (not 14 .i.c2?
J:lxfl+ 1 5 1rxf1 l:tf8 and the game is already
over) 1 4 ...l:txf8 1 5 .i.e3 exd3 16 1Vxd3 l:tf6
and Black has the advantage.
b2) After 1 3 'ii'e 1 Black should exchange
with 1 3 ...1Vxet. Instead 13 ...'ii'h5?1 proved
futile after 14 fxe4 dxe4 1 5 .i.c4 f4 1 6 d5
'it>b8 17 c6 .i.c8 1 8 cxb7 .i.g4 t 9 lbxf4!? gxf4
20 .i.xf4 and White was virtually winning,
Mannion-Kobese, Yerevan Olympiad 1996.
b3) 1 3 a4 l:tde8 with a final split:
b31) 14 g3 lllxg31? 1 5 hxg3 (15 lbxg3 and now:
..ixd4+ 16 �h 1 f4 1 7 lbe2 .i.h3 gives Black a a) 1 5 fxg4? ..i.xd4+ 1 6 �hl .i.e5 17 .i.f4
good attack) 1 5...'ii'h3 1 6 c6! (16 l:lf2?! .i.xd4! (even worse is 1 7 h3? 1lg3 18 �g1 'ifh2+ 1 9
17 lllxd4 1Vxg3+ 18 Wfl f4 is very bad for � f2 .i.g3+! 20 �e3 - 2 0 W f3 .i.h4! -
White) 16....i.xc6 17 l:tf2 .i.xd4 1 8 lbxd4 20...1le8+ 21 �d3 .i.bS+ and Black wins)
'irxg3+ 19 �fl 'ii'h3+ with a draw. 1 7 ....i.xf4 1 8 l:lxf4 gxf4 19 1rxd5 l:lcHI 20
b32) 1 4 .et 'ii'xet 1 5 Axel f41 16 fxe4 •d 1 .i.c6 21 1i'f1 :C3 22 :C1 1re7 23 .to
tlxc4 1 7 .i.c4 f3 18 .i.e3 fxe2 1 9 l:lxe2 c6 20 .i.xf3 24 gx£3 lle2 and Black had an edge in
dS cxd5 21 .i.xd5 h6 22 l:td2 l:td8 with an Sherzer-Halasz, Budapest 1990.
level ending, Sax-Salov, Brussels 1 988. b) 15 .i.f4 lbf2!? 16 .i.g3?! (16 J..xg5
1 2 f3 li)f6 1 3 .td2 lbh3+! 1 7 gxh3 1Vxg5+ 18 �h1 1Vf4 with

121
 The Petroff Defence

unclear play is stronger) 16...llhd1 17 ..i.xh4 lle8. Although Black is better, the game is
..i.xd4+ 18 �ht lt)e3 (we prefer Black's posi­ still open.
tion after the stronger 18...lbxb2!? 1 9 ..i.xgS 21 .. Jlf8 22 ..txh7?1
l:le8) 19 ..i.xgS l:le8 20 ..i.xc:3 llxe3 21 l:lfe1 Now it is all over. 22 c6 bxc6 23 llee1
.i.xb2 22 ..i.fl .id4 with an unclear endgame, .ie6 was the last chance.
Movsesian-Raetsky, Pardubicc; 1 992. 22.....te6 23 Ae1 ..te3 24 g4 AhB 25
.i.f5 �dB 26 Le6 fxe& 27 ..txe6 Lh2
28 Ad1 Ah1 + 0-1
White resigned because of 29 �e2 l:lxd1
30 Wxd1 c6 31 gS ..i.xc5.

Game 51
Kasparov-lvanchuk
Debrecen 1992

1 e4 e5 2 lLlf3 lbf6 3 d4 lbxe4 4 ..td3 d5

5 lbxe5 lbd7 6 lbxd7 .i.xd7 7 0-0 Wh4 8
c4 0-0-0 9 c5 g5

1 4...c!bh51?
14...Wxd4+ 1 5 .te3 WeS 16 Wd2 is !>imply
great for White.
1 5 1if2 •xf2+ 1 6 �xf2
1 6 :Xt2 ltlf4 17 .i.xf4 gxf4 1 8 c6 .i.c6 1 9
cxb7+ Wb8 with unclear play (Yusupov)
is probably a better option.
1 6. . .1i�f4 1 7 ..txf4 gxf4 18 Afe1 .txd4+
1 9 �1 ..txb2
19 ... ..i.xc5 20 J:leS ..i.e6 21 J:lhS allows
counterplay.
20 Aab1 ..td4
1 0 ..te3
White has a couple of alternatives here:
a) 10 g3 Wh3 11 lbc3 t5!? 12 .ie2 l:Zg8 13
..i.f3 .ie8 14 a4 �6 15 lla3 1i'f6 16 b4 ..i.g7
17 lbe2 hS with chances to both sides,
Pilipovic-Nikcevic, Tivat 1995.
b) 10 lbd2 lbxd2 (also possible is
lO ... i.g7 11 ltlf3 �5 12 lbcs Wxd1 13
l:lxdt .i.e6 14 i.c2 f6 1S ltl.n t5 16 lbe5 f4
when Black is no worse) 11 i.xd2 llg8 12
.i.c3 (or 12 l:lct !? Wxd4 13 ..i.c3 1i'h4 14
..tf6 l:le8 15 f4 i.e7 16 c6!? with unclear
play, Pinkas-Kuczynski, Wroclaw 1987; one
21 Ae7?! continuation is 16...i.xc6 17 .it5+ Wb8 18
More resilient is 2 1 l:lect .ic3 22 l:lc2 llxc6! bxc6 19 1ib3+ �a8 20 i.d7 .icS+ 21

122

"'h I .i.b6 22 .i.xc6+ 'iPb8 23 .i.xe8 llxcB
wilh a messy position) 1 2...g4 1 3 b4 Wb8 1 4
b 5 1ff6 1 S .i.d2!? ..i f5 (1 S...1i'xd4 1 6 .i.c3
l'c5 17 1fa4 gives White good compensa­
tlon) 1 6 .i.xfS 11fxf5 17 a4 11fd3 1 8 .i.f4
I'Kd1 1 9 llfxd1 llg6 20 aS WeB with an
c;umplicated struggle, Komeev-Raetsky,
Cannes 1 994.
10 Ae8
...
Black should probably look carefuUy at
rhc alternatives to this:
a) 10....i.g7 1 1 f3 lbf6 transposes to 10 f3,
while I t ...lbxcS?! 1 2 dxcS .i.xb2 1 3 lbd2
.1Ka1 14 'lrxa1 is fantastic for White.
b) 10...lbf6 is certainly a playable altcma­
rive: 1 1 lbd2 J:lgs (the main point) 1 2 lbf3
WhS 13 lDeS 11fxd1 1 4 :axd1 .i.e6 l S f4
gxf4 16 .i.xf4 (1 6 :xf4 .i.h6 1 7 :o .i.xe3+
i H :xe3 lbe4 1 9 :n with a slight edge,
, lbbS Wb8 1 S 11'd2 g4 16 11'aS .i.xbS 17
�hould be considered) 1 6 ...lDe4 17 c6 f6 1 8
cxb7+ 'iPxb7 1 9 lbc4 f5 (l.Gurevich-Barua,
Hastings 1 993/94). Now White could
probably have been slighdy better after 20
li)e3 .i.g7 21 .i.eS .i.xcS 22 dxeS .rigS 23
J:lr4.
c) lO .. f5 is also enticing - we see no fault
. :Xe4 (13 ...dxe4 1 4 lbxgS irg6 1 S dS h6 16
with this move. For example, 1 1 f3 lbf6 (not
l l ...f4?! 12 fxe4 fxe3 1 3 g3 Wh3 1 4 cSI hS?I
1 5 c6! and Black was in deep trouble,
Emcste-Goldmane, Riga 1 989) 1 2 lDcJ (1 2
c6 .i.xc6 13 ..ixf5+ i.d7 1 4 g3 'lfhs 1 S
i.xd7+ :xd7 1 6 lbc3 .i.g7 gives both sides
chances) 12 ... f4 1 3 .i.f2
3 d4: The Main L in e
and now:
ct) n...Whs t4 lbbs Wb8 1 s 11'et !? g4 1 6
'ireS 'irxeS (Black should play 1 6....i.xbS 1 7
1i'xhS lbxhS 1 8 .i.xbS gx f3 1 9 gx f3 .i.e7 20
:ret ..if6 with a slighdy worse position) 1 7
dxeS lbhS 1 8 fxg4 .i.xg4 (Khait-Raetsky,
Upetsk 1 993) and now after 19 c61? bxc6 20
lbd4 Wb7 21 lbb3 Black's position looks
uncomfortable.
c2) 1 3...'1'h61? is stronger: 14 l:tet?! g4
(Black now has the initiative) 1 5 c6 .i.xc6 1 6
.i. f5+ Wb8 1 7 fxg4 .i.d6 1 8 •o 1i'g7 1 9
l:te2?1 ( 19 .i.h4 is better, although Black's
position remains preferable after 1 9 .:d£B 20
..
gS lbd7) 1 9 ... hS 20 gxhS J:lxhS 21 .i.h3 l:tdh8
22 Wf1 &4 and Black was close to winning
in Werner-Raetsky, Cappelle Ia Grande 1 999.
Instead of 14 l:te1?1, White shouW play 14
'iVxbS (but not 1 7 .i.xbS?? gxf3 and Black
has a winning attack) 17 ...c6 1 8 11'a5 .tg7
with chances for both sides.
1 1 .!i)d2 .i.g7 1 2 �ta -.s 1 3 lt)xg5
This leads to a slighdy dull position, but
one that is preferable for Black. 1 3 .i.xe4
c6 is also reaUy messy) 1 4 lbxgS l:lg4 1 S f4
(IS lbm llxg2+1 16 Wxg2 .i.h3+ 1 7 Wh 1
.i.g4 and Black wins) 1 S...h6 1 6 lbf3 J:lg8 is
certainly very unclear.
13 •xd1
•••
1 3....i.g4?! is strongly met by 1 4 f3 lbxgS
1 S fxg4 1Wh6 1 6 .i.f5+ Wb8 1 7 .i.f4 and
123
The Petroff Defence
White is dominating the board.
1 4 llfxd1 lDxg5 1 5 .i.xg5
1 5....i.g4
1 5 ....i.xd4 1 6 c6 .i.c6 1 7 cxb7+ Wxb7 1 8
i.bS Aeg8 19 Axd4 AxgS 20 llct i s a bit
unpleasant for Black.
1 6 lld2 .i.xd4 1 7 c6 .i.e5 18 .i.b5 b6 1 9
.i.h4 llhg8
The alternative 1 9 ...d4 20 Lg3 .i.g7 2 1
.i.a6+ WbB 22 a4 would be dangerous for
Black (Kasparov).
20 .i.a6+ � 21 .i.g3 .i.xg3 22 hxg3
lidS 23 lld4 lld6 24 llc1 .i.c8 25 .i.d371
25 .ixc8 is a better move: 25...'�xc8 26 b4
a6 27 a4 h5 28 bS and White has some pres-
sure.
25 ...llg41? 26 llxg4 .i.xg4 27 f3
27 .i.xh7 d4 28 .i.d3 .i.c6 29 b3 .i.dS
gives Black sufficient countcrplay (Kas­
pamv).
124
27 ....i.e& 28 b4 d4 29 a4 a5 30 b5 .i.b3
31 lla1 �c8 32 �2 h6
32 ...h51? is probably stronger: 33 lit>e2
Wd8 34 c;t>dz c;t>e7 35 .i.c2 .ixc2 36 c;t>xc2
llg6 37 c;t>d3 llxg3 38 lla2 c;t>d6 39 c;t>xd4
.J:lgS and Black should be okay.
33 We2 � 34 �2 �e7 35 .i.c2 .i.c4.
36 llh1 d371
In time trouble lvanchuk pushes his
pawn, but he only succeeds in weakening it .
36. .1le6 37 .i.d3 .i.b3 38 llh4 lld6 39 .i.c2
.
i.c4 40 lle4+ c;t>f8 41 g4 is slightly better for
White (Kasparov).
37 .i.d1 �8 38 llh4 .i.a2 39 lle4 llg6
40 g4 h5
Or 40 ..:d6 41 llc3 .i.b I 42 c;t>c3 and
.
White has a great advantage.
41 f41? hxg4 42 f5 lld6 43 llxg4 .i.d5
44 llh4 �g7 45 g3
45 ...�6?1
This loses and should be avoided, though
4S....ia2 46 llc4 Cit>f8 47 lle3 is also nice for
White.
46 llh8 'ite5 47 .Z:c8 �d4 48 llxc7
llh6 49 l:ld7 llh2+ 50 �1 1 -0
Game 52
Lastin-Najer
Elista 2000
1 e4 e5 2 ltlf3 ltlf6 3 d4 ltlxe4 4 .i.d3 d5
5 ltlxe5 ltld7 6 ltlxd7 hd7 7 0-0 'ifh4 8
c4 0-0-0 9 c5 g5 1 0 f3 ltlf6 1 1 .i.e3

11 .tg7
.•.
In this rabiya Black has many interesting
c 'PP<>rtunities:
:•) 1 1 ...g4 12 g3 Wlh5 13 f4 :Cs 14 Jlet
.i.g7 1 5 �3 Jle7 1 6 Wd2 Jlhe8 17 b4 .i.f5
( 1 7...tDc4 1 8 .i.xe4 dxe4 19 b5 Jld8 20 Wb2
looks very dangerous for Black) 1 8 c6!? bxc6
I IJ ll'la4 .i.xd3 20 Wxd3 ltle4 21 Jlact with
compensation, M:usuura-Kapelari, Sao Paulo
1 997.
b) I t ...llg8 1 2 ll'lc3 g4 looks very att:rac­
livc:
b1) 13 Wet ?! is strongly met by t3 ...g31 1 4
hxg] l:txg3 1 5 lLlez (even worse is 15 Wd2?
.ixcS 16 dxc5 :dgS 17 l:tfd1 d4 1 8 c6 dxe.3
1 9 cxd7+ Wd8 0-1 Dolmatov-Makarychev,
Palma de MaUorca 1 989) 1 5....i.d61 1 6 'l'f2
( 1 6 cxd6? l:txg2+! and it is all over) 1 6 ....i.h3
1 7 ll'lxg.3 .i.xg3 18 'l'c2 l%g8 with a fantastic
nrtuck.
b2) 1 3 g.3 Wh3 14 f4 lllh5 1 5 Wet (weaker
is 1 5 .i.f2? ll'lx£41 16 gxf4 g3 1 7 .i.xg3 l:txg3+
1 8 hxg3 ..xg3+ 19 Wh1 'ifh4+ 20 Wg1 .i.g7
nnd Ulack wins, despite being a rook down)
I S .. .lle8 1 6 'l'f2 ltlf6 1 7 Jlae 1 hS with an
unclear game.
c) 1 t...l:te8 12 'lfd2 (12 .i.f2 1Wh6 1 3 ll'lc3
g4 1 4 f4 - 14 l:tct c6 1 5 b4 gxt1 16 Wxt1
lL'l,.,r4 also gives Black considerable counter­
play - 14 ...11'xf4 1 5 .i.h4 .xd4+ 16 Wh1
.ig7 with unclear play - Makarychev)
12 ....i.g7 (12...:Xc3? is too optimistic: 13
Wxd lllhS 14 g3! ll'lxg3 1 5 'IreS l:tg8 16
3 d4: The Main L in e
hxg3 1l'h3 17 l:td1 .i.g7 1 8 .i.ft 1-0,
Tomashevic-Kondali, correspondence 1 991)
13 .i.f2 'lfh6 14 'l'a5 Wb8 1 5 lLlcl l:te6 1 6 a4
(or 16 ltlb5 .i.xbS 1 7 1i'xb5 ll'lb5 18 g3 '..f6
19 Jlae1 c6 20 'ifa5 l:the8 and Black has de­
cent counterplay) 1 6...:C8 17 Jla3 g4 1 8 ll'lb5
l:ta6 19 Wb4 brxf3 20 ll'lxc7!? :Xc7 21 .i.xa6
fxg2 22 Wxg2 lllh5 with a real mess, Mor­
gado-Gottardi, correspondence 1 997.
1 2 l003
1 2 g3 is worth considering: 12...'1Fh5 1 3
ll'lc3 D.he8 14 l:te 1 g4 1 5 f4 ll'lc4!? 1 6 li)xe4
{16 1Vc2 ll'lxc5! 17 dxc5 d4 1 8 .i.f2 dxc.3 19
bxc3 .i.c6 gives Black a fine game) 1 6...dxe4
1 7 .i.e2 1Vg6 1 8 Wb3 (1 8 dS .i.c6 19 .i.xg4+
�b8 20 f5 1Wf6 21 d6 cxd6 22 cxd6 1Wxb2
isn't clear ac aU) 1 8....i.c6 1 9 .l\.c4 'if£5 20
:Cd1 (20 .i.xfl l:te7 21 dS .i.xd5 22 .i.xdS
'l'xdS 23 'l'a4 a6 shouldn't be worse for
Black) 20....i.d5 21 c6 bxc6 22 lLct llc6 2.3
1Wa3 .i.xc4 24 :Xc4 Wbs 25 l:tdct and White
had fantastic play for the pawn in Tiviakov­
Rozentalis, Groningen 1997.
1 2 ...Dhe8 1 3 .i.f2 1ih6 14 irb3
Another try is 1 4 ll'lb5 Wb8 1 5 a4 ll'lh5 1 6
Wd2 a6 1 7 l:tfcl!? li)f4 (not 17...axb5? 1 8
axbS b6 1 9 1Wb4 with a winning attack) 1 8
.i.fl l:le6 1 9 'I'aS l:tc6 20 lL3 .i.c8 21 lle 1
..tf6 22 1Vd2 'ifg7 with chances for both
sides, Zulfurgarli-Bayramov, Baku 1 998.
14 ...g4 1 5 �b5 �8 1 6 1lfe1?1
White should have played 1 6 .l\.g3 lieS 17
.i.eS; after l 7...g.3!? 1 8 .l\.xg3 ll'lhS 1 9 .i.f2
125
The Petroff Defence
/l)f4 Black has brood play for his pawn.
1 6 ...gxf3 1 7 ltld61?
White is trying to complicate matters; after
1 7 gxO /l)h5 18 �hl 1i'd2 Black is in con­
trol.
1 7 . . .cxd6 18 J..a& b6?!
This makes no sense - Black probably
overlooked the note to White's 20th move.
After 18 ....i.c6 1 9 .i.xb7 .i.a4 (even
1 9....ixb7 20 c6 Axel+ 21 llxel l:ld7 22
cxd7 �xd7 23 1Wc3 a6 is comfortable for
Black) 20 1i'b4 (20 ..xa4? /l)g4 and Black
wins) 20 ..a5 21 'irb6 li)d7 22 1Wxa5 llxel+
. Anand-Hubner
23 l:txe1 Wxb7 24 11fxd8 dxc5 Black has the
advantage.
1 9 cxb& i.e&
20 bxa7+
Much stronger is 20 1Wc3t lbg4 21 h4 tbe5
22 dxe5 dxe5 23 llact and Black is under
attack.
126
20...�xa7
After 20 ..�a8? 21 1i'b6! Black cannot de­
.
fend himself.
21 1ta3
21 .. .i.b7
21 ...1la8? would be met by 22 1Wa51 .i.b7
23 llact!! l:lxct+ 24 ..i.fl+ 'it>b8 25 'ii'c7+
<tla7 26 llc3 and White wins.
22 .txb7 + �b7 23 111b4 + � 24 'lra5 +
�7 25 M»5 + �7 26 '1ra5 + %-%
Gan1e 53
Dortmund 1992
1 e4 e5 2 ltlf3 ltlf& 3 d4 ltlxe4 4 J..d3 d5
5 lZ':ixe5 lbd7 6 ltlxd7 J..xd7 7 0-0 1Wh4 8
c4 0-0-0 9 c5 g61?
This move keeps the g5-square vacant as a
possible retreat square for the knight.
 3 d4: The Main Lin e

10 c!i.:lc3 .i.gS+ � f8 1 7 1Wc2 according to Rozcntalis)

I 0 f3?! ll)g5 1 1 J.e3 .*.g7 1 2 g3 'ilh5 gives 14 lL!c6+1 (the move Rozentalis overlooked?)
Black considCfllble counterplay. 1 4...Wc8 15 lL!xa7+ Wb8 1 6 lLlc6+ Wc8 1 7 f3
10 .tg7 , , g3
.•• Ld4 18 .i.c3 llxd3 19 ..xd3 lllxg3 20 .i.f41
Or: 1 -0 lvanchuk-Rozentalis, Debrecen 1 992.
a) 1 1 llk2 :hetl

1 2 i.e3
and now: White has alw tried 1 2 lL!xe4 dxe4 1 3
a l ) 1 2 a4 /.l)g5!? 1 3 :a3 h6! 1 4 J.c2 (1 4 .ixe4 i.h3 1 4 1Wb3 11fa6 (or 1 4...c6!? 1 5 d5
i.c3 �6 1 5 1i'd2 with unclear play is proba­ cxd5 16 J.xdS Ld5 17 "tlfxd5 :dB 18 "tlfb3
hly better) 14 ...llxe2! 1 5 1i'xe2 �6 1 6 1Wd1 9c6 19 f3 J.xfl 20 Wxfl :d 5 with unclear
li:)xd4 and Black had enonnous play for the play) 1 5 :et :xd4 16 .i.f4 J.e6 17 9f3 c6
exchange, I vanchuk-Rozentalis, Dcbrecen 1 8 .id6 l:.d2 19 Wf4 l:r.d4 and the position
1 992. was very messy, Sax-Skembris, Burgas 1992
n2) 1 2 .i.e3 J.h6 1 3 .i.xh6 9xh6 1 4 1Wct
and White might have a very slight edge.
b) 1 1 '*.e3 lLixc5 1 2 g3 1i'h3 13 .i.e2 (1 3
dxcS d4 14 lLld5 dxe3 1 5 fxe3 - 15 c6? exf2+
I (l llxf2 .i.d4 17 cxd7+ Ld7 is bad for
White - 1 5...-*.e5 1 6 Wc2 �b8 leads to an
unclear position) 13 ...h5 14 lL!xd5 J.a4! 1 5
b.l .D.xd5 1 6 bxa4 :es (but not l 6 ...:hd8?!
17 i.f3 llxd4 1 8 Wc21 with a clear edge for
White, J.Polgar-Skembris, Moscow 1 994) 17
'i'c2 lL!e4 18 :ret Wd7 and the position is
very complex.
1 1 ..f&
...

1 1 ...'ilh3?! was cleverly refuted by board 1 2 .tf5?

...

nne of the 2004 Olympiad Champions, 1 2...lllg5 is necessary. Now 1 3 lL!xd5?

Ukraine: 12 lL!xdS J.g4 1 3 lLle7+! (not 1 3
'i'b3?! i.f3 1 4 lLif4 lL!g51 and the position is
less clear - Ivanchuk) 1 3 ...Wb8? (Black
should settle for a bad position after
LL�d7 14 1Wa4+ �c7 1 5 .i.xe4 :xd4 1 6
lllh3+ 14 Wg2 .i.c6 15 .g4+ �b8 1 6 Wxh3
:xd5! gives Black a clear advantage for Black
(Anand). Instead White plays 13 f4 when
Black has the foUowing tries:
a) 1 3..."ti'e6?1 1 4 l:r.c1 lLle4 1 5 J.xe4 clxe4

127
The Petroff Defence
16 dS 1i'a6 17 .i.c.I4 with a considerable ad­
vantage (Yusupov).
b) 1 3 ...1i'e7?! 14 lle1 �6 1 5 .i.ft ! (a nice
refutation of Alexander's idea; I S c!LlxdS 1We8
gives Black enough counterplay, Sasha's
preparation running 16 f5 gxf5 1 7 .i.xf5 .i.a4
18 1i'g4 llxdS 1 9 .i.xe6+ fxe6 20 1i'xg7 l:g8)
1S...c6 16 'i'a4 Wb8 17 f5! (another strong
move from the strongest player of our little
region) 17 ...gxf5 1 8 .i.f4+ WaS 19 .lle3! c!LlxcS
(19 ...'i'f6 would fail to 20 'i'xa7+!! �xa7 21
c!LlbS+! cxb5 22 ltt3 mate) 20 .llxe7 c!Llxa4 21
c!Llxa4 .i.xd4+ 22 .i.e3 .i.xe3+ 23 :Xe3 and
White went on to win with his extra piece,
Delikov-Raetsky, Voronezh 2004.
c) 1 3...c!Llh3+! (the best move) 1 4 �g2
.llhe8 1 5 9d2 (Leko-S.Farago, Dudapest
1 993). Now after 1 S ....i.g4 1 6 b4 hS Black
has counterplay, but it's difficult to teU
whether it is enough for equality.
1 3 lilb51 J.h3
Black is in a bad way. After 1 3... a6 14
c!Llxc7! �xc7 (14...9c6 1 5 c!LlaB! is horrible
too) 1 S .i.f4+ �d7 1 6 .i.e5 ...e6 17 .i.xg7
.llhg8 18 .i.eS c!LlxcS 19 .i.xf5 ...xf5 20 g4
9e6 21 .lle 1 �4 22 f3 c!Lld6 23 1ib3 White
won in Har Zvi-Lev, Tel Aviv 199S.
14 ltlxa7 + �8 1 5 ltlb5 J.xf1 1 6 J.xf1
White has a clear edge despite the missing
exchange - Black's king is simply too weak.
1 &. . .:heS 1 7 ita4 ita& 18 itb4 Was
Or l B ...1i'c6 19 .i.f4 :C7 20 WaS and
White wins.
128
19 a4 g5 20 a5
20 f3? would aUow 20...c!Llxg3 21 .i.xg5
c!Llxft 22 .i.xd8 .llxd8 23 Wxft 1i'a6 when:
BL'lck suddenly has counterplay.
20 c& 21 c!ilc3 ltlxc3 22 bxc3 h& 23 a&
•.•
Now White is winning.
23 . . .f5 24 J.h3 l:lfS 25 a7+ �cS 26
'8b1 g4 27 J.f1 �7 28 ..i.d3 h& 29
J.f4 l:tf7 30 itc2 ..i.fS 31 'lle2+ 1 -0
Game 54
Elizarov-Raetsky
Belorechensk 1989
1 e4 e5 2 lilf3 lilt& 3 d4 c!l:lxe4 4 J.d3 d5
5 ltlxe5 ltld7 6 ltlxd7 J.xd7 7 0-0 itf617
This is probably a little bit better for
White, but it's stiU playable.
8 J.xe4
Probably the critical move, but White has­
some alternatives:
a) 8 .i.e3!? is also a decent option: 8 ... .i.d6.
9 c4 c6 10 cxd5 (1 0 llk3 c!Llxc3 l 1 bxc3 dxc4
1 2 Lc4 0-0 13 9hs .llfe8 is fine for Black)
l O...cxdS 1 1 Wb3 .i.c6 12 .i.xc4 dxc4 1 3
llk3 0-0 14 dS .i.d7 1 S c!Llxe4 (IS 11'xb7?1
1We5 1 6 g3 .i.h3 with an attack on the light
S<JUares) 1S....i.xh2+ 16 Wxh2 1i'h4+ 1 7 �gl
1i'xe4 1 8 1i'xb7 .i.e6 1 9 .llfd1 (Macieja­
Rowson, Duisburg 1 992). Now after
19 . ...11fb8 20 9c6 ll.d8 21 lld4 .i.xdS 22
.llxe4 .i.xc6 23 .lle7 White might have
slightly better chances in tllis endgame.
 3 d4: The Main Lin e

b) 8 c4 is very aggressive, but not clear at
nil. Black plays 8...1Wxd4
'ife3 lld6 19 .xa7 l:txc6 20 .aS+ �c7 21
llab 1 .C.b6 22 :Xb6 �b6 23 .C.b1+ �c5 24
1i'a3+ �d4 25 .e3+ �e5 26 i.xg6+ 1-0
Kotronias-Atalik, Pucarevo 1 987.
8 dxe4 9�3 0-0-0 10 lllxe4 1tg6
...

and now:
b1) 9 ltk3 �c5 1 0 Xlel+ i.e6?1 (1 0...�d8
iN necessary; after 1 1 i.e2 1Wxd1 1 2 l:txdt
dxc4 13 i.xc4 White has compensation, but
nothing more) 1 1 .C.xe6+! �xe6 12 cxd5 i.c5
1 3 i.e3 •e5 1 4 dxe6 i.xe3 1 5 i.bS+ �f8 1 6
.f3 .xe6 17 fx.e3 c 6 1 8 .L4 �18 1 9 i.b3
'fkc7 20 .C.ft f6 21 llk2 and Wh_ite had a clear
nlgc in the game Tseshkovsky-Bareev, Kiev,
1 1)R6.
b2) 9 cxd5 is probably less dangerous:
11...0-0-0 1 0 .c2?1 (10 i.xe4!? .xe4 1 1 �c3
'fkh4 1 2 i.e3 i.d6 13 g3 with unclear play
was necessary) 10 ... �5 1 1 i.c4 i.d6 1 2
1 1 f3
The alternative 1 1 /i)g5 is probably best
met by 1 1...£6, for example 1 2 �f3 i.h3!?
(1 2... h5 13 c4 h4 14 l&1 h3 15 g3 i.g4 16 f3
i.cS 1 7 i.e3 llhe8 1 8 i.f2 i.f5 1 9 d5 i.d6
20 1i'd2 gives White a slight edge, S7jebcrt­
Raetsky, CappeUe Ia Grande 1 999) 1 3 �e1
i.c5 14 d l:the8 15 i.e3 i.d6 with compen­
sation for the pawn.

Ac3 'ireS 1 3 g3 i. f5 1 4 'ird2 1We4 1 5 �a3

.C.hd� 1 6 llfd1 i.h3 17 i.ft i.xft 1 8 :Xfl
Q)d3 and Black dominated events in Chudi­
novskikh-Raetsky, Briansk 1995.
c) 8 � is pretty tame:
cl) 8 ....xd4 9 �xe4 dxe4 1 0 i.xe4 1Wxd1
I I llxd 1 0-0-0 12 i.g5 f6 13 i.f4 i.cS 1 4
l:c.l5 i. b6 1 5 llad1 g6 16 a3 c6 1 7 .C.Sd3 i.g4
I K :Xd8+ llxd8 1 9 .C.xd8+ �xd8 1/:�-1/z Ken­
�is- Rozentalis, Vilnius 1 984.
c2) 8.. �d 9 bxc3 0-0-0 is possible, but
.

White seems to be slightly better: 10 'irh5 11 h5

•••

'fkc6 1 1 i.d2 g6 1 2 'iff3 f6 13 l:tfel 'lff7 1 4
c4 dxc4 1 5 i.e4 c6 1 6 d5 i.e8?1 (16...i. f5 1 7
i.x£5+ gx f5 1 8 .xf5+ 'ird7 19 1i'xd7+ �xd7
211 .C.ab1 is better for White, but the game
WI 1uld still be undecided) 17 dxc6! l:txd2 1 8
Black ha.c; no path to absolute equality, for
example:
a) I 1 ...f5 1 2 �f2 i.bS 13 .C.e1 .i.d6 14
�h3 .C.dc8 1 5 i.f4 :Xel+ 1 6 1i'xe1 Xle8 17
'ifg3 1i'xg3 1 8 hxg3 is a bit better for White.

129
The Petroff Defence
b) 11 ...-*.5 12 c3 h5 (or 12...i.xe4 1 3 fxe4
...xe4 1 4 :Xf7 i.d6 1 5 .*.g5 lldf8 16 llxf8+
ltx£8 1 7 .d2 and we prefer White due to the
extra pawn) 13 'ifa4 'ii1b8 14 .*.f4 h4 1 5 llae1
h3 16 g4 i.d7 (Klovans-Ro?.entalis, USSR
1 985). Now after 1 7 'ifc4 llc8 18 �5 i.d6
19 .*.xd6 ..xg5 20 �e7 i.b5 21 .*.xg5 .*.xc4
22 llf2 .*.xa2 Black is worse but he has &JOOd
drawing chances in the endgame.
1 2 ..tg5
12 .*.f4 is also strong after 1 2..h4 13 'ifd3
and now:
a) 1 3 ... 'ifb6 14 a4 .i.e6 1 S llfd t h3 1 6 aS
..c6 1 7 a6 b6 1 8 c4 f5 19 l'Df2 (dtis is
stronger than 19 dS?l fxe4 20 ..xe4, Raetsky­
Kvt:inys, correspondence 1 987) 19 ...hxg2 20
dS ..cS 21 Wxg2 and White is clearly better.
b) 1 3...h3 14 g4 'fib6 15 a4 i.e6 1 6 llfd1
and here Black must choose between dte
following options:
b1) 16 ... 5 17 aS 'ifc6 (or 17 ... fxe4 18 axb6
exd3 1 9 bxa7 �d7 20 a8'ii' lLa8 21 llxa8
dxc2 22 llct i.b3 23 .*.eS c5 24 lL.3 c4 25
l:la8 and the advantage is definitely widt
White) 1 8 gxf5 i.xf5 19 a6 b6 20 .i.g3 ..g6
21 'ii'c4 lld7 22 d5 Le4 23 'ii'xc4 l'lh6 24
lla4 .i.cS+ 25 'ii1f1 and White was clearly
better in Raetsky-Sivets, correspondence. In
fact, following 2S...:C7? 26 d6! it was already
rime to resign.
b2) 16 ...a6?! 17 aS 'ii'xb2?! (17 ...'ii'c6 is bet­
ter, even dtough after 1 8 ...c3 'ii'xc3 1 9
.!Dxc3 .i.b4 20 ltk4 Black has no compen.<�a-
130
lion for dte pawn) t 8 .!Dc3 'ifb4 19 l:la4 'ii'e7
20 d5 g5 21 .ig3 f5 22 gxf5 .tf7 23 ltc4 with
a dose-to-winning advantage for White,
A.Ivanov-Kochicv, Kosttoma 1 985.
b3) 16 ...a5!? seems to be necessary, al­
though after 1 7 'lrb5!? llxd4 1 8 1ixb6
l'lxdl+ 19 l:lxd1 cxb6 20 .i.e5 dte endgame is
more pleasant for White.
12 f6 1 3 ..th4 �
•••
Or 1 3...•£7 14 1t'd2 g5 1 5 .tf2 h4 16 d5
h3 1 7 g3 and White is better.
14 c4 g5 1 5 ..tf2 f5 16 .!i)c3 h4 1 7 d5 g4
1 8 ..td4 .i.d6
White cannot be allowed to play 1 9 f4, for
example 1 8...:g8 19 f4 .td6 20 ...d2 c5 21
dxc6 .ixc6 22 t'Dd5 and White has all dte
chances.
1 9 fxg4
1 9 .ixh8? ...e3+ 20 l'lf2 g3 21 .i.d4 gx£2+
22 �ft 1t'f4 is bad for White.
1 9 :hg81
••.
Not 19 ...fxg4 20 .txh8 (20 l'lf6!? 1Whs 21
c5 .ieS 22 c61 is equally strong) 20...'ifxh8 21
llk4 .ixh2+ 22 �xh2 'ifeS+ 23 �g1 Wxe4
24 Wd2 and White the advantage, and a sub­
stantial one at dtat.
20 gxf5?
For some reason White decides to assist
Black's attack with his next two moves. Here
he should play 20 c51 .i.cS 21 .i.xe5 11fc3+ 22
l'lf2 'irxe5 23 'iffl fxg4 24 c6 bxc6 25 ...a6+
�b8 26 dxc6 .i.c8 and Black is only a bit
worse in this messy position.
 3 d4: The Main L ine

23 g3 is met strongly by 23....txg3! 24

hxg3 :xg3+ 25 �h2 :gZ+ 26 �h 1 :Xd5! 27
cxd5 ..d6 with mate to come.
23 . . .11rh4 24 Wh1 llxg4

20. . . c5!
131ack grabs his chance instantly. l11e key
Idea is 21 .i.f2 .f41 22 .i.g3 ..g5 23 lbe4
Wg3+ and the queen's dance wins a piece.
21l... .i.xh2+1? 21 Wxh2 .d6+ 22 Wh1 ll.xg2! 25 ..i.f6
h:;tds to a draw as 23 .i.g1 ?? Wg3! 24 ll.f3 After 25 .i.f2 ..g5 26 11t'f3 :lxc4 White is
l:.xgt+ 25 ..xgt 1i'xf3+ 26 ..g2 ..xf5 is finished.
clc;trly better for BL1ck. 25 ...1lg1 + 1 26 Wxg1
21 dxc67 Or 26 :xgt ..c4+.
A terrible mistake, and suddenly Black has 26 ..tc5+ 27 �1 •xc4 0-1
.••

ufantastic attack. After 21 lbe4! cxd4 22

Wxd4 :dtll the game would have remained Game 55
unclear. Tiviakov-Miles
21 ...i.xc6 22 lilll6 l.inam 1998

1 e4 e5 2 ltlf3 ltlf6 3 cl4 ltlxe4 4 ..i.d3 d5

5 ltlxe5 ttJd7 6 ltlxd7 ..i.xd7 7 0-0 ..i.e7
11tis is a bit passive, and so is 7. ..lt'lf6, for
example:
a) 8 .i.g5 i.e? 9 c3 c6 1 0 lbd2 0-0 11 ..c2
h6 12 .th4 li)h5 13 .i.xe7 Wxe7 14 ll.fet
..d6 15 li)f3 li)f4 16 .th7+ �h8 17 .tf5
.i.xf5 1 8 Wxf5 was very slightly better for
White in GcUer-Smyslov, Moscow 1 981,
although here the players agreed a draw.
b) S ltct+ .te7 9 .tf4 (or 9 Wc2 .t� 10
f4 g6 1 1 li)c3 'ird7 12 f5!? gxf5 - t2 ... Lf5?
22 . . .h31 13 .txf5 gxf5 14 11t'e5 would of course be
The right way to attack. After bad - 13 .i.f4 lbc4 14lbb5 .i.d8 with un­
.txh2+?1 23 �xh2 ll.xd5 24 cxd5 ..d6+
22... clear play) 9....i.g4 (9...0-0? is met strongly by
2S Wh 1 l:.xg2! 26 .i.c5! Wxe5 27 Wxg2 10 .i.xc7 Wxc7 1 1 :xe7 '1Vb6 12 li)c3 ..xb2
Wg3+ all Black has is a draw. 1 3 lbxd5 li)xd5 14 l:lxd7 li)c3 15 ...ct
23 g4 ..xct+ 16 llxctlbxa2 17 l:la1 li)b4 18 .i.e4

131
The Pe tro ff Defence

and Black is struggling, possibly in vain) t 0 hxgS. Now 19 d5!? should en.or;ure some ad­
'il'd2 0-0 t t li:lc3 c6 12 o!ik2 ...d7 13 lbg3 vantage. Black has to be careful, for example
li:lh5 1 4 .!fuh5 hh5 1 5 .i.f5 ...d8 (not 1 9..Jife8 20 lZxe8+ lZxe8 21 dxc6 .ixc6 22
1 5......xf5?! 16 :Xc7 lZfc8 17 lZact lZxc7 1 8 h41 1i:lb5 23 li:lxb5 .i.xb5 24 hxgS with a clear
:Xe7 with a clear edge, a s t 8...b6? t 9 llc5 edge for White.
...g6 20 llg5 is winning) 1 6 :C3 i.g6 17 10 c5 0-0 1 1 .i.f4 c& 1 2 b4
Lg6 fxg6 1 8 1lac1 lZf7 1 9 g3 with a slight
edge for White, Palac-Zaja, Pula 2000.
8 c4
8 lZe1 doesn't really work: 8...0-0 (8...1i:lf6
transposes to 7...lbf6 above) 9 .i.xe4 dxe4 10
D.x'--4 .tf6 1 1 :C t i.c6 1 2 .i.c3 ...d5 1 3 0
l:tfe8 1 4 �3 (14 c3?! i.M! 1 5 .i.f2 .i.xf2+
1 6 Wxf2 :Xet 17 ...xe1 :Cs 18 ...d1 .tb5
with a strong initiative) 14......d6 1 5 Wh1
ltad8 1 6 Wd2 i.xd4 1 7 .i.xd4 1i'xd4 1 8
:XeS+ .ixe8 19 ...xd4 lZxd4 and Black is
ever-so-slightly better (Skatchkov).

It's advisable for White to get going here,

Svidler-Yusupov, Ka7.an., 1997 continued
more slowly with 1 2 h3 lZe8 13 :Ct o!L'ld7 1 4
•d2 (or 1 4 b4 .i.g5 1 5 .i.g3 •f6 and Black
is okay) 14 ....if6 1 5 .i.c2 lt'lf8 16 D.e2 b6 1 7
Le 1 bxc5 1 8 dxcS .i.d7 1 9 lZxe8 .*.xeS 20
b4 li:le6 21 .ld6 g6 22 f4. Here Svidler gives
the line 22... a51 23 f5 b>Xf5 24 .txf5 .ig7 25
.ixe6 fxe6 26 llxe6 'ii'h4 27 o!be2 and as­
sesses the position as unclear.
1 2 1Wd7 1 3 1Wc2 g6
••.

8 lLif6
•.•

Or 8 ...c6 9 cxd5 cxd5 1 0 o!L'lc3 lt'lf6

(IO...Ii:lxc3 1 1 bxc3 Wc7 12 :Ct is good for
white; 12 ......xc3 1 3 .td2 ...xd3? 1 4 lZxe7+! is
even winning!) 1 1 .i.g5 .i.c6 12 .i.c2 (1 2 lZet
0-0 1 3 •n looks even stronger) 12 .. 0-0 1 3
.

n :es 1 4 •d3 g6 1 5 :ret •d6 t 6 •d2

li:lh5 17 .i.xc7 lZxe7 1 8 lZxe7 ...xe7 1 9 lZc1
'iff6 20 lZc5 and White had an edge in Alek­
seev-Skatchkov, St Petersburg 2002.
9 ll:lc3 .i.e&
Zapata-Perdomo, Colombia 1 998 contin­
ued 9 ...dxc4 10 .i.xc4 0-0 1 1 :Ct c6 1 2 .*.g5 14 J:lfe1
h6 1 3 .th4 lDd5 1 4 .txe7 li:lxe7 1 5 'ilrh5 14 b51? :res (14...o!L'lh5 1 5 .tM {;jjg7 1 6
li:lf5 t 6 l:tad I �6 t 7 .i.b3 1i'g5 18 Wxg5 a4 .if6 17 �2 also l1x>ks a touch better for

132

WhiLe) 1 5 a4 li)hS 1 6 �e3 gives White a
Hlight plus. Now in J.Polgar-Van der Sterrcn,
Wijk aan Zce 1 998 Black went astray with
I Cl ...i.d8?! 1 7 aS a6 1 8 bxa6 bxa6 19 li)a4!
. 34...1te7 35 Lc6 .i.xd4 36 lbb6 .i.c3 37
i.xa5 20 li)b6 .i.xb6 21 cxb6 i.f5 22 i.xf5
.xfS 23 l:lfcl Vxc2 24 :Xc2 aS 2S l:lxc6,
�iving White a clear plus.
14 . . .lbh5 1 5 .i.e5
l S .i.h6 l:lfe8 16 bS .i.f6 17 bxc6 bxc6 1 8
.:�4 li)g7 with the idea o f ...ll)f5 gives Black
�ufficiem countcrplay.
1 5...f6 1 6 .i.g3 lbxg3 1 7 hxg3 i.f7 1 8
b5 1tfe8
18... f5 19 l:labt .i.f6 20 li)e2 l:.fe8, with
chances for both sides, is probably a better
c 11>tjon.
19 a4
1 9...i.d8?!
19 ... f5 is stiU correct.
20 ltxe8+ 'lfxe8 21 aS a6 22 bxa6 bxa6
23 'lfd2 f5 24 ll:la4 .i.f6 25 lte1 '1Vd8 26
ll'lb6
Or 26 ..tn ..tg7 27 l:tbl Vf6 28 l:tb4,
when White is more comformble.
26 ...1ta7 27 ..if1 �g7 28 ltb1 h5 29
11rc3 'lfha 30 ltb4 �7
Black should probably play 30... h41? 31
1,rxh4 Vxh4 32 g3 Vhs, although White does
keep an edge.
31 'lfd2 'lfe8 32 lba4 '1Vd8 33 ltb6 'lfh8?!
33...Vc8 is a better option, although after
14 1ib4 White has a strong pressure.
34 La6
3 d4: The Main L ine
Or 34 :Xc6 i.xd4 3S :Xa6 l:lxa6 36
.i.xa6 .i.e8 37 li)b6 1WeS and the endgame is
bad for Black.
1i'g5
37 'fre57
•••
37...l:le1 38 li)xdS .i.xa5 39 �3 is obvi­
ously nice for White, but at least it's not over!
Now, though, White strikes a winning blow.
38 lbxd51 1tb7
Or 38...'.xd5 39 Vxe7.
39 lbxc3 11'xc3 40 11'f6 11'xa5 41 ltc8
.i.g8 42 i.c4 1 -0
Game 56
Sorokin-Raetsky
Kra.modar 1984
1 982, Vorone1.h State University, De­
partment of Mathematics, 23 years before the
publication of J>ehvjf Defmce by Everyman
Chess. Wasting no time during a physics
lecture, the future author makes an impormm
discovery on his pocket chess computer (a
version of the computer software Fritz 0.03):
the Petroff Defence is stiU a]jvc!
He has just found a beautiful combina­
tional refutation of ECO'r critical assessment.
The bible of that time insists on '14 �12
with a slight edge' as in the game Mortensen­
Borik played at the t 980 Chess Olympiad,
but his intuition whispers 'it's not that sim­
ple' and 'Eureka!'...
133
The Petroff Defence

We had to wait for two years until my uni­

versity analysis was included into all books
on the Petroff Defence. My 'co-author' was
the future GM Sorokin.
1 e4 e5 2 o!i:lf3 o!i:lf6 3 d4 �xe4 4 -*.d3 d5
5 o!Lixe5 o!Lid7 6 ••2

and now:
a) 9 WI? 1fd4 1 0 'irbS+ c6 1 1 ..xb7
ltc8 1 2 .ie3 11'xe5 t 3 .xa7 .ib4 1 4 'iVd4 ·

and White was a bit better in Suetin-Radulov,

Athens 1 984.
b) 9 11rxe4 is less clear: 9._.id5 tO 'irg4 hS
This is not really considered dangerous 1 1 1Wh3 'ire7 12 f4 ..e6 t3 ..xe6+ fxe6 1 4
now. � f2 (1 4 0-0 .icS+ 1 5 �h1 0-0-0 1 6 lbd2
6 .1tlxe5
•. :thffi gives Black excellent play) 1 4...g5!? 1 5
Another option is 6 ...1Ve7 7 .ixe4 (I fxg5 ( 1 5 �c3 .icS+ 1 6 .ie3 .ixe3+ 17 �xe3
lDxf7!? ITansposes to 6 �xf7 'iVe7 7 1t'e2) .*.xg2 18 lthgt gxf4+ 19 �f4 llg8 looks
7...dxe4 8 .if4 lbxe5 9 .ixe5 and now: k.-vel) t 5.. ic5+ 1 6 �g3 (1 6 .ie3 0-0+ 17
a) 9 .. f6 is weak: 10 .ig3 f5 1 1 �3 c6 1 2
. �c2 .ixe3 18 �xe3 ltfS 1 9 lbc3 :XeS+ 20
0-0-0 g6 (even worse is 1 2...1Vg5+?! 1 3 �b1 �f4 ltfS+ 21 �e3 ltxg5 and Black is no
.ib4? 14 �bS f4?1 - t 4_.cxb5 1 5 1t'xb5+ worse) 16 ... h4+ 1 7 �h3 .ie4 and Black has
�f7 1 6 1Vxb4 f4 1 7 h4 is more resilient al­ good counrerplay.
though White keeps an advantage - 1 5 �7+
�ffi 1 6 1t'xe4 and Black resigned in Dcly­
Malich, Pees 1 964) 1 3 .ie5 .ih6+ t 4 �b 1
0-0 1 5 h4 and White clearly has all the fun.
b) 9....if5 10 �3 0-0-0 1 t 0-0-0 'iVe6 12
'iVc3 h5 1 3 h3 f6 1 4 .ih2 g6 1 5 �b1 .ih6 1 6
1Vg3 lth7 17 lthe1 'irb6 18 ltlxe4 :Xd4 1 9
ltxd4 1Vxd4 with level chances, Karpov­
Hort, Amsterdam 1 980. More energetic is t 4
.ixc7!? �xc7 15 d5 11rc5 16 1Vxa7 .id6 1 7
ltlbS+ �d7 (17...�c8 18 ltd4 is dangerous)
1 8 1Vxb7+ �e8 19 exg7 lth7 20 ltlxd6+
'iVxd6 21 Wg3 1Vxg3 22 fxg3 with a very
diffiCult endgame for both players. s •xe4 �e6
7 �xe4 dxe4 8..id6 9 dxe5 'fle7 to 0-0 0-0 1 1 .if4
A weaker option is the move 7....ie6?1 8 ltc8 12 :kt f6 13 �2 fxe5 14 .ig3 (or 14
dxe5 dxe4 .ixc5 .ixe5 1 5 'iVdS+ �h8 1 6 f4 .ie6 17

134

..xeS 1tb4 with a slight edge for White)
1 4... .i.d7 1 S lbf3 .i.c6 16 1i'c4+ Wt7 17
..xt7+ �xt7 1 8 lbxe5+ .i.xeS 19 .i.xeS and
White was a pawn up in Suetin-Bcx, Biel
1 99S.
9 1rxe5 ..d7
10 .te3
White has two main alternatives here:
a) to lbc3 0-0-0 1 1 .i.e3 .i.b4 12 0-0 f6 t 3
1Vg3?! (1 3 1fe4 AfS 1 4 Wf3 .i.xc2 1 5 :act
11rf5, with equal chances, is better) t 3.. ..i.xc3
14 bxc3 h5 1 5 h4 g5 1 6 f3 (1 6 hxg5 h4 1 7
1Ihz h3 1 8 g3 fxgS 1 9 .i.xg5 lldg8 20 f4
1fc6 also gives Black a strong attack)
1 6....1ldg8 17 llt2 Wc6 18 .i.d2 g4 19 f4 .i.c4
and Black has an attack. The game Karpov­
l .arsen, Ttlburg 1 980 continued 20 dS .i.xdS
21 f5 :C:8 22 a3 lle4 2.1 %let l:the8 24 llxe4
Axc4 25 �h2 1i'c5 and Black was clearly
better.
b) 1 0 0-0 0-0-0 1 1 .i.e3 .i.b4 1 2 c3 f6 13
1fg3 with a further split:
bl) 1 3....i.d6 14 .tf4 .i.f8 1 S .d3 gS 16
.i.c3 h5 17 lbd2 (we prefer 1 7 c41? .i.fS 18
ti'd2 h4 19 dS g4 20 lbc3 h3 21 g3 and
White is slighdy better) 1 7 ... h4 1 8 f3 .J:lg8 19
c4 f5 20 f4 gxf4 21 llxf4 Wg7 with an un­
clear position, Magcm Badals-Macieja, Ba­
lumi 1 999.
b2) 1 3 ....i.e7 t4 lbd2 hS 1 5 f3 h4 1 6 ..t2
hJ ( 1 6...g5 1 7 c4 is better for White - Yusu­
pov) 1 7 g.l 'Wb5 1 8 b3 llhe8 1 9 c4 •aS 20
n3 'li'c3 21 dS .i.fS with chances for both
3 d4: The Msin Lin e
sides, Krakops-Raetsky, Apolda 1994.
1 0 ....tb4+
10...0-0-0 1 1 ..a5
gives Black a wide, but unsatisfactory,
choice:
a) t t ....i.d5 1 2 lbc3 .i.xg2 1 3 llg1 b6 1 4
1i'xa7 .i.b7 1 S 0-0-0 .i.d6 1 6 h4 f6 17 d5
l:the8 18 lld3 h6 19 .i.xb6! cxb6 20 lba4
.i.c7 21 d6! .xd6 22 llxd6 llxd6 23 b3 and
Black was much worse, Udalov-Raetsky,
correspondence 1982.
b) 1 t ...Wc6 1 2 lbc3 'lt'xg2 Qnstead of
12 ...b6 13 1fa6+ �b8 14 lbb5 .i.c4 1 5 a4
.i.b4+ 1 6 c3 .i.d6 17 'lt'xa7+ �c8 18 0-0-0
1i'xg2 19 d51 .i.xdS 20 llhgt 1i'xh2 21 aS
with a raging attack, Hort-Short, Bundesliga
1 986) 1 3 0-0-0 a6 1 4 dS .i.h31? 1 S .i.f4 lld7
1 6 llhet .i.c7 1 7 .i.g3 and White's position
looks preferable.
c) 1 1 ...�b81? is probably best: 1 2 lbc3 b6
1 3 'lfa6 .i.b4 (13...Wc6 1 4 lbb5 1l'b7 1 5
1Wxb7+ �xb7 16 0-0-0 is better for White) 1 4
0-0 .i.xc3 15 bxc3 Wc6 and Black has some
compensation for the pawn.
1 1 c3 i.d6 1 2 1ta5
White has other options, but none that is
comforting:
a) 1 2 � 0-0 1 3 0-0 .i.g4 14 Wh4 llfe8
1 5 tnd2 f5 16 h.l .i.e2 17 life 1 'Wb5 1 8 c4?!
.i.xc4 1 9 b3 .i.dS 20 .i.f4 llxet+ 21 llxe1
1i'd3 22 .i.xd6 Wxd2 23 lieS was agreed
drawn, Movsesian-Haba, Pardubice 1998,
although of course Black is to be preferred
135
The Petroff Defence
here. Instead White should play 18 b3 .i.d3
1 9 c4 ..d7, although Black retains excellent
compensation.
b) 1 2 ..xg7 0-0-0 1 3 �2 was played in
00-Khalifman, Sochi, 1 984. Now after
1 3 .. .'Vc6!? 1 4 f3 :hg8 1 5 1fxh7 :Xg2 Black
has excellent play for the pawns.
1 2...•c6 13 f3
1 3 0-0?! �d5 gives Black a clear edge.
One game continued 14 f3? b6 1 5 1fa6 .i.c4
1 6 d5 1fxd5 (16....i.xh2+ 17 Wf2 .i.g3+ 1 8
Wgl 1fxd5 19 .a4+ b S 20 .c2 0-0 also
wins for Black) 1 7 ..a4+ bS 1 8 ..d1 •es 1 9
W f2 'irxh2 20 f4 .i.xfl 0- 1 Klinger-Wolff,
Baguio 1 987.
1 3 .i.d5 14 llXI2 0-0 1 5 0-0
..•
1 5 Wf2 b6 1 6 .a6 f5 is also unappcrising
for White.
1 5 ...:fe8
1 6 .i.g5?
136
Other moves played here include:
a) 16 :ret b6 1 7 •a6 .i.xh2+ 1 8 Wxh2
:xe3 1 9 Wgt 1le6 20 "lffl :ea 21 :xe3
..xc3+ 22 Wf2 Wd3 and Black was slightly
better, Barcenilla-Ye Rongguang, Beijing
1 992.
b) 1 6 .i.f2 b6 1 7 1fa6 Lh2+ 1 8 Wxh2
1fh6+ 1 9 Wgl 1fxd2 looks good for Black,
bast..-d on 20 c4?! .i.xf3! when White is in
trouble.
18 Jie2 1 7 c4
.•
1 7 :ret is met by 1 7 ..b6 t R 1fa6 :Xd2 1 9
.
.i.xd2 .i.c4 20 d 5 .i.xh2+ 2 1 Wh1 ..xdS and
Black will win on the kingside.
1 7 ....i.xh2+
1 8 Wh 1 ?
The last chance was with 1 8 Wxh2 1t"d6+
19 Wht "llg3 20 D.gt WxgS 21 11fxc7 (not 21
�fl Wh4+ 22 �2 .i.xf3! 23 1fxc7 .i.e4 and
Black should win) 21 ...lLd2 22 cxdS :xd4
2..'\ :ge 1 :d2 24 1fh2 ...xdS, although the
endgame a pawn down leaves few drawing
chances.
1 8 ...•g6 1 9 trxd5 .i.f4 20 g4 .i.xg5 0-1
Game 57
Dolmatov-Mamedyarov
Moscow 2002
1 e4 e5 2 ll:}f3 ll:}f6 3 d4 ll:}xe4 4 .i.d3 d5
5 ll:}xe5 lDd7 6 ll:}c3
White also has a tame option in 6 0-0
�xeS 7 dxeS �cS 8 �c3 (8 .i.e2 .i.e7 9 .i.c3

0-0 10 f4 f6!? 1 1 cxf6 l:xf6 12 c4 .le6 1 3
CKdS 'ii'xdS 1 4 lbc3 1i'xd1 1 5 l:axdl c6 1 6
:02 aS 1 7 b3 l:ff8 1 8 l: fd l l:ae8 lead to
even chances in Wedberg-Rozenailis, Vasby
2000)
and now:
a) IL.c6 (the solid choice) 9 f4 f51? 10 llk2
( 1 0 exf6?1 is weaker: 1 0...1Wxf6 1 1 f5 - 11
llc l+ .lc7 1 2 ..c2 �xd3 1 3 cxd3 �fl and
us White has no �5+ coming, Black is bet­
ter - 1 1 ...�xd3 1 2 Wxd3 .le7 13 �2 0-0 1 4
ltlg3 'ilf7 1 5 .le3 b6 and Black was slightly
hettcr, Rozentalis-Turov, Montr'->al 2001)
I O .le7 11 .le3 0-0 with chances for both
•••
sides.
b) 8 ...�xd3 9 .xd3 c6 10 �2 f!fJ 1 1 Wg3
.i.g7 1 2 �f4 0-0 (12....lxe5?! looks risky: 13
J%c 1 f6 1 4 lDd3 0-0 1 5 �xeS fxe5 1 6 l:xe5
"nd Black is weak on the dark squan.-s) 1 3
lbh5!? gxh5 1 4 .lh6 .lg4 1 5 .lxg7 �xg7 1 6
h3 •gs 1 7 f3 f5 18 exf6+ l:xf6 1 9 fxg4 l:affi
20 :Xf6 l:xf6 21 :Ct hxg4 22 hxg4 :n with
level prospects, Tiviakov-Van Wely, Leeu­
wardcn 2002.
6 lllxc3
...
Black has a couple of valid alternatives ttl
1his natural move:
a) 6 ... .tb4 wins a pawn, but is quite dan­
gerous: 7 0-0 �xc3 8 bxd .lxc3 and now:
at) 9 .la3?! is too optimistic: 9...�xc5
�Jut not 9 ... .lxa1?1 10 �c6!? bxc6 1 l Wc2+
lbc::S 1 2 .xeS+ .lc6 1 3 l:xa1 with an attack)
I U dxcS .lxa 1 1 1 •xa1 and we feel that
3 d4: The Main L in e
White hasn't quite got the most from his
position.
a2) 9 llb1 .lxd4?! (9. ..c!Dxe5, transposing
to 6...�xe5, is better) 1 0 �xd7 J.xd7! 1 1
:Xb7 0-0 ( 1 1 ....lc6? loses to 1 2 .la31 .lxb7
13 llel+ �d7 1 4 .lf5+ �c6 1 5 Wxd4 and
mate is imminent) 1 2 .lxh7+ �xh7 1 3 Wxd4
and White has a dear edge.
b) 6...�xe5 7 dxeS .lb4 (7...�4?! 8
.lxe4 dxe4 9 �5 'ild8 10 .lgS 'ilc.l7 1 1 e6
fxe6 12 �5+ g6 1 3 lbf6+ �f7 14 �4 gives
Black serious problems, while 7...lbc5?1 R
�xd5! �xd3+ 9 1i'xd3 J.e6 10 �f4 obvi­
ously favours White) 8 0-0
and now:
b1) 8 ... .lxc3 9 bxc3 .lc6 with a funher
branch:
bt l) 1 0 ...el!? �c5 (1 0...f5 1 1 exf6 ...xf6
1 2 .lxe4 dxe4 1 3 •xe4 0-0 14 .ta3 l:fe8 1 5
Wxb7 1Wxc3 1 6 .lb2 1t'c4 and White is only
slightly better) 1 1 .lbS+ c6 1 2 .la3 �d7 1 3
J.d3 c5 14 f4 g6 15 c4 1Wc7 16 cxd5 J.xd5
17 c4 R.e6 1 8 l:b1 with the initiative for
White, Reefat-1-lossain, Dhaka 2003.
bt2) to f4 f5 1 1 exf6 Wxf6 1 2 f5!? (12
R.xe4 dxe4 1 3 �5+ 9f7 14 'iWbS+ c6 1 5
Wb4 J.dS 16 a4 b6 was equal in Palac­
Arkhipov, Belgrade 1 988) 1 2 ... .tf7 (not
12 ...lxf5?1 1 3 J.bS+ c6 1 4 1t'xd5 l:ffi 1 5
.la3 lld8 16 9c4 and Black is under serious
attack) 1 3 J.xe4 dxe4 14 J.e3 0-0 and Black
is probably not worse in this complex posi­
tion.
137
The Petroff Defence
b2) 8...lt�xc3 9 bxc3 .i.xc3 10 llb1 0-0
(1 0...1t'e7 1 1 llb3 .i.xe5? - Sax gives
1 t ....i.b4 12 f4 with unclear play - 1 2 lle1
leaves Black in trouble; Sax-Nunn, Brussels
1985 concluded 12 ..0-0? 13 1rh5 f5 14 .i.f4
t -O) I t .i.xh7+ Wxh7 12 .d3+ Wg8 1 3
'lfxc.l d4 1 4 'iWg3 ..d7!? 1 5 c3 d3 16 lld1
11'g4 17 ..xg4 (17 llxd3 .xg3 18 llxg3 .i.fS
19 lla 1 llfd8 gives Black enough play for the
pawn) 1 7 ....i.xg4 1 8 f3 .i.c8 19 llb3 cS is
level, Roiz-D.Fridman, Pardubice 2002.
7 bxc3 c!bxe5
7 ... .i.d6 transposes to 5....i.d6 6 0-0 0-0 7
llk3 lbxc3 8 bxc3 lbd7.
8 dxe5 J.e7
Other moves:
a) 8....i.e6 9 ltb1 1l'c8 10 .i.gS h6 1 1 .i.h4
.i.cS 1 2 0-0 c6 13 Wht gS 14 .i.g3 11'd7 1 5 f4
gxf4 1 6 .i.h4 .i.e7 1 7 :Xf4 0-0-0? (Movse­
sian-Wc:.:glarz, litomysl 1995) and now 18
.i.a61 would have won after 1 8...bxa6 19 11'fl
..c7 20 1Wxa6+ 'it'd? 21 llb7. Instead Black
should play 17 ..ixh4 1 8 1lxh4 11'e7 with an
unclear position.
b) 8....i.c5 9 0-0 .i.e6 (9...1ih4?! loses a
pawn to 10 .i.bS+ c6 11 1Wxd5) 10 D.bt .i.b6
1 1 Whs h6 12 Wh1 •e7 1 3 f4 J!fl 14 1t'e2
0-0-0 1 5 a4 •cs 16 llbS 11t'c6 (or 16. ..xd..
17 .i.d2 •c6 1 8 a5 .i.cS 19 l:lfb1 with an
attack) 1 7 aS!? a6 18 axb6 axbS 1 9 bxc7 with
compensation for the exchange, Reefat­
Vakhidov, Dhaka 2003.
9 'Wh5 .i.e& 1 0 f4
138
Or 10 llb1 ..d7 1 1 .i.gS 0-0-0 12 0-0 h6
1 3 Le7 11'xe7 1 4 11'e2 (14 l:lb3 cS 1 5 l:lbS
c4 16 .i.fS llhe8 with equal) 1 4...1ic5 1 5
1id2 d4 1 6 cxd4 llxd4 17 .e3 llhd8 and
Black was no worse in Kremcnietsky-Pripis,
Moscow 1 977.
10 g6 1 1 •t3 f5 1 2 exf6
•..
1 2 llbt 'ireS 1 3 .i.e3 0-0 14 0-0 cS also
gives Black decent counterplay.
1 2 -*.xf& 1 3 0-0 0-0 1 4 J.a3
.•.
1 4 l:lb1 c5 1 5 llxb7 1ic8 16 llb1 .i.xc3
offers chances to both sides.
1 4 .i.xc3 1 5 JZad1 .i.d4+ 1 6 c;th1 c5 1 7
•.•
c3 ...a5
After 17 ....i.xc3 18 .i.xc5 l:lf6 19 .i.c4
1ia5 20 llxd5 .i.xd5 21 1ixd5+ 'it'h8 22 .i.d4
..xd5 23 .i.xf6+ .i.xf6 24 .i.xd5 the game
would end in a draw (but not 18..1tf7 19
.i.xg6 hxg6 20 11'xc3 when Black has vulner­
able dark squares).
1 8 cxd4 •••3 1 9 .i.f5! ••6 20 .i.xe&+
••e6 21 dxc5 JZad8 22 l:lfe1 •c6 23
l:le7 l:ld7
23 ..1lde8 24 l:lxc8 llxe8 25 f5 'IVxcS 26
.
fxg6 hxg6 with unclear play was a more chal­
lenging try.
24 l:lxd7 •xd7 25 l:lxd5 •a4 26 f5
•••2 27 h3 l:lxf5 28 l:ld8+ Wg7 29
•xb7+ Wh& 30 •e4 ••1 +
Or 30...1lxc5 31 •f4+ llg5 32 •f8+ 'it'hS
33 ..f3+ with perpetual check.
31 Wh2 ••5 + 32 1rxe5 :XeS 33 l:lc8 a5
34 Wg3 a4 35 Wt4 l:le1 36 :as l:lc1 %-%

Game 58
Timofeev-Raetsky
Correspondence 1982
•------•
1 e4 e5 2 Q)f3 Q)f& 3 d4 Q)xe4 4 .i.d3 d5
15 ttlxe5 Q)d7 6 it:lxf7
This sacrifice is an attempt to achieve a
draw directly from the opening, but Black
has enough resources to play on.
e �xf7
...
Also to be considered is 6 ...'1fe7!?, which
is bit risky but playable. Now we have a
a
many of variations to look atl
a) 7 lDcs is good for Black: 7 ...lDxe5 8
dxe5 •xeS 9 'lfc2 .td6 I 0 llx12 llk5 1 1
..xeS+ .i.xeS 1 2 lD£3 lDxd3+ 1 3 cxd3 .i.f6
with a better endgame due to the two bish­
c ops ami the stronger pawn structure.
b) 7 lDxh8 lDc3+ 8 �d2 lDxdl 9 J:le1
liJxf2 10 .ltxh7 (but not to :Xe7+? .ixe7 1 1
lLic3 lDf6 and Black wins) tO...lDc5!?
(IO...lDc4+ 1 1 J:lxe4 dxe4 12 .ig6+ �dK 13
liJf7+ �e8 1 4 lDd6+ 'h-'12, I.Zailliev-Karpov,
L eningrad 1966, is safer) 1 1 J:lxeS .ie6 and
now:
b1) 12 .i.g6+?! �d7 1 3 .tf7 (1 3 .i.f5?
loses to 13...'1fg5+ 14 �e2 'ifxct 15 .ixe6+
Wc6 16 .i.xcl5+ �b6 - Yusupov) 13 ...lDc4+
1 4 �e1 (14 �d3 is met by 1 4...llk5+1! 15
clxc5 .i.f5+! 16 :Xf5 'lfe4+ and 81ack wins)
1 4...'ifh4+ 1 5 g3 (Smerclon-Solomon, Gold
< :oast 1999) and now Black can play
3 d4: The Msin Line
15...'ifxh8 1 6 .i.xe6+ �d6 17 .i.xtiS (or 17
h3 :CB 1 8 .i.f4 gS 19 .if5 gxf4 20 J:lxe8
'ifxd4 21 .txe4 fxg3 and Black wins)
17...'ifxh2 1 8 .1xe4 'ifgt+ 1 9 We2 'lfxct
when White's position is critical.
b2) 1 2 .ig8! 'ffh4 (or 12...lDe4+!? 1 3 �d1
.ixg8 14 lDg6 'ifxeS 1 5 lDxe5 .id6 and
Black has some compensation for his pawn)
1 3 .if7+1 �dB 14 :Xe6 'i!fgS+ 1 5 �c2 'lfxg2
and the position is completely unclear.
c) 7 'lfe2
and now:
ct) 7...�f7?! 8 'lfhS+ �f6 (after B �e6..•
9 .ltxc4 dxe4 1 0 dS+ �d6 1 1 .tf4+ lDe5 12
llk3 the attack also seems to be very danger­
ous) 9 0-0 'iff7 1 0 Wh4+ g5 1 1 .ltxg5+1
lDxgS 1 2 f4 We6 (1 2...'1fe7 1 3 fxg5+ �g7 14
lDc3 also gives White a winning attack) 13
fxg5 'lfg7 14 lDc3 lDb6 15 J:lf6+ 'iPcl7 16
11af1 �e8 1 7 J:lf7 Wxf7 1 8 J:lxf7 �f7 19
g6+ �c6 20 'ifet+ 1 -0 Gurgenid7.e-Bcllin,
Thilisi/Sukhumi 1977.
c2) 7...'1fxf7 8 f3 lDl'lf6 9 lDd2! Qcss ener­
getic is 9 fxe4 .i.g4 1 0 'ife3 dxe4 1 1 .i.xe4
0-0-0 12 0-0 11fcl7!? 13 .i.d3 - but not 13 c3?
:Cs 14 �12 lDxc4 1 5 lDxe4 .tf5 16 .l%f4 gS
and White is on the ropes - 1 3.. Wxd4 14
.
'ifxd4 J:lxd4 1 5 .i.e3 .i.c5 with level chances)
9 ...'1fh5 10 fxe4 ..tg4 1 1 'ife3 dxe4 t 2 lDxe4
0-0-0 with a further split:
c21) 13 lDxf6 gxf6 1 4 0-0 .i.d6 1 5 g.� (or
15 h3 llhg8!? 16 'lfh6 - but not 1 6 hxg4
Wh2+ 17 �f2 'lfh4+ 1 8 �e2 J:lde8 1 9 h4
139
The Petroff Defence

:lxg4 20 J.£5+ �d8 21 .Lg4 Wxg4+ 22 .13+ Wc6 could still transpose to the game,
�d3 :xe}t 23 J.xc3 f5 and White has a lot though it looks risky with the king on f6) 9
of defending to do - 1 6..ixh3! 17 Wxh5 Wh5+ with a repetition.
:Xg2+ 1 8 �h 1 l:th2+ with a draw) 15 ....ZZ.he8
(15...Ade8 16 '1Vh6 would benefit White).
After 1 5 ...:he8 Black has enough play, for
example 1 6 Wh6 J.c5! 17 dxc5 (17 'Wxh5
.i.xd4+ 1 8 �g2 hh5 is equaQ 17 ...Wxc5+
1 8 �h 1 •d5+ and Black delivers perpetual
check.
c22) 1 3 0-0 is stronger: 1 3...lnd5 (after
13...tbxe4?! 14 .i.xe4 .i.d6 1 5 h3 l:[hf8 16
J.d2 White is simply a pawn ahead) 14 9g5
li)b4 1 5 h3 (15 tb£2 tbxd3 1 6 tbxd3 D.xd4
W"dS played in Zhao Zhong Yuan-Solomon,

Gold Coast 1 999; following 1 7 1i'xh5 J.xh5

18 .i.e3 lle4 1 9 l:tact the position looks
even) 1 5.....te2 1 6 .i.xe2 .xe2 1 7 'Wf5+ �b8
1 8 .i.f4 and we slightly prefer White.
8 1i'e2
8 .i.xe4?1 is less reliable: 8 ...dxe4 9 d5+
�e7 10 ..tgS+ tbf6 1 1 tbc3 .irs 1 2 0-0-0
�d7 1 3 Wh4 J.d6 and Black is a piece up.
A more serious option is 8 .g4+ �d6
(8...Wf7 draws) and now:
a) 9 .if4+? Wc6 lO .e6+ (on t o .i.xe4
luckily for Black he has 1 O li)f61 1 1 'il'g5 h6,
•.•

winning) 1 o....i.d6 1 1 .ixe4 tbf6 and White's

queen is trapped.
b) 9 ..txe4?l dxl.-4 10 1i'xe4 (10 .if4+ We7
1 1 1i'g3 .!£Jf6 and Black is close to winning)
10....!£Jf6 1 1 .i.f4+ Wd7 12 1i'e2 .i.b4+ 1 3 c3
lle8 1 4 ..te5 .i.d6 and Black is certainly better.
c) 9 'il'e2 11h4 10 g3 'il'e7 1 1 c4 when
White has compensation, but is it enough?
7 'tih5+ We61? 8 .i.d6 9 f3 1i'h4+
.••

The adventurous approach. Also possible

is 7 ...rile7 and now:
a) 8 1i'xd5?l is known to be bad after
8 ...tbdf6 9 1ib3 (9 'ilfc5+? �fl 1 0 .i.xe4
.i.b4+ 1 1 c3 J:lcs and Black wins) 9 ... .i.e6 1 0
Wxb7 �f7!? (Yusupov) 1 1 0-0 (11 .i.xe4
l:tb8 12 •c6 l:lb6 1 3 •as 1i'xd4 and Black
wins) 1 t ..id6l 1 2 f3 .ixh2+! 1 3 �h2
•d6+ 1 4 �g1 'ilfxd4+ 1 5 Wh2 •e5+ 1 6 �gl
tbg3 and Black has the advantage.
b) 8 .e2 �fl (8...C�d6? 9 .i.f4+ Wc6 10
.Lc4 dxe4 11 tbc3 a6 12 1i'c4+ �b6 13
.i.xc7+! and White wins - Olthof; 8...Wf6 9

140
 3 d4: The Msin L ine

10 g3? Another option is 1 8....td61? 19 cxd5 l£if6.

This is already the losing mistake - White 1 9 llf1 i.f4 20 llxf2 llxe3+ 21 Wf1 g5
must move his king:
a) 1 0 �ft .i.xh2 1 1 c4 c6 12 ltkl2 .i.e5 1 3
:xh4 li)g3+ 1 4 �f2 l£ixe2 1 5 �xe2 .i.f6 1 6
:xh7 l:txh7 1 7 .i.xh7 .i.xd4 would probably
lead to a draw.
b) 10 �d1 �fl (or to...lC!c5 1 1 fxe4
lbxd3 12 exdS+ l12 cxd3 %lf8 is a mess]
1 2 ... �d7 1 3 1fxd3 1Wg4+ 14 'lrt1 'ifxd4+ 15
..td2 with chances for both sides) 1 1 fxe4
ll'lf6 (after 1 l.. .dxe4?! 12 l:tft+ l£if6 1 3 J..c4+
rize7 14 h3 White is better) 1 2 .!0d2 dxe4 13
lC!xe4 i..,_t4 14 lCigS+ �fB 15 l£lt1 1Wh5 and
Black has good play for the pawn.
10 ...i.xg3+ 1 1 �d1 The endgame is horrible for White and, in
After 1 1 �ft .i.f4 White is dojng very correspondence chess. almost impossible to
badly. hold.
1 1 ...lDdf6 1 2 fxe4 �7 1 3 •g2 i.g4+ 22 'Dc3 dxc4 23 'De2 lle4 24 llc1
1 4 i.e2 llhe8 25 'Dxf4 gxf4 26 llxc4 c6 27 J:lc5
14 �d2 .i.f4+ 1 5 �c3 .i.h3 1 6 '1Vf3 .i.xct llxd4 28 llf5+ �g6 29 ll5xf4 llxf4 30
1 7 l::txct dxc4 18 i.c4+ �fB 1 9 'lrg3 11rxg3+ llxf4 lle5 31 llg4+ Wf6 32 l:tg2 a5 33
20 hxg3 l:te8 also gives Black a material edge. b3 l:tc5 34 �1 llg5 35 J:lf2+ �6 36
14 ...i.xe2+ 1 5 �xe2 11rxe4+ �d2 J:lh5 37 �3 b5 38 J:ld2 c5 39 llg2
Or 1 5... lC!xc4 1 6 l:tfl+ �8 1 7 ltkl2 .xh2 J:lh3+ 40 �c2 �5 41 lla2 b4 42 �2
I R 1Vxh2 i.xh2 1 9 lClxe4 l:te8 and Black �c& 43 a3 a41 44 axb4 llxb3+ 45 Wa2
should win, though it will take some work. cxb4 46 J:le4 �c5 47 J:le5+ �4 48 lla5
1 6 11rxe4 �xe4 1 7 c4 llae8 1 8 i.e3 'Df2 J:lh3 49 llxa4 Wc3 0-1

141
The Petroff Defence

Summary
In the main line Black has very much switched to 7 ....i.d6 recendy. However, the annotated
games prove that 7 ..11rh4 remains interesting and gives Black excellent counterplay after 8 c4
.

0-0-0 9 c5 g6, and especially after 9...g5. Nevertheless, the truth is that 7 ...'1fh4 leads to ex­
tremely sharp positions that are not to everyone's taste.
As Game 54 shows, 6 ltlxd7 .i.xd7 7 0-0 ..f6 allows White to win a pawn after 8 .i.xe4
dxe4 9 llk3 1i'g6 10 ltlxe4; Black has some compen.ution, but no more than that. However, if
we insert the moves 7...'1fh4 8 g3 and now 8...'1rf61 9 .i.xe4 dxe4 1 0 ltlc3 'lrg6 t 1 lLlxe4 0-0-0,
Black has a full compensation due to the weakened light squares on the kingside.
The brave 6 ltlxf7 (Game 58) has not been played much rccendy. Black should accept the
'gift' with 6...'1Pxf7 and following 7 'lfhS+ he should move forward with 7 ...We6! - Black•s
chances arc preferable in the arising complications. On the other hand, White has a plus aftet
6.....e7 7 'lrc2! 'lrxf7 8 f3 ltlf6 9 lLld2 1i'h5 1 0 fxe4 .i.g4 1 1 ..e3. The assessment of the Zait­
sev-Karpov game has changed and become more precise. At present the best line is considered
to be 6 ...'1rc7 7 ltlxh8 lLlc3+ 8 Wd2 ltlxd1 9 l:te1 ltlxf2 10 .i.xh7 ltle5 1 1 1lxc5 J.c6 1 2 .i.gS
1i'h4 l3 .i.f7+ �d8 14 :xe6 'lrg5+ 1 5 �e2 'lrxg2, which is 'a hell of a mess'.

1 e4 e5 2 lilf3 lilf6 3 d4 lilxe4 4 .tell d5 5 lilxe5 lild7 (D) 6 lilxd7

6 'lre2 - Game56; 6 lLlc3 - Game 57; 6 lLlxf7 Game 58
-

6 . . . .i.xd7 7 0-0 .td6

7...'1rf6 - Game54; 7 .ie7 - Game55
.

7...1i'h4 8 c4 0-0-0 9 c5 (D)

9...g6 - Gt1111e5J
. 9. .g5
10 J.d - Ga1nt 51
10 f3 - Game52
1 0 lLlc3 .i.g7: 1 1 g3 - Gt111e1 49; 1 1 lLle2 - Game50
B c4
8 lLlc3 Game 48
-

B ..c6 9 cxd5 cxd5 (D) 10 lDc3

.

10 Wbs 0--0 1 1 1i'xd5 .i.c6 12 1ih5 g6 1 3 11'h3: 1 3....tb4 - Gt����t 46; 1 3. lbg5 - Game47
..

10 ...lilxc3 1 1 bxc3 0-0 1 2 1rh5 f5 - GanH 44

12 ...g6 - Gt111e1 45

5 .ltid7
.. 9 c5 9... cxd5

142
 CHAPTER SEVEN
3 d4: 5 .i.d6 and
. . .
Fifth Move Alternatives
1 e4 e5 2 �f3 tl)f6 3 d4 ltlxe4 4 i.d3 d5
5 �xe5
In the previous chapter we looked at the
main line with 1 e4 e5 2 lbf3 lbf6 3 d4 lbxc4
4 .id3 d5 5 lbxe5 lbd7. Now it is time to
:;rudy other 5th move options for Black.
Until quite recently the symmetrical re­
sponse 5...�d6 (Games 59-62) was as popu­
lar as 5 ..lLld7. After mutual castling White
. Game 63), while 6 li)xc6 bxc6 7 0-0 .i.d6
:macks the centre, but this can be done in
more than one way. If White develops his
lJUeen's knight for this purpose, he should
prefer 7 lbd2 (Game 61) to 7 lbc3 (Game
62) because it avoids doubled pawns.
However, the main way of attacking the
centre is with 7 c4 (Games 59-60). Black has
numerous responses against this. Strengthen­
ing the d5-pawn with 7 ...c6 is a bit passive,
while the counterattack on the d4-pawn with
7...l'lk6 (Game 60) isn't sufficient to equalise
in view of 8 lbxc6 bxc6 9 c5 �e7 I 0 �12 or
1 0 lbc3 - the pawn doubling is unfavourable
ti1r Black. A sharp and critical position (de­
spite a queen exchange) is reached after
7...�xc5 8 dxc5 lbc6 (!L�e6 docs not
ClJUalise) 9 cxd5 (the only try for an advan­
tage) 9 ...'iWxd5 10 'iWc2 lbb4 1 1 �xe4 lbxc2
1 2 �xd5, and this has been considered a
tabiya of the whole 5...�d6 system for many
years. The fate of Black's knight driven into
I
the comer is the main question of this key
position, one which is studied in Game 59.
5 ...lbc6 attacks both White's centralised
knight and the d4-pawn. l lowever, this is not
tJ1c best idea as White can simply swap tl1c
knights to inflict doubled pawns on his op­
ponent. Now 6 lbxc6 bxc6 7 0-0 �e7 trans­
poses to 5 ....i.e7 6 0-0 lbc6 7 lbxc6 bxc6 (sec
transposes to 5...�d6 6 0-0 lbc6 7 lbxc6
bxc6 (sec Game 61). However, Bil!,ruer
demonstrated that White docs not have to
castle; 7 'iWe2 creates problems for Black
along the e-file, and that's why 5 ..lbc6 is
.
almost out of usc.
The continuation S...�c7 (Game 63) docs
not challenge the c5-knight and is dictated by
the wish to complete dL-vclopmcnt as soon as
possible. A bit passive, 5 ... �e7 has never
drawn much attention. However, even after
tl1e most natural 6 0-0 0-0 7 c4 c6 it is un­
clear how White can 1-,rain a visible advantage.
Probably White should consider Stcinitz's
recommendation of7 ltcl !?.
Gaf!Je 59
Van Der Wiei-Mellado
Elgoibar 1998
1 e4 e5 2 tl)f3 tl)f6 3 d4 �e4 4 i.d3 d5
143
The Petroff Defence
5 .lt»te5 .i.d6
Other options are considered in Game 63.
6 0-0 0-0
Sixth move alternatives for both sides are
studied in Game 61.
7 c4 .i.xe5 8 dxe5 lbc&
8... .i.e6 is less reliable. White can put
Black under immediate pressure with 9 cxdS
1i'xd5 1 0 1i'c2 f5 1 1 exf6 lihf6 1 2 liX.1 1i'e5
(1 2...9c6 1 3 b3 lilbd7 1 4 .i.a3 ltlc5 15 .i.bS
1IVb6 1 6 Let also looks promising for
White) 1 3 lllc4 ll:lg4 1 4 lllg3 hS (14 ...h6? 1 5
.i.xh61 lilxh6 1 6 llae1 9f6 17 :xe61 would
t."'ld the show) 1 5 .LIZ .i.dS (1 5...h4? 1 6
:net and White takes on e6 again) t 6 .i.c3
1i'g5 1 7 .i.h7+ �h8 1 8 ...g6 1i'xg6 19 .i.xg6
with a clear plus for White, Liberzon-
1-lcnnings, Debreccn 1 968.
9 cxd5
White needs to try to refute Black's set-up
altogether in order to play for an advantage.
Other moves are less critical:
a) 9 f3 lllcS 1 0 cxdS lllxd3 1 t 1i'xd3 ltlb4
(1 1 ...lllxe5 1 2 Wd4 :Cs 1 3 lilc3 .i.f5 1 4 .i.e3
fie7 also looks okay for Black) 12 -.,3
lllxdS 1 3 lldt c6 1 4 lllc3 1l'b6+ 1 5 ..xb6
lllxb6 1 6 b3 Ji.e6 17 �f2 .rl.fd8 1 S :xd8+
ltxdS is drawish, Raetsky-Kuznctsov, corre­
spondence t 9S3,
b) 9 f4 .i.f5 (maybe an <.oven safer way to
equalise is 9...ltlb4 to cxdS WxdS 1 1 .i.xe4
1i'xe4 1 2 ltlc3 9g6 1 3 .i.e3 .i.g4 1 4 llldS!?
lllxdS 1 5 ..xdS b6) 1 0 lllc3 ltlxc3 1 1 bxc3
144
i.xd3 12 Wxd3 dxc4 1 3 Wxc4 ltla5 14 We4
...e7 and Black is okay (Kapengut).
c) 9 .i.f4 lllb4 to cxd5 (or 10 ll:la3 lllc5 1 1
i.bt dxc4 1 2 lllxc4 i.c6 1 3 b3 ..xd1 1 4
:xdt :adS with equality - Yusupov)
lO.. :tl'xdS t 1 .i.c2 .i.f5 t 2 llla3 l:tad8 1 3
.i.c4 1i'c6 1 4 -.,3 aS 1 5 Ldt 1Wg6 1 6 :xd8
ll.xdS 17 lld 1 ltf8 18 We3 h6 19 ll:lb5 ltlc2
20 We2 c6 21 lild6 lllc5 with an unclear
game, Raetsky-rilatov, cortespondcncc t 982.
9 11'xd5 1 0 .c2
.•.
1 0 'lff3? is much weaker as Black has
1 0 ... .i.f5l 1 1 'lfxf5 1fxd3 12 lllc3 lllc5 1 3
9hs :res 1 4 f4 •d4+ 1 5 �ht :a<�s 16 •f3
1i'c4 when his position is preferable. To
make matters worse, White only needs to
play l 1 :C 1 ?! for Black to grab the advantage
with 1 t ...LdS 1 2 i.xe4 .i.xe4 1 3 ...g3 .i.xb1
14 :xb1 lllxeS 1 5 .i.e3 :res, when Black
should convert his pawn, Zelcic-Ascic, Rabac
2003.
1 0. . .tnb4
Absolutely the main line, but not the only
move in the position because Black can also
play 1 o... .tf5. For example, 1 1 lllc3 lllxc3 1 2
.i.xf5 llld4 13 .i.xh7+ �h8 14 'lfd3 li:ke2+
1 5 �h1 'ifxe5 16 9h3 lilf4 17 .i.xf4 (17
'ifh4 g5 18 1i'h6 9g7 1 9 ..xg7+ �g7 20
.te4 c6 is very close to equal, if not simply
cquaQ 1 7...•xf4 1 8 Ldt 'irh6 (1 S...Ilfd8?1
1 9 .tc4+ 9h6 20 'ifc3 lllc6 21 :d5 and
Black was under pressure in Mi.Tseitlin­
Kondali, correspondence 1 990) t 9 1ixh6
 3 d4: 5 . . �d6 and Fifth Mo ve Alterna tives
.
wch6 20 IZ.xd4 �h7 21 lld7 and White has a
preferable rook endgame. However, whether
this is enough to win is hard to tell.
1 1 �xe4 �xc2 1 2 i.xd5 i.f5
t2...lt�xa1 gives White a slight advantage
after 1 3 i.e4 (13 c6 lbc2 1 4 exf7+ llxf7 1 5
lldl .t f5 1 6 .i.xf7+ *xf7 17 ltk3 c6 1eads to
immediate equality) 1 3..l[c8 14 l&3 llxc5
1 5 .i.d2 llxe4 1 6 �xe4 �c2 1 7 llct i.f5 1 8
f3 li)d4 1 9 llxc7 .lxe4 20 fxe4 b6 21 *f2.
The endgame is uncomfortable for Black,
lhough not necessarily impossible to defend.
Note that after 21 ...�6 White retains the
pressure with 22 lle7! *£8 23 .i.b4, when
23 ...�5 loses to 24 llc7.
13 g41
1 3 . i.xg4
..
Or 1 3....i.g6 14 f4 and now:
a) 1 4....td3? 1 5 .Ddt i.a6 16 .le4! (16 e6
�xa 1 17 exf7+ Wh8 1 8 .le3 :ads 19 lbc3
also looks very promising for White)
1 6...�xa1 1 7 .le3 .i.e2 1 8 :Ct .i.xg4 1 9
�3 and White has good winning chances.
b) 14 ... c6 1 5 .i.c4 bS with a further split:
b1) 1 6 .le21? hS 1 7 f5 .th7 1 8 gS �xal
19 c6 fxe6 20 b>6 .i.xg6 21 fxg6 llxfl+ 22
�fl probably leads to a draw, but this line
is by no means forced and Black needs to do
defend accurately.
b2) 1 6 f5 �xal?, as in Ginda-Witt, Galati
1 973, should have lost quickly to 17 ..i.d3
l:lad8 1 8 .i.e4 lld4 19 �2 when Black will
end up with material losses. Instead Black
should accept an inferior position with
1 6...bxc4.
14 i.f4
14 .i.e4 �xa1 1 5 �3 is the alternative,
but probably less critical (1 5 .tf4 simply
transposes): 15 ... £5 16 exf6 .i.h3 17 llet
llae8 18 .L12 (1 8 .i.e3 IZ.xe4 19 �xe4 �2
20 llct �xe3 21 fxe3 c6 22 �gS .tf5 23 f7+
*h8 with equality, Tal-1imman, Reykjavik
1 987) 18..l[xe4 1 9 �xc4 �c2 20 :C1 li)d4
21 llxc7 �f3+ (21 ...gxf6?! 22 .i.e3 �2+ 23
Whl looks better; 21 ..1lf71? 22 llxt7 �f3+
23 c;i;>ht �xf7 24 fxg7 *xg7 25 .i.e3 b6 26
�12 �xd2 27 .i.xd2 aS 28 f3 a4 29 ¢>g1
.i.t.-6 30 a3 '12-1/z Tai-Karpov, Milan 1 975 -
Black will place his king on f7 and White can
never make any progress) 22 �h1 �xd2 23
llxg7+ c;i;>h8 24 lbgs .i.f51 (the most precise;
24.. .:Xf6 25 IZ.xh7+ Wg8 26 1Z.xh3 IZ.xf2 still
gives White chances - Gipslis) 25 �f7+ llxf7
26 llxf7 .i.e4+ 27 f3 .lxf3+ 28 c;i;>gt Wg8 29
lld7 lDc4 30 Wf2 .i.c6 3t llg7+ wm 32
llxh7 li)d6 33 We3 �e8 and the endgame is
drawn, Raetsky-Bclomestnykh, correspon­
dence 1 982
14 ...�xa1 1 5 .i.e4
1 5 llct has also been tried a few times,
but it is less dangerous for Black: 1 5 ...c6 1 6
.le4 (or 1 6 .i.g2 f6!? 1 7 exf6 IZ.xf6 1 8 .i.e3
llg6 19 *h 1 .tc<, 20 �d2 .lxa2 21 llxa 1
i.dS 22 .txdS+ cxdS and Black should not
have any trouble at all) 1 6 .. £5 (or 1 6 ... f6 1 7
.
�3 fxcS 1 8 .lxeS llae8 1 9 f4 gS 20 llxa1
145
The Petroff Defence
gxf4 21 .i.d4 a6 with compk-tc equality) 17
cxf6 :Xf6 1 8 .i.d (Sveshnikov-Tischbierck,
Budapest 1 988). Now Black could have con­
tinued 1 8 .. .:C:8 1 9 lL!c3 lL!b3 (1 9...bS 20 .i.g2
llg6 21 �h1 b4 22 lL!e4 .i.fS 23 l0g3 :xg3
24 hxg3 lLlc2 25 i.xc6 looks pnmusing for
White - two bishops and a weakened black
queenside) 20 axb3 a6 21 i.d3 .i.fS and
Black's position is at worst marginally infe­
rior; in fact we think it's equal.
1 5...f5?1
After this move we cannot find a route to
fuU equality. Probably Black should try
lS ...f61? 16 lL!c3 fx�:5, when his results have
been encouraging:
a) 1 7 .i.g3 :ad8 1 8 l:lxa1 .i.£3 1 9 .ixe5
J:td2 gives good compensation (Yusupov).
b) l7 .i.e3 .i.f3 1 8 l:lxa1 .i.x�:4 1 9 lL!xe4
b6 20 b4 aS 21 b5 :ad8 22 l:.ct ll£7 23 a4 h6
24 Wg2 J:td3 and Black had sufficient
counterplay in Kasparov-1imman, Paris
(rapid) 1 991 .
1 6 -td5+
Much weaker is 16 .i.xb7?! l:lab8 1 7 .i.dS+
�h8 1 8 f3 (18 lL!a3?1 l:lfd8 19 .i.c4 J:[d4 20
.i.e3 lL!c21 21 .i.xd4 lL!xd4 and Black's more
active pieces mean that he enjoys an advan­
tage in the endgame) 1 8. ..th3 19 lld1 Lb2
. position. Black seemed to lose th�: foUowing
20 e6 l:.g2+ 21 Wh 1 l:.c8 22 .i.c4 llg6 23
.i.g3 f4 24 e7 l:ld6 25 llxd6 cxd6 26 .i.xf4
.i.d7, when the responsibility of holding this
endgame lies entirely with White.
1 6...�8 1 7 llc1 c6 18 ..tg2
146
Oddly enough the bishop is best placed
here, where it seemingly has little influence.
The reason for this is found in the foUowing
line: 1 8 .i.e6 g5! 1 9 ..LgS l:lae8 20 .i.c4 (or
20 .i.h6 :xc6 21 .i.xf8 .i.h3 22 :C3 %Zg6+ 23
llg3 �g8 24 .i.e7 f4 25 ltxg6+ hxg6 26 lL!a3
f3 27 .i.c5 b6 28 .i.c.l4 .i.£5 29 b4 ru 30
lL!xc2 .i.xc2 '/z-'/z OU-Khalifman, Vilnius
1988) 20...b5 21 .i.ft f4 (Rm:entalis­
Ivanchuk, Minsk 1986). Mter 22 .i.f6+ llxf6
23 exf6 .i.fS 24 f7 llf8 Black is no worse.
1 8 .. .llfd8 1 9 �d2
This is the critical move. 1 9 f3 .i.hS 20
lLhl :U4 21 .i.e3 lZ.b4 22 lL!c4 (22 llxa1 f4
23 .i.f2 llxb2 24 lL!c4 llc2 2S lL!d6 b6 is not
clear) 22...l:la4 23 lL!a3 leads to a repetition,
Sax-Yusupov, Thessaloniki 1 984.
1 9 ...h6
1 9...llxd2?1 has an awful score. FoUowing
20 .i.xd2 :dB 21 .i.c3 :d i+ 22 llxd1 .i.xd l
White should play the foUowing plan: 23
.i.fl I (23 f4 lLlc2 24 Wf2 �PB 25 a4 aS 26
.i.xa5 lL!d4 27 .i.ft .i.b3 1/z-1/.t Kasparov­
Anand, Unares 1991) 23 ...g6 24 .i.c4 �g7 25
b4! (25 �:6+ Wf8 26 .i.f6 'iPe8 27 e7 lLlc2 is
unclear - Yusupov) 25...W 26 .i.b3 when it
looks impossible for Black to improve his
endgame without any real chance: 26 ...�f8
27 .i.d2 We7 28 f3 �e8 2<J �f2 b6 30 .i.f4
cS 31 bxcS bxcS 32 c6 c4 33 i..xc4 lL!a3 34
.i.dS .i.a4 3S .i.e3 a6 36 .i.c5 lL!c2 37 i..c4
aS 38 �g3 lLle 1 39 Wf4 i..c6 40 .i.e2 h5 41
 3 d4: 5 . . . �d6 and Fifth Move Alterna tives
h3 �dB 42 �gS .i.e8 43 c7+ �c7 44 f4 li)g2
45 .i.f1 �c6 46 .i.f2 �d6 47 .i.xg2 1-0 Mru­
gala-Wisc, correspondence 1 999.
20 h4 :.d3 21 .i.f1 !
This was Timman's improvement on his
play in an earlier matc�>ame with Yusupov:
21 Axa1 gS! 22 hxgS hxgS 23 .i.xgS (23 .i.fl
hrxf4 24 .i.xt.l3 l%d8 25 c6 l:txd3 26 e7 .i.hS
27 li.)b3 b6 28 ltet .i.cB and Black is fine
Yusupov) 23 .Jlg8 24 .i.f6+ Wh7 25 li.)ft f4
.
26 �h2 Jlg61 27 :C t l:th6+ 28 �gl l:tg6
with equality, Timman-Yusupov, 2nd match­
game, Linares 1 992
21 ...:.d4 22 .i.e3 :.d5 23 :.Xa1 :.Xe5 24
.!Dc4 :.e6
In the stem game Black played 24...l:td5
25 .i.g2 l:tb5?1 26 l:tc1 l:td8 27 .i.xa7 l:td1 28
l:lxd 1 �xdl 29 �d4 f4 30 .i.c4 and White
enjoyed a clear superiority, Timman­
Yusupov, 6th matchgame, Linares 1992.
L ater Black improved with 2S..ltdd8 26 l:te1
'it>h7 27 �h2 �hS 28 .i.cS b6 29 .i.e? l:td4
30 lbcl6 l:tc:12 31 �xc6 l:tb8 32 Wg3 f4+ 33
Wxf4 llxf2+ 34 �g3 ltxb2 with n:asonable
drawing chances even though it is stiU a
tough defence, Bucher-Girtz, Biel 1998.
25 f41 .i.f3
25 ...:C4 26 .i.d3 l:tac8 looks tticky, but af­
lCr 27 �f2! l:t4e6 28 �5 White has the ad­
vantage.
26 c!Lie5 .i.d5
26...�h7 27 Wf2 .i.hS 28 l:tc1 also fa­
vours White.
27 h5 :.aea 28 .i.d3 �f3
Also possible is 28...b6 29 i.f2! (29
.i.xf5?? l:txeS! wins for Black). Now White
has the edge after 29....i.c4 30 �c4 .i.dS 31
.i.xdS cxdS 32 l:tdt .
29 �2 .i.xh5 30 :.h1 .i.g4 3 1 .i.xa7
:.Xe57
This is simply a blunder. Black needs to
play 31 ...lla8 32 i.d4 l:ta4 33 �c3 :xf4+ 34
�g3 l:ta4 35 �c2 liaS 36 .i.b3 when after
36...llxe5 37 i.xeS White has some chances
to win the game, though nothing is clear.
32 fxe5 :.Xe5 33 lbh6+ ! �g8 34 .i.c4+
�8 35 :.h8+ r/;e7 36 lZ.b8 b5 37 :r.b7+
wdB
37...�f6 38 .i.ft l:td5 39 :C7 also wins for
White.
38 .i.b6+ �c8 39 :.c7+ Wb8 40 :.Xg7
1 -0
Game 60
- Shirov-Vusupov
TerApe/ 1997
1 e4 e5 2 �f3 c!Lif6 3 d4 o!bxe4 4 .i.d3 d5
5 �xe5 .i.d6 6 0-0 0-0 7 c4 �c6
This particular variation, like so many in
the Petroff, is a speciality of Yusupov's.
7...c6 is also possible, after which White
has a wide range of choices:
a) 8 1i'c2 1Wh4 9 li.)f3 1i'hs 1 0 cxdS cxdS
1 1 ..ixe4 dxc4 1 2 'ifxc4 li.)c6 and Black has
compensation (Yusupov).
147
The Petroff Defence
b) 8 cxdS cxdS 9 'irc2 .IL:8 10 f3 (10 R.xe4
dxc4 1 1 .xe4 .f6 1 2 .i.£4 �c6 looks fine
for Black) to ...�f6 1 1 .i.gS h6 1 2 .i.h4 �a6
1 3 a3 -*.xeS 14 dxeS lr.xeS t S R.f2 and White
has compensation for his pawn, although it is
difficult to see how Black should ever end up
being worse.
c) 8 llXJI seems to give White the edge af­
ter 8 ...�xc3 9 bxc3 and now:
c1) 9 ... R.e6 10 £4 .i.xeS 1 1 fxeS dxc4 1 2
-*.xh7+!? Wxh7 1 3 1Wh5+ �g8 t4 R.gs •as
1 5 .J:r.£3 �7 1 6 lr.g3 looks very dangerous
for Black.
c2) 9...lbd7 1 0 £4!? (tO �xd7 R.xd7 trans­
poses to S...llkl7 6 �xd7 R.xd7 7 0-0 .i.d6 8
c4 c6 9 W �xc3 10 bxc3 0-0) 10 ...�f6 1 1
cS J..e7 t 2 5 J..d7 1 3 g4 gave White an at­
tack in Korneev-Y.Hernandez, Mondariz
1 997.
c3) 9....i.xe5 10 dxeS dxc4 l t R.xc4 fle7
(1 t ...'Wxd1 1 2 lr.xd1 J.. 5 1 3 .L3 lr.eB 1 4 f4
is clearly better for White; it gets even better
after 14 ...�7? 1 5 e6! when White was win­
ning in Maroczy-Marshall, Paris 1900) 1 2 a4
lr.d8 1 3 1Wh5 l:te8?! (13...g6 was sadly neces­
sary although after 14 J..gS gxhS 1 5 .lxe7
:CB 1 6 J..d6 Black is in a bad state) 14 J..a3
'l'd7 1 5 l:adt '1'5 16 W'h4 .xeS 17 f4 'Wf6
(17 ...1i'e3+ 1 8 �h1 .i.e6 1 9 .i.d3 h6 20 aS!
and Black's queen is trapped) 18 .x£6 gxf6
t9 .J:r.£3 J..e6 20 lr.g3+ �h8 21 J..e7! hS 22
J..x f6+ t -0 Chigorin-Lebedev, Moscow 1900.
S lOxc&
This is the right path to an opening advan­
tage. Mter 8 cxdS �xd4 9 .i.xc4 (9 �c4
1Wh4 1 0 �xd6 t'Oxd6 t 1 �c3 .i.5 and Black
has equalised - Euwe) 9 ....ixe5 1 0 �3 .IL:8
1 1 lr.e1 eo5 12 •n g6 1 3 -*.£4 ..i.x£4 1 4
'ffx f4 �6 1 5 :C3 J..d7 1 6 lr.act 5 1 7 .LI3
..f6 t8 a4 lr.xe3 1 9 lr.xe3 lr.e8 Black is very
close to equality, Zelcic-Pavasovic, Nova
Gorica t 997.
s. . .bxc& 9 c5 j.e7 1 0 &3
This seems to be the soundest way of
playing for White, but it is not the only way:
a) After tO f3 �gS t t �c3 :Cs t 2 •a4
148
R.d7 13 .i.d2 lr.b8 14 l:tab1 .i.£6 15 �h1 h5
t6 �2 h4 17 h3 'l'c8 18 J..a6 'Wd8 t9 .i.d3
..c8 20 .i.a6 the game finished with a draw
in Kaspamv-Yusupov, Horgen 1995. 1 8
�f41? � 19 lr.fet is interesting, but we
believe that Black is okay after the foUowing
line: 19...-*.gS!? 20 .i.a6 ..dB 21 �xe6 lr.xe6
22 :Xe6 R.xe6 23 .i.xgS ..xgS 24 ..xc6
.i.xh3! 25 R.n ! .i.xg2+! 26 .i.xg2 ..5 27
l:lg1 h3 - the game will end in a draw.
b) 10 l0d2!? looks stronger: 1 0.. ..i.f6 1 1
�xc4 dxe4 1 2 .i.xe4 .i.a6 (t2....i.xd4 1 3
.d3 �h8 14 .i.g51? .i.xf2+ 1 5 lr.xf2 .xgS
16 Lh7 looks slightly better for White) 1 3
:C 1 .i.xd4 1 4 1i'c2 lr.eR 1 5 .i.e3 .i.xb2 1 6
.i.xh7+ �h8 (Sanche-.r.-Morgado, correspon­
dence 1 978) and now after 1 7 •xb2 �xh7
1 8 l:ad1 .e7 19 1i'c3 we think White has
some pressure.
1 0. . .f5
Black seems to be unable to equalise here,
and it's not due to a lack of trying:
a) 10....i.5 1 1 f3 �xc3 1 2 bxc3 .i.xd3 13
'l'xd3 aS (13....i.g5?! 14 ..a6! .i.xct 1 5
Let .d7 16 lr.fel lr.fe8 17 lr.xc8+ l:r.xe8 18
h3 h5 19 .xa7 and Black did not have
enough for the pawn Shirov-Hiibner, Frnnk·
furt [rapid[ 1996) 1 4 l:tb1 lr.e8 1 5 .i.d2 and
White is a bit better.
b) to....i.f6 1 1 .c2 (1 1 �xe41? dxe4 1 2
.i.xe4 transposes to 10 �d2 .i. f6 l 1 �xe4
dxc4 1 2 .i.xe4) t l ....i.xd4 1 2 �c4 dxe4 13
R.xe4 W'h4 (1 3...R.a6?! 14 .J:r.dt with a plus;
 3 d4: 5 . �d6 and Fifth Mo ve A lternatives
. .

14....i.xf2+? 15 Wxf2 'i'h4+ 16 g31 - 16

�g) ?l llae8! is Black's point - 16. ..xh2+ 17
..

.tg2 and White is winning - Kasparov) 14 g3

..f6 1 5 .i.e3! .ixe3 (1 5 ....i.xb2 1 6 llab1 .i.eS
1 7 .i.xh7+ Wh8 18 .i.e4 should also favour
White) 16 fxe3 1i'h6 1 7 llf4 g5 1 M llf2 .ih3
1 9 .if5 and in this position White is some­
what better, lvanchuk-Yusupov, Hoq,�n
I IJ95.
1 1 f3 tDg5 1 2 1Wa4 �d7 13 .i.f4

1 7 o!Llxf4 1 8 g3 o!Llh3+ 19 �g2 1Wh6 20

..•

�d5 f4?
Black should play 20 .:ae8 21 �7+ �h8
..

22 ..i.xf5! llxe7 23 .i.xd7! (but not 23 llxc7?!

9d2+ 24 Wxh3 llxf5 25 llxf5 .ixf5+ 26 g4
Lg4+ 27 Wg3 .ih51 28 lle5 ..d3+ and
Black attains a draw) 23. .ltxft 24 Wxft llf7+
.

25 llfS llx5+ 26 .ixf5 and White is only

slightly better (Shirov).
21 �e7+ �h8 22 ltlf5!
Or 1 3 llk2 .if6 1 4 ti)f4 'lfe7 1 5 .i.d2 g6 It is important that it's the knight that
1 6 lLe1 ..g7 17 .i.c3 hS 18 ..i.a6 llab8 1 9 goes to 5: after 22 .i.f5? fxg3 23 hxg3 g6 24
ti)cB h4 20 ti)e5 with a slight advantagt:, .ixd7 ..d2+ 25 �xh3 l:lxft White is missing
Timman-Yusupov, Bth matchgume, Linares his bishop in defence.
1 992. 22...1i'g5
13 �f6 14 llae1 lDe6
••. 22. .ixf5 23 llxf5 1lae8 24 ..d t l wins for
..

14 ..ll:lf7 is punished by 1 5 ti)e2 g6 1 6

. White according to Shirov - the knight is
'liaS, when Black cannot defend himself simply ttapped.
properly. 23 o!Llh4 1Wh6 24 �f51
1 5 �e5 .be5 1 6 :Xes -...4
16. . ti)f4 1 7 .ib1 tl)g6 1 8 lleet Ae8 was
.

an alternative, although White's position still

looks more promising.
17 f41?
Shirov, the great calculator of complicated
lines, naturaUy takes the chance to enter this
wildly complex position. He suggests that
after 17 .i.xf5 llxf5! (17 ...ti)xc5? 18 ..i.xh7+
Wxh7 1 9 dxc5 is good for White) 1 8 llxf5
lt:lxcS t 9 dxcS ..xa4 20 �a4 .ixf5 White
has a slight edge.
After 17 f4 Black has little choice but to
take the challenge. 24. . .fxg3

149
 The Petroff Defence

The last chance was probably 24....i.xf5 25 0-0-0 cS with a very complex position in
lbxf5 wrg6 26 'itc2 'iig4 27 'ii'd1, when which White might have an edge) 9 wrxd3 c6
White retains a large advantage but Black can 10 0-0 .i.g4 11 lbg5 (the less obvious 11
struggle still. .i.d2 llXI7 12lbd4 � 13 h3 .i.e6 14lbxe6
25 hxg3 g6 26 i.xd7 1rd2+ 27 �h3 fxe6 is not easy to evaluate) 11...h6 12 'itd4
l:lxf1 28 i.xc6 l:laf8 29 i.g2 .i.e6 13lLlxe6 fxe6 14 c4 1fe7 15 .i.d2 c5 16
Also winning is 29 .lie7 l:te1 30 Axe 1 Wg4 d4 and Black has no obvious worries in
'itxe1 31 'itb3 Wg7 32 g4, but the game con­ this unclear position, Raetsky-Y.Aleksandrov,
tinuation looks safer despite the complex correspondence 1983.
lines. d) 6 c4
29.. Jl1 f2
Black also loses after 29...l:t1 fl 30 l:le8 gS
31 .i.d5 llf6 32l:.xf8+ llx£8 33lbf3, but not
32 lbm g4+ 33 Wxg4 l:lg6+ 34 Wh3 :h6+
35 lbh4 l:lxh4+! 36 gxh4 'itd3+ whc::n Black
escapes with a draw.
30 1rd7 g5 31 l:le7 1rc2 32 ..i..e4 l:lh2+
33 �Lh4+
Black's attack also runs out of steam after
33...'fle2+ 34lbf3 34...h5+ 35 Wxg5 We3+ 36
Wg6 ltg8+ 37 ll.g7 1fxe4+ 38 Wh6 'ife3+ 39
lDgs, when despite his extra exchange Black
is 'out of bullets'.
34 gxh4 'ife2+ 35 �g5 l:lg8+ 36 This allows Black to act quickly with
�6 1-0 6 ....i.xe5 7 dxeS lbc6 8 0-0 lbcS 9 cxdS
r------. 'itxdS 10 .i.bS 'ifxdl (weaker is 10....i.e6 11
Game6 1 li'xdS .i.xdS 12 lbc3 0-0-0 13 .i.e3 lbe4 14
Timoscenko-Yusupov lbxcl5l:.xd5 15 .i.xc6 bxc6, when White had
F'mn� 1979 an endgame edge in Lindoerfcr-Meijers,
Schwabisch Gmund 1998) 11 llxd1 .i.d7 12
1 e4 e5 2 ll:!f3 lDf6 3 d4 lbxe4 4 i.d3 d5 lbc3 lbc6 13 f4lbcd4 14 .i.a4 0-0-0 and here
5 llJxe5 i.d6 6 0-0 we can see no argument supporting a white
White has a extensive list of alternatives: edge.
a) 6 lbc3 lbxc3 7 bxc3 0-0 8 0-0 ll:X17 6 ...0-0
transposes to Game 62. Or:
b) 6 'irf3 0-0 7 0-0 cS!? 8 .i.xe4 dxe4 9 a) 6 ....i.xe5 7 dxcS lbc5 is a slightly dubi-
Wxe4 'fle7 10 lle1 lle8 11 lbd2 cxd4 12 ous favourite of the solid GM from Uthua-
lbet1 (12lbec4i.b4 13 'iixe71lxe7 14 :xe7 nia, F�uardas Rozentalis. Now 8 .i.e2! is
.i.xe7 1Slbc4lbc6 would give Black a pref- probably the right path for White here, for
crable endgame) 12 ...lbc6 13 1fxe7 llxe7 14 example 8.. 0-0 9 b31? lbc6 (9...c6 10 .i.a3
.

lbe4 .i.c7 15i.d2 .i.b>416lbc5 .i.d6 17lbe4 'iie7 11 f4 aS 12 lbc3 lblxl7 13 wrd4 gives
.i.c7 with a draw, Raetsky-Mironov, corre- White a powerful initiative) 10 L3 b6 11 f4
spondencc 1985. lte8 t2lbc3 d4 13i.f3 .i.b7 14 .i.xcS bxcS
c) 6lbd2 .i.xeS 7 dxe5lbcs 8lbf3lbxd3+ 15 lba4 wrc7 16 We2 ll:XI8 17 'ifb5 and
(or 8....i.g4!? 9 h3 lbxd3+ 10 Wxd3 .i.xf3 11 White has strong prt:ssure, C.l-lanscn-
'irxf3 0-0 12 'irg3 Wh8 13 i.gS 'ii'd7 14 Ro:r.entalis, Malmo 1997.

150
 3 d4: 5 .. . �d6 and Fifth Move Alternatives
b) 6 ...Qk6 7 lfu.c6 bxc6 8 c4 Wlt4?!
(fL0-0, transposing to 6...0-0, is preferable) 9
g.� 1i'h3 (9 ...'ii'f6 is refutoo by to i..xe4 dxe4
l tltlc3 i..f5 12 f31 ..g6 13 fxe4 i..h3 14 e51
i..xft 15 exd6 i..h3 16 'ii'e2+ 'l'e6 17 'ii'xe6+
i..xe6 18 dxc7 i..xc4 19 i..f4 with a clear
advantage for White - analysed by the very
reliable duo Yusupov and Dvoretsky) to c5
i..g4 11 i..e2 i..xe2 12 1i'xe2 i..e7 13 f3ltlf6
14 i..f4 and Black has achieved little witlt his
kingside actions - White has a slight advan­
laJ.,>e.
c) 6...c51? allows White to gain an edge
with 7 i..bS+ltld7 8 dxc5 i..xeS 9 9xd5 0-0!
(9...1i'e7?1 10 c6! would be an unpleasant
surprise) to ..xe4 'ii'c7 11 i..d3 g6 12 'ii'h4
ltlxc5 13 i..e 2 i..f5 when Black has some
compensation for the pawn, but hardly
enough. White could even consider 14 Qk3!?
here.
7 lDd2
7 ltc 1 is a respectable alternative, after
which we have the foUowing possibilities:
a) 7...i..xe5 8 dxe5 Qk6 9 i..£4 f5 to f3
ltlc5 11 i..b5 ltle6 with unclear play (Yusu­
JX>v).
b) 7 ... c5 8 c4!? cxd4 9 cxd5 i..xe5 (9 ltlc5
.
..
10 i..c4 :C8 11 i..f4 i..f5 12 llXl2 looks
hrood for White, as after 12 ... £6 13 ltlc6!
Ulack is under attack) 10 i..xe4 f5 11 i..O
1i'd6 12 h3 giving a position balanced on the
edge between a slight white plus and �-quality.
In other words, something along the lines of
equality but Black needs to prove it! In Slo­
bodjan-Forintos, Germany 1992, Black failed
to do so and foUowing 12...ltld7 13ltla3 i..£6
14llk2ltle5 15ltlxd4ltlxf3+ 16ltlxf3 b6 17
i..g5 i..b7 18 i..xf6 .llx£6 19 :C5 White held
a clear advantage.
c) Lltld7 8 Le4 (8 ltlxd7 leads to an
immediate draw after 8...i..xh2+ 9 Wxh2
1i'h4+ 10 �g1 1i'xf2+) 8...dxe4 9 ltlc4 ltlb6
10 ltlxd6 1i'xd6 11 llxe4 i..f5 12 l:leS Wg6
and Black has good counterplay for the
pawn.
7 �xeS 8 dxe5 �c5 9 l0b3
••.
9 'ii'h5 does not look dangerous here,
9...ltlxd3 10 cxd3 c5 being the must natural
reaction. Now 1 1 b4?! looks unjustified:
11...cxb4 12 a3 bxa3 13 i..xa3 l:le8 14 i..d6
ltlc6 15 l:la4 'ii'd7 16 l:lf4 1Ve6 17 d4 'ii'g6 18
1i'f3 i..c6 with a clear edge for Black, J. Pol­
gar-Kamsky, Groningen 1993. After the
more sensible 11 ltlb3 'ii'c7 12 i..f 4 i..e6
Black shouldn't be worse.
9 l0xd3 1 0 '1Fxd3 tOe&
•.•
1 1 �f4
11 f4 should probably be met by 11...f6 1 2
ex£6 1Vxf6!? (12.. .l:lxf6 13 f51? ltle5 14 'ii'h3
.ffi 15 g4ltlc4 16ltld4 is very complex and
difficult to evaluate) 13 .xdS+ i..c6 14 'ii'c5
llad8 15 i..d21? 1i'xb2 16 i..c3 •xc2 17
i..xg7! 1i'xc5+ 18ltlxc5 l:lfe8 19 i..c3 (Ljubo­
jevic-Razuvaev, Amsterdam 1975), and now
19...i..c4! 20 ltlxb7 i..xft 21 ltlxd8 i..xg2 22
�xg2 .:.Xd8 leads to an endgame that Black
757
The Petroff Defence

will have to defend, but rtln defcm.l with ac­ 23 :ata 24 .i..e3 1tg6 25 �2 lL!f5 26
...

curate play. i.f47

11 1id7 1 2l:tad1
••.

12 .i.g3 leads to a hannless position after

12...b6 13 :ret �b4 14 'ifd2 c5 15 �4
'it"g4 16 c3 cxd4 17 cxb4 d3 18 'it"xd3 Wxb4
19 a3 1t"c4 20 ..d2 (1/2-1/2 Griinfeld­
Yusupov, Amsterdam 1982).
1 2...1fg4 1 3 1i'e3 liJe7 14 h3
14 'ifc5 �fS 15 g."\ c6 docs not cause any
real problems for Black.
14 1tg6 1 5 c3
..•

15 1t"c5 invites Black to play a pleasing

piece sacrifice with 15....ixh3!? 16 .ig3 �f5!
17 gxh3 �xg3 18 fxg3 'ii'xg3+ 19 'it'h1
1t"xh3+ 20 Wgl 'it"g3+ 21 �h1 l:lae8!? (of
course a draw was available) 22 :d3 Wh4+
23 �g2 :e6, when we fear for the safety of
the white king.
15 b6
... 1 6 e6?!
Tougher resistance was possible with 26
l:lft, though after 26 ....ia4 27 llc I Wc6 28
�2 .ibS 29 c4 .ia6 Black is very close to a
winning position.
26 lL!h4 0-1
...

Game62
Anand-Mishra
India 1988
1 e4 e5 2 lLJf3 lL!f6 3 d4 �xe4 4 i.d3 d5
5 lL!xe5 i.d6 6 0-0 0-0 7 lL!c3 lL!xc3
The colourful 7...f5!? 8 f3 .ixeS 9 dxcS
�xc3 10 bxc3 11c7 1 I .D.ct .ic6 12 a4 cS
lead to unclear play in Geller-Yusupov, Vil­
nius 1980.
8 bxc3 &iJJJ7

I\ poor strategic mistake. Black will enjoy

the open f-filc and develop very l]uickly. 16
�d4 cS 17 �f3 .i.b7 18 �h4 We6 would
give Black plenty of counterplay, but this was
stiU the best option.
1 6 c5 1 7 exf7 + l:txf7 1 8 l:tfe1 i.d7 1 9
•••

1fg3 1fc6 20 i.g5 lLJf5 2 1 1ff3 h6 22

i.c1 lLle7 23 1fe271
Here the queen is not very weU placed. 23
Wg3 is stronger; foUowing 23 ...1%af8 White
.

has 24 .if4 �f5 25 'ilfh2 gS 26 .ie5 with

some counterplay, even though Black is bet­
ter after 26...'ilfa4. ll..c5!? has also been suggested: 9 ..h5 g6

152
 3 d4: 5 ... �d6 and Fifth Move Alternatives
10 'lt'h6llk61? 11 .i.gS .i.e? (1 l...f6 12 .i.xg6
- 12 o!ilig6? llt71 - 12.....c7 13 .i.xf6 llxf6
14 .ixh7+ ..xh7 15 ..xf6 might favour
White slightly) 12 .i.xe7 ..xe7 l 3 llae1 1ff6
14 �xc6 bxc6 1S dxcS ..xc3 16 f4 ..xc5+
17 �h1 f5 with unclear play (Yusupov).
However both 14 f4!? and even 14 .i.bS!?
look like reasonable improvements for
White.
91le1
Or:
a) 9 li:)xd7 .ixd7 transposes to 5...�d7 6
�d7 J.xd7 7 0-0 J.d6 8 etk3 �xc3 9 bxc3
IJ-0.
b) 9 f4 cS 10 c4!? cxd4 11 �xd7 (1 1
cxdS?I �f6 12 .tc4 .i.cS 13 1rf.3?!- 13lbd3
'ikc7 only leaves Black slightly better -
13....tf5 14 lle1 WaS! 1S .i.b2 d3+ 16 �h1
d2 and White is in deep trouble, Kupreichik­
Dvoretsky, Leningrad 1974) 1 L..i.xd7 12
cxdS Was 13 1Whs f5 14 •a .i.bS and Black
should not be worse.
9 . .li�xe5
9....txeS I 0 dxeS llks is a respectable al­
ternative,
after which White can play:
a) 11 llbl �xd3 12 cxd3 b6 looks very
sc 11id for Black. I r is unlikely that White can
slrueeze an advantage out of the position, for
example 13 :b4 f5 14 cxf6 (14 d4 .i.e6 15
.lla3 f4 16 llb2 llf7 with chances for both
sides) 14.....xf6 1S l:lf4 .i.fS 16 .i.d2 Wd6 17
1i'f3 .id7 18 1Whs c6 and Black had equal-
ised, A. Sokolov-Schandorff, Bundesliga
2001.
b) 11 .tfl :C8 t21Whs 1i'd7 13 J.e3 1i'c6
14 llab1 a6 1S llb4 b6 (1S...�e4 16 c4 .i.e6
17 cxdS .i.xdS 18 c4 .i.e6 19 .i.d3 would
give White real threats) 16 llh4 h6?! (Chibur­
danidze-Schussler, Haninge 1988), and now
White had the chance to play 17 c4! dxc4 18
.i.xc4 .i.e6 19 .i.xh6! .i.xc4 20 .i.e3 with a
close-to-winning position - 20... £6 21 cxf6
.xf6 22 1lxc4, if forced, simply leaves White
a piece up. Instead of 16...h6, Black should
play 16...11t'g6 17 1i'xg6 hxg6 when maybe
White has a small plus.
1 0 dxe5 .ic5 1 1 1ih6 g6 1 2 1ih6 lieS
1 3 �g5 .if8 14 ....4 1rd7
14 ... .te7 1S .tf6 eSt? 16 c4 d4 does not
look clear at all. All these positions are really
difficult to evaluate, as time after time every­
thing hangs on one move.
1 5 .lab1 1lb8 1 6 c4 d4 1 7 f4
1 7 •c6?
...
First of at� it is not obvious what the
queen is doing here. Secondly, Black should
not give White a free hand to continue his
attack on the kingsidc. After 17...'iVg4 18
•f2 cS Black would at worst only be slightly
worse.
1 8 �f6 �g7?
Black should not voluntarily part with his
only defensive piece. After 18...h6 19 1Vg3
Wh7 20 :n .tfS 21 .i.xfS gxfS 22 'llh3 •d7
23 llf3 things look bk"llk for Black, but this is
1 53
The Petroff Defence

still favourable to the game. 6 lbd2 might be a good way to tight for
1 9 f5 11t'c5 an advant:tf.,re. Now Black has two main ways
Or 19....txf6 20 exf6 .i.d7 21 l:tftl and togo:
the double threat of fxg6 followed by .i.xg6 a) 6...lbd6 7 'iff3 (l 111h5 g6 8 .i.xg6?1
combined with 'ti'h6 ends aU speculation does not work on account of 8 ... fxg6 9lbxg6·
about the result. .i.g4! to 'ifxg4 l:tg8 II 'ilh5 Axg6 1 2 ltl£3
20 -*.xg7 'it>xg7 21 •f6+ �g8 22 e6 1 -0 lbd7, when White doesn't have enough for
the piece) 7 ...c6 8 ltlftl? (Kapcngut) 8... 0-0 9
Game63 lbg3 lbd7 10 ltlxd7 'ifxd7 1 1 0-0 1tg4 1 2
Aagaard-Legky 'ifxbr4 .i.xg4 1 3 .i.f4 and White might have a
Budapest 1996 very tiny edge.
b) 6... ltlxd2 7 .i.xd2 ltlc6 (1...0-0 8 'ifh5
1 e4 e5 2 �f3 �f6 3 d4 �e4 4 .i.d3 d5 g6 9 1i'h6 ltlc6 1 0 lfuc6 bxc6 1 1 0-0-0 l:te8
5 �xeS -*.e7 1 2l%dc1 :bs 13 :CS looks better for White)
8 lbxc6 bxc6 9 0-0 0-0 10 'iVhs g6 1 1 'ti'h6
llb8 1 2 b3 .i.f6 1 3 c3:c8 t 4:fe1 :xe l+ 1 5
:xe1 .i.e6 1 6 ..f 4 .i.e7 1 7 'ii'gJ .i.d6 1 8 .i.f4
.i.xf4 19 1txf4 'ifd6 20 1if6 and White had a
bit of pressure in Hort-Spassky, Reykjavik
1 977.
6 ...0-0
The saf est choice. Others include:
a) 6...lLld7 7 .i.f4 (l c4 is less dangerous:
7...lbxc5 8 dxcS c6 9 cxd5 1txd5 10 Wn
.i.fS! 1 1 ltlc3 t£lxc3 12 'ifxf5 br6 1 3 'iVh3 :d8
with unclear play- Yusupov; 7lbxd7 .i.xd7
ttansposes to s. .ltld7 6 t£lxd7 .i.xd7 7 0-0
.

This is a bit passive and Black can easily .i.e7) 7 ...t£lxe5 8 .i.xeS 0-0 9 c4 c6 10 1tc2
end up in a slightly worse position. and White has an edge according to Euwc.
Another sideline that is no longer popular b) 6...t£!c6 also doesn't fully equalise: 7
is 5 ...ltlc6 6 ltlxc6 bxc6 7 'ife2!? (l 0-0 .i.d6 ltlxc6 bxc6 8 c4 0-0 9 t£lc3 ltlxc3 10 bxc3
ttansposes to 5....i.d6 6 0-0 ltlc6 7 ltlxc6 dxc4 1 1 .i.xc4
bxc6, and 7 0-0 JJ..e7 transposes to 5....i.e7 6
0-0 ltlc6 7 lbxc6 bxc6) 7...'ife7 8 0-0 g6
(8...lLld6 may be safer though after 9 :c 1
1txe2 1 0 ltxe2+ .i.e6 1 1 ltld2 �d7 1 2 ltlf3
f6 1 3 c3 .ifS 14 ltlc1 g5 1 5 .i.xfS+ lDxfS 16
lLld3 White retains a very slight edge) 9 .i.xe4
..xc4 1 0 1td2!? (this looks more logical than
lO 1txt:4+ dxc4 1 1 :e1 f5 1 2 f3 .i.g7 1 3 c3
0-0 1 4 .i.f4 c51? 1 5 dxc5 :b8, which gave
Black considerable compensation in PiUs­
bury-Schlechter, Munich 1900) IO....tc7 1 1
:e1 1tf5 12 'iff4 and White has St.-vera1
threats.
6 0-0 and now:

154
 3 d4: 5
bt) 11....LI6 12 1fd3 lZ.b8 13 h3 gives
White a s(jght edge. Holzke-Yusupov,
Hundesliga 2000 continued 13...c5?! 14 dxc5
.lxc5, an(l here White can play 15 .ixt7+!
<bxt7 16 'tl'c4+ .i.e6 17 'tl'xc5 'ird6 18 'ifxd6
cxd6 19 .ia3 with good chances of convert­
ing the extra pawn into a full point.
b2) 1t ....tf5 12 .if4 .td6 13 .tg3 .JZ.e8 14
'ira4 .ie4 15 lLe1 .ixg3 (15 ... h5?! is reck­
less: 16 f3 .i.d5 17 llxe8+ Wxe8 18 lle1 1t'd8
19 .i.xd5 cxd5 20 1Vc6 .i.xg3 21 hxg3 lZ.b8
22 :C5 with a clear plus for White, Ivan­
chuk-Yusupov, Novgorod 1995) 16 fxg3
i.d5 17 .i.d3 'irg5 18 'ifc2 h6 19 Wf2 and
here we believe there is no argument about
White's edge.
7 c4
Maybe the best idea for White is 7 llel!?,
as first suggested by Steinitz. After 7 ... f6 8
ll'ID f5 9 c4 .i.c6 10 cxd5 .ixd5 11 ll'lc3
ll'lxc3 12 bxc3 ll'lc6 13 ll'lc5 ll'lxe5 14 llxe5
White was a bit better in Ed.Lasker-Kupchik,
New York, 1915.
7 ...c6
The passive 7...ll'lf6 does not give real
chances for equality after 8ll'lc3 and now:
a) 8 ...ll'lc6 9 cxd5!? ll'lb4 10 .ic4 ll'lbxd5
It Wb3 c6 (11....ie6?! is a silly invitation: 12
WVxb7 .JZ.b8 13 9xa7 .JZ.as 14 �7 .JZ.b8 15
1fc6 ll'lb4 16 1ff3 J.xc4 17 ll'lxc4 1fxd4 18
ll'lc3 ..e5 19 ll'lfS and White is much better,
G.Guscinov-Ekdyshman, St Petersburg
2000) 12 lZ.e1 .i.d6 13 .i.g5 widu plus.
. . . .J.d6 and Fifth Move Alternatives
b) 8...c6 9 lZ.e1 ll'lbd7 10 .i.g5 dxc4 11
.i.xc4 ll'lb6 (11...ll'ld5?! is punished by 12
.ixd5 .ixg5 13 .ixt7+1 :Xt7 14ll'lxt7 �xt7
15 1Vh5+ *fB 16 li'xh7 ll'lf6 17 'lt'h8+ll'lg8
18 d5 .id7 19 llad1 with a strong attack,
Yurtaev-D.Frolov, Tomsk 1998) 12 .ib3
ll'lbd5 13 li'f3 .ic6 14 llad1 WaS and Black
is only slighdy worse (Yusupov).
8&3
8 cxdS cxd5 9 .i.xc4 dxe4 10ll'lc3 .ifS 11
l:le1 ll'ld7 12 ll'lxd7 1Vxd7 13 ll'lxe4 lZ.ad8
gives Black excellent compensation for the
pawn, a draw being the likely result.
8 1fc2 ll'lf6 9 c5 is less forcing and there­
fore also more ambitious: 9 ...ll'lbd7 toll'lc3!?
ll'lxe5 11 dxe5ll'ld7 12 .ixh7+ *h8 13 .if5
ll'lxe5 14 lle1 .i.f6 (14...f6 is wt:aker; after 15
.i.e3 .i.xfS 16 1fxf5 d4 17 :Cd1 i.xc5 18
lL\.14 ..e7 19 1Vh3+ *g8 20 ll'lxc5 dxe3 21
1fxe3 White enjoys slight pressure) 15 .if4
lZ.eS 16 lZ.e3 .i.xfS 17 1fxf5 lL\g6 18 llh3+
*g8 19 1Vh5ll'lh4 20 .i.g3 1fe7 with a com­
plex batde, Mi.Tseidin-Karasev, Leningrad
1970.
B �c3 9 bxc3 dxc4
...
Also possible is 9...ll'ld7 10 cxd5 (10 f4
dxc4 1 I .i.xc4 transposes to 9...dxc4)
to...ll'lxe5!? 11 dxe5 ..xd5 12 1fe2 .tc6 13
l:d1 1Wa5 with decent counterplay.
10 .i.xc4 liJ47
1 1 ltlxf71?
A surprising sacrifice. 11 f4 is generally
recommended: 12.)t)f6 12 l:lb1 (12 1fb3
155
The Petroff Defence
llkls 1 3 f5 f6 1 4 .!i)g4 bS 1S .i.e2 aS with
unclear play is not something Black should
avoid) 12...1Vc7 13 9b3 .i.d6 14 .i.a3 .i.xa3
lS irxa3 .i.fS 16 libel .i.e4 17 lL\g4 .!i)xg4
18 llxe4 l:lae8 19 llxe8llxe8 20 irxa7 'ii'e7
with chances for both sides, Popiel-Von
Gottschall, Hanover 1902.
1 1 .. Jtxf7 1 2 .txf7+
Also possible is 12 '1Vb3!? 'ii'ffi 13 l:let
.i.d6 14 .i.g5ltJb6 lS .i.e7 .i.xe7 16 .i.xf7+
'i'xf7 17 llxe7 1i'xb3 18 axb3 �ffi 19 lieS
with a double-edged endgame.
1 2 �xf7 1 3 Wb3+ �8 14 lle1 b6!
.•.
The only move, but good enough to f�ght
on equal terms.
1 6 c4
lb.is might be too rash a decision. After
lS .i.a3!? .i.xa3 16 'ii'xa3+ cS 17 c4 the posi­
tion is less clear, though it is difficult to be­
lieve that Black should be worse.
1 5 ...lL!f6 1 6 .tb2 .i.f5 1 7 d5 ..d7 1 8
Jlad1 Ilea 1 9 dxc6 •xc6 20 .ta3 .te6
156
21 lld4 � 22 .txe7 1lxe7
Black is a bit better here.
23 h3 lld7 24 llxd7 + l0xd7 26 lld1 �5
26 ••3 ltld7?1
This inaccuracy allows White's queen to
penetrate Black's queensidc and create suffi­
cient countctplay to draw. Though White
might have been able to save the draw any­
way, it was definitely worth torturing him
with 26...ltJg6 27llct 'ii'cS with a slight, but
enduring, edge.
27 1i'f4+ lL!f6 28 -.& a5 29 lld6 •xc4
30 llxb6 •xa2 31 ec7+ .td7
White also achieves a draw after 31.../lkl7
32lla6 'i'bt+ 33 �h2 'i'e1 34 'ii'f4+ ltJ£6 3S
lla7+.
32 llxf6+ �f6 33 •xd7 ft1 + 34 �2
� 35 .d8+ � 36 .g8+ �6 37"
•d8+ �c6 38 11'c8+ �5 39 •d7 + Wc4
40 •xg7 1lrd6+ 41 g3 a4 42 •xh7 a3:
43 •c2+ Wb4 44 �1 + �a5 45 ft3·
�%-%
 3 d4: 5 ... Jl.d6 and Fifth Move Alternatives

Summary
After 1 e4 eS 2 �£3 �f6 3 d4 �xe4 4 .i.d3 dS 5 �xeS, the move S...i0c6 is unattractive in
vit:w of the strategic problems Black faces after the doubling of the pawns. In our view, an
interest in S....i.e7 will be maintained; this continuation is not ambitious and docs not have any
positional flaws.
After S....i.d6 interest in the forced line 6 0-0 0-0 7 c4 .i.xeS 8 dxeS �c6 9 cxdS 'ifxdS 10
'tl'c2 �b4 t 1 .i.xe4 �xc2 12 .i.xdS will die away as it is thoroughly analysed and docs not
appear to be more promising for White in terms of gaining an advantage than, for instance, 7
ltlc.l2!?.

1 e4 e5 2 �f3 �f6 3 d4 �xe4 4 il.d3 d5 5 �e5 (D) il.d6

S....i.e7- Gan1e 6J
6 0-0 0-0 (D) 7 c4
7 �2- Gmne 62
7 �3 Gt1111e 61
-

7 Jl.xe5
...

7...�c6 Game 60
-

8 dxe5 (D) - Game 59

5�e5 6 ... 0-0 8dxe5

157
 CHAPTER EIGHT I
3 d4: Fourth Move
Alternatives

1 e4 e5 2 �f3 lt!f6 3 d4 lt!xe4 It is more na tu ral to arrack the centralised

Af te r the basic opening moves 1 e4 eS 2 knight with 5lL!bd2 (Game 66), which is one
lLif3 lLif6 3 d4 lL!xe4 the position arising of the most fashionable variations in the
after 4 .i.d3 dS 5lL!xeS is considered to be a Petroff Defence these days. Black can swap
tabiya. This brief chap ter cove rs deviations on d2, place his knight to c5, improve i ts
by White and Black on move four. position by means of ... f7-f5 or ... .i.fS, o r
Afte r 4 .i.d3 i t first appears tha t Black ca rry on with the development via 5...lL!c6 o r
must defend his knight wi th 4...d5. Howeve r, S... .i.e7 (p ro tec ting the yueen i n case o f an
in 1993 g randmaster Jacob Mu rey demon­ exchange on e4). Game 66 shows tha t all(!)
strated a pa radoxical al ternative in 4...lL!c6!? these con tinuations are yui te reasonable.
(Game 64). White can cap tu re the knigh t but
is unable to retain i t (5 .i.xe4 d5 and ... e5-e4, GaPJe64
or 5 d5 lL!c5 6 dxc6 e4). Since 1993 Morey's Tiviakov-Forintos
idea has been employed on many occasions Porto San Gimgio 1994
by m<xlem h'Tllllll masters.
4 dxe5 is often played instead of 4 .i.d3. 1 e4 e5 2 lDf3 ll)f6 3 d4 lDxe4 4 ..id3
Curiously 4 ... .i.c5, trying to sci7.c the initiative lt!c6!?
by a ttacking f2, is unpopular. Prom our pe r­
spective, Jines from opening manuals aimed a t
persuading readers of Whi te's advantage are
no t particularly convincing (see Game 65).
Tiu: continuation 4...d6 transposes to the
Philido r Defence ( 1 e4 c5 2 lLif3 d6 3 d4
lLlxe4 4 Jxe5lL!e4) and is not discussed he re.
In the main line wi th 4... d5 the move 5 .i.d3
(Game 65) is identical to 4 .i.d3 d5 5 dxeS,
which is pret ty harmless regardless of Black's
response (S...lL!c5, S... .i.e7 o r S...lL!c6). The
whi te pawn on e5 is almost no inconvenience
and often becomes an object of a ttack.

158

This move was a shocking novelty when it
first appeared in 1993. Nowadays, of course.,
ir has lost its surprise value.
6 ..txe4
White has a large number of options:
a) 5 .!Dxe5 .!Dxe5 (5....!Dc5 6 .i.c4 .!Dxe5 7
dxe5 d6 8 1i'f3 1Wd7 9 .i.e3 c6 10 Lc5 dxc5
II 0-0 favours White, whilst 5...d5 transposes
In 4 .i.d3 d5 5 .!Dxe5 .!Dc6) 6 .i.xe4 d5! 7
dxe5 dxe4 8 11'xd8+ �cl8 9 .!Dc3 .i.b4 10
.ig5+ �e8 11 0-0-0 .i.g4 12 l:d4 .i.xc3 13
hxc3 h6 14 llxe4 .i.h3! (a typical Shirov
move, here played against him) 15 gxh3 hxg5
16 l:tg1 llxh3 17 llxg5 �fB 18 .J:leg4 :C8 19
llxg7ltxe5 20 llg8+ �c7 with a likely draw,
Shirov-Timman, Wijk aan Zee 1998.
b) 5 dxe5 and now:
b1) 5...d5 6 exd61Dxd6 7 0-0 .i.e7 8.!Dc3
Jtg4 9 .!Dd5 .!Dd4 10 i..e2 (10 :C1 i..x£3 11
hrxf3lbc6 12 f4 f!f.l13 f5.!Dxf5 14 i..xf5 gxf5
15 'ti'£3 leads to a wildly complex position
with chances for both sides; note 15....!Dd4??
16.!Df6+ �fB 17 i..h6 mate!) to...lbc6 11 h3
.i.h5 12 :C1 0-0 13 .!Dxc7+ 11'xc7 14 .!Dg5
..i..xe2 15 1i'xe2 Wf6 16.!Dxe6 .J:lfe8 17 Wg4
Axc6 18 l:lxe6 fxe6 19 c3 l:lf8 with level
chances, Yakovich-Makarychev, Elista 1995.
b2) Also fine is s....!Dc5 6 0-0 .!Dxd3 7
'iixd3 d6 8 ..i..f4 dxe5 9 '1Vxd8+ 1Dxd8 10
i.xe5 lbc6 11 :C1 i..d7 12 i..xc7 .J:lc8 13
..i..e5 llxc2 14 lDbd2 i..c6 15.!Db3 .i.x£3 (an
improvement over 15...i..b4?! 16 .!Dfd4 .J:lc4
17 :Ccl l:lxct+ 18 .J:lxcl �d7 19 a3 i..e7 20
.!Dxc6 bxc6 21 l:ldt+ �c8 22lDa5 c5 23lDc6
with a clear edge for White in Bezgodov­
Cs.Horvath, Ljubljana 1995) 16 gxf3 i..b4 17
l:tc4 i..c5 18.!Dxc5 l:lxc5 with an <.>tJual posi­
tion.
c) 5 d5 was meant to give White an edge
once upon a time, but we cannot see how:
5....!Dc5 6 dxc6 e4 7 cxb7 (7 .i.c4 ex£3 8
cxd7+ i..xd7 9 Wxf3 1Ve7+ 10 .i.e3 .ic6 11
i.d5 .i.xd5 12 WxdS 1i'e6 13 1i'xe6+ .!Dxe6
14 .!Dd2 0-0-0 15 0-0-0 was drawn in V. Ko­
rnv-Frolyanov, Russia 2003) 7..ixb7 8 .i.e2
cxf3 9 .i.x£3 i..x£3 10 1fxf3 .i.d6 11 0-0 0-0
3 d4: Fourth Move Alternatives
t2.!Dc3 c6 13 :C1 ..c7 14 g3lDe6 15 1Wd3
i..e5 16 .i.e3 d5 (Sadvakasov-Koneru, Jodh­
pur 2003) and now 17 .i.d2 l:lab8 18 l:labt
.i.f6, with a level position, is a possible con-
tinuation.
5 ...d5
6 ..txh7
6 .!Dxe5 dxe4 7 .!Dxc6 bxc6 8 0-0 .i.d6 9
:Ct 0-0 10.!Dd2 (10 l:lxe4 i..f5 11 :C1 1i'h4
12 g3 'ti'h3 looks dangerous) 10...f5 11 .!Dc4
.i.di 12 b3 Wh413.!Dxd6 cxd6 14 1fd2 gives
a dynamicaUy balanced position, Hracek­
Barua, Moscow 1994.
6 i..g5 leaves Black with a wide range of
choices:
a) 6...f6 7 .!Dxe5 dxe4 8 11Vh5+ (8 .!Dxc6?!
bxc6 9 .i.e3 .id6 10 c4 0-0 11 c5 .i.e7 12
.!Dc3 f5 13 Wb3+ �h8 14 g3 f41? 15 gxf4- or
15 .ixf4 ..xd4 16 .i.e3 We5 with attacking
chances according to Frolyanov - 15..ig4
16 llg1 .i.f3 and Black had an attack in Yur­
taev-Frolyanov, Moscow 2003) 8...b>6 9.!Dxg6
hxf!f.l10 1fxg6+ �d7 11 i..xf6 (11 1Vf5+ �e8
with a draw is of course possible) 11..lth6 12
Wxh6 i..xh6 13 i..xd8 �xd8 14.!Da3 and we
evaluate this endgame as slighdy favourable
for White.
b) 6...1Vd6 7 dxe5 Wh4+ 8lDc3 dxc4 9 a3
WaS?! 10 .!Dd4 .!Dxe5 11 0-0 .i.d7 12 .!Dxe4
li)g6? (Black was in a bad way, but resistance
was still possible) 13 :Ct .i.e7 14 b41i'b6 15
..txe71Dxe7 t6.!Dc5 l:ld8 17 We2 1-0 Palac­
Kos. Fcldbach 1997. Instead of 9...'ifas.
159
 The Petroff Defence
.
Black can try 9.. 1Wxb2!? 10 ll:kls .tcs 11
llb1 1Wxa3 12lbxc7+ �fB 13lbxa8 exf3 14
'iVxf3 'ttxf3 15 gxf3 lbxe5 with a very un­
clear position.
c) 6...1Wd7 (the normal move) 7 .td3 e4 8
0-0 f6 9lle1
with another branch:
.
ct) 9. .fxg5 to c4 .tb4 11 lbc3 (11 cxd5
.txel 12 'lrxe1 lbe7 13 .txe4 g4 14lbc5
'iVd6 15 lba3 a6 with an unclear game is
given by Murey) 11...0-0 12 cxd5 lbxd4 13
lbxd4 exd3 14 llX.o6 llt7 15 1fxd3 favours
White.
.
c2) 9. ..te7 10 .tf4 cxd3 11 'ilfxd3 0-0 12
lbc3 .tb4 13 llc2 (Tunman-Hiibner, France
1993) and after 13...1Wt7 14lbb5 .taS the
chances are level (Hubner).
6 ...Axh7
7 dxe5
7lbxc5?! is weak: 7...'ttc7 8 0-0 lllxc5 9
1 60
llc1 (or 9 dxc5? 'iVh4 10 h3 .txh3! and Black:
wins) 9...lllf3+ 10 gxf3 .te6 1 I 'iVd3 llh3
with a very strong attack for Black.
7 ...i.g4 8 i.f4 11d7
8... g51? is also enticing: 9 .tg3 f5 10 exf6
1Wxf6 11 lbc3 .i.xf3 12 'iVxf3 .xf3 13 gxf3
.i.b4 14 0-0 .i.xc3 15 bxc3 0-0-0 with an
equal ending; or 9 1t'd3 .i.xf3 to gxf3 (Aiek•
seev-Bezgodov, Hoogcvecn 2002) and aftCJ:
10..llh6 11 .tg3 .i.g7 12 f4 gxf4 13 .i.xf4
llc6 Black should have sufficient counter­
play.
9 �bd2 11f5 1 0 i.g3 0..()-0 1 1 0-0 i.c5
1 2 a3
12 llct lZ\ct4 13 lllxd4 .i.xd4 14 lZIO
i.xb2 15 llb1 .i.c3 161lc3 d4 17lld3 11'g6
18 h3 lldh8 gave Black a strong attack in'
Svcshnikov-Pavasovic, Nova Gorica 1996.
1 2...i.b6
Or 12...�51? 13 c3 (13 b4? does not
work on account of 13...i.d4 , with a clear
advantage for Black) l3...d4 14 1Wa4 d3 with
substantial counterplay.
1 3 b4 �5 14 l:.e1 �d4 1 5 a4 a6 1 8
l:.a3 �f51
1 7 �f1
17 .tf4 would be met by 17...g51 18 .i.xgS
J:lgs with a strong attack (Forintos).
1 7 ...trucg3 1 8 �xg3 i.xf3 19 •xf3,
'W'xh2+ 20 �1 g6 21 l:.d3 I
Or 21 We2 1Wh4 22 llh1 11'c4+ 23 '1Vd3
llxh1 24 1Wxc4 dxc4 25 lbxh1 llc8 and th ,
endgame favours Black.
J

21 ...1rh41 22 llxd5 irc4+ 23 lld3 llxd3
24 •xd3 1rxb4 25 e671
'Ibis accelerates an uncomfortable posi­
tion into more trouble. After 2S l:te4 1WcS 26
..Wd2 :J,8 White is still only 'somewhat'
worse.
25...1rf4 26 l:le2?
The last chance was 26 1i'f3 'l'x£3 Zl gxf3
l:th2 28 � fxe6, though Black has every
chance of winning the endgame.
26 ...1rxg31 27 lle3 lZh1 + 28 We2
..g4+ 0-1
Game65
Nevednichy-Collas
Montpellier 2003
1 e4 e5 2 li:lf3 /t)f6 3 d4 li:lxe4 4 dxe5
An unusual move. One further option is 4
�xeS, when 4 ...d6 S �£3 transposes to 3
�xeS d6 4 �f3lbxe4 S d4 and 4...dS S ..t.d3
1ransposes back to 4 ..t.d3 dS S �xeS.
4 . d5
. .
4...d6 transposes to the Philidor Defence
(I c4 eS 2 �f3 d6 3 d4 �f6 4 dxeSlbxe4),
which is outside the scope of this book. We
will say, however, that White is generally
1hought to have a slight advantage.
4.....t.cS!? leads to very sharp play:
a) S .i.c4 �f2 (S....i.xf2+!? 6 �e2 ..e7 7
WJ3 f5 8 �3 c6 9 ..t.c3 .i.xe3 10 1i'xe3 with
unclear play is also possible) 6 ..t.xn+ �f8
(or 6...�n 7 ..dS+ �e8 8 'ifxcS �xh 1 9
3 d4: Fourth Move Alternatives
..t.gS d6 10 1i'e3 1i'd7 11 �c3 h6 12 ..t.h4 gS
13 �S �ffi and Black is under attack, but
nothing is clear) 7 'iVdSlbxh1 8 .i.hS fle7 9
..t.gS .i.f2+ 10 We2 'l'e6 11 �c3 h6 12
..t.d8!? lbc6 13 1Wxe6 dxe6 14 .i.xc7 with a
position almost impossible to understand
without dedicating days or weeks of your life
to it!
b) 5 1i'dS ..t.xf2+ (S...�xf2? 6 WxcS �xht
7 .i.gS f6 8 cxf6 hrxf6 9 ..t.h4 is probably los­
ing for Black)
and now:
bl) 6 �d11? f5 7 .i.c4 .z:tf8 (after 7...'fle7 8
llk3 c6 9 1i'd3 ..t.cs to :n bS 11 .i.b3 �f2+
12 .z:txf2 .i.xf2 13 'ii'xfS White has a strong
initiative for the exchange) 8 lblxl2 c6 9
�xe41? (after 9 'ii'd3 dS 10 cxd6 'ii'xcl6 11
�xe4 'iVxd3+ 12 ..t.xd3 fxe4 13 Le4 ..t.fS
14 .i.xfS llxfS the position has rapidly turned
drawish) 9. .cxdS 10 �d6+ �e7 11 ..t.gS+
.
.z:tf6 (1t...'.tc6?! 12 ..t.xd8 l1xd8 13 �gS+
�xeS 14 �dn+ Wd4 1S �xd8 �xc4 16 JZ.ft
and White has the advantage) 12 .i.xdS �c6
13 exf6+ gxf6 14 �xfS+ �e8 15 �d6+ �e7
16 .i.f4 with very unclear play. Of course
White could take a draw by perperual with 16
�fS+.
b2) 6 <.te2 f5 7 �3 (Black would get
away too easily after 7 exf6 �xf6 8 WeS+
�f8 9 .i.gS ..t.b6 10 �3 �6 It 1i'f4 h6 12
.i.xf6 'iVxf6 13 'ilrxf6+ gxf6, after which
White cannot fully justify the loss of a pawn)
7...c6 (7...�xc3+ 8 bxc3 .i.h4 9 �xh4 'iVxh4
161
The Petroff Defence
10 g3 'ile7 11 �f2/tk6 12 .lc4 'ilc6 13 'ilcS
with unclear play might be an improvement)
81Wd3 0-0 9 lh'te4 fxe4 10 Wxe4 dS 11 exd6
.l:te8 12 ...xeS+ ..xe8+ 13 Wxf2 -'.g4
(13....i.e61? looks safer; after 14 .i.d3 the
position remains very unclear) 14 -'.c4+ �h8
IS .J:r.e1 WfB 16 :C7l0d7 17 -'.d2 and White
had good attacking chances in Lozcnko­
Titlianov, Sverdlovsk 1974.
5 .i.d3
Sl0bd2 is considered in the next game.
5 exd6 makes no sense: 5 ... -'.xd6 6 .i.e2
0-0 7l0bd2 -'.f5 8 0-0l0c6 9l0xe4 -'.xe4 10
lOgs .lg6 11 .id3 •f6 12 .i.xg6 1fxg6 13 c3
ltad8 14 1Wb3 b6 and Black was a bit better,
Karpov-Noakh, Leningrad 1966.
5.-.lDc:S
This is the most active move, but Black
has reasonable alternatives:
a) s...lOc..c; 6 0-0 l0xd3 7 ..xd3 .i.e? 8
l0c3 c6 9 l0d4 0-0 to f4 f6 1 1 .i.d2 l0a6 12
a3 -'.cS 13 .i.e3 (13 Wh1 .i.xd4 14 •xd4
.i.f5 15 .l:tactl0c7 16l0e2 fxeS 17 fxe5l0e6
18 'ilb4 .d7 gave Black good play in Schif­
fers-Mason, Breslau 1889) 13...'ile7 14 b4
.i.b6 15 exf61bf6 is unclear.
b) 5... .i.e7 6 0-0 0-0 7 h3 f61? 8 c4 c6 9
'iVc2lOa6 to .i..xe4 (10 cxd5 cxdS 11 -'.xe4
dxe4 12 IZ.d1 'ilc7 13 ...xe4 .i.e6 is messy)
10...dxe4 11 'iVxe4 fxeS t2l0c3 ...c7 13l:<.tt
.i.f5 14 11'xe5 Wxe5 15l0xe5 .i.cS and Black
has sufficient counterplay for his pawn,
Drazic-Lanzani, Milan 2003.
162
6 0-0 .i.g4
6....i.e7 is also safe: 7 l0c3 .i.f5 8 'iletl?
l0b4 9l0d4 (9l0xe4 dxe4 10 .i.xe4l0xc2 11 j
.i.xc2 .i.xc2 is simply equal, but 9 e6!?l0xd3
to cxd3l0xc3 I 1 exf7+ Wxf7 12l0e5+ �g8
13 ..xc3 .i.f6 might give White slight pres­
sure) 9....i.g6 to f4 cS (or IO...l0xc31? 11 fS
l0cxa2l? 12 fxg6 fxg6 13 l0e6 Wd7 14
lOxg7+ Wd8 15 .i.h6 with a very complex
position) 11 f5 l0xd3 12 cxd3 cxd4 13 fxg6
hxg6 14l0xe4 dxe4 15 1Wxe41Wb6 16 b3 0-0
17 .i.b2 l:lad8 with level chances, Neved­
nichy-Piket, World Championship, New
Delhi 2002.
7 1Llc3
7 l0bd2 l0xd2 8 .i.xd2 l0d4 9 .i.e2
l0xe2+ 10 'ilxe2 'ild7 11 h3 .i.xf3 12 ..xf3
.i.. c5 13 c4 0-0-0 14 .l:tadt •e6 15 .i.g5 ltd7
16 .llxd5 .l:txdS 17 cxd5 was agreed drawn,
Romanishin-Makarychev, Frunze 1985.
7 �c3 8 bxc3 .i.c5
•.•
Or 8....i.e7 9 .i.£4 0-0 10 h3 .i.hS 11 llbl
'ilc8 12 g4 .i.g6 13 .i.xg6 (Kumosov­
Skatchkov, St Petersburg 2001) and now best
is 13...hxg6 14 1Wxd5 1We6 15 .l:tfd1 .l:tad8 16
1Wb3 l0a5 17 1Wa4 b6 with an unclear posi­
tion. Even so, the text move seems more
prudent.
9 .i.f4 0-0 1 0 h3 .i.h5
1 1 g471
This i.<> very risky. A more modest ap­
proach with 11 ltb l .i.b6 12 llbS �h8
(12..l0c7? 13 .i.xh7+! would be bad news

Indeed for Black) 13 g4 .i.g6 14 .i.xg6 fxg6
15 llxd5 'fle7 was called for, reaching a posi­
tion with chances for both sides.
1 1 .....tg6 1 2 .IZ.e1 ..tb6 1 3 ..tg3
1 3 ...f5
13...d4 is weaker: 14 cxd4lilid4 15 �xd4
Wxd4 16 .i.xg6 hxg6 17 'flxd4 .i.xd4 18
%lad1 .i.b6 and White has many ways to se­
cure the advantage in the endgame.
14 �2 ¢'h8 1 5 .i.h4 •ea 1 6 e6 txg4 1 7
hxg4.1Z.f4
Black can also play the safer 17 ... .taS! 18
11td2 d4 19 �xd4 (not 19 e7? llxf31 20 .ixg6
hxg6 21 Wxf3 .i.xc3 22 'flf4 .ixat 23 llxa1
�xe7 24 9xc7 �6 and Black has the ad­
vantage) 19....ixd3 20 11Vxd3 �xd4 21 1Vxd4
1rc6+ 22 'fle4 .ixc3 �' 1rxc6 bxc6 when
nlthough White will have enough compensa­
lion for a draw, he will never be able to win.
18 �g3 ..txt2+ !?
3 d4: Fourth Move Alternatives
A valid alternative is 18...lla41? 19 .i.b5
lla3 20 'irxd5 :Xc3, when it is probably
more pleasant to be Black.
19 �xf2?
An understandable mistake as White is
scared of allowing his king to go too far into
the open. However, this is the worst of the
two evils and Black now has a very strong
attack.
The alternative is 19 Wxf4 ...ffi+ 20 .te;
'fld6+ and now:
a) 21 lle5lilie5 22 .L.f2 �xg4+ 23 Wxg4
h5+ 24 Wh3 .i.x£5+ 25 Wg2 .ie4 offers
Black compensation, but it is very likely that
White wiU survive the attack.
b) 21 WgS! suggests to Black that he
should settle for a draw, as after 2L..'flc7+ 22
Wf4 .i.xh4?! (22.. .1t'd6+ is perpetual check)
23 �xh4 1fxh4 24 llht 1ff6 25 1fxd5 l:l.d8
26 1t'c5 it is not so obvious that Black has
sufficient compensation for the exchange.
The White king looks strange, but Black's
piccc:s are not that active.
19 .. ...te4 20 ..tg5
20 .i.e2 1i'xe6 21 .i.g3 llf7, with ... l:laffi
to follow, does not help White at aU.
20 ....1Z.xg4
Or 20... llxf3+!? 21 Wxf3 .i.xf3 22 Wxf3
d4 and Black has goo d chances. However,
the game continuation makes more sense.
21 llg1 •xe6 22 .IZ.xg4
Black's attack is also irresistible after 22
.i.xc4 dxe4 23 �h4 e3+ 24 We2 llf8.
163
The Petroff Defence

22...1rxg4 23 .i.e2 llf8 24 1Wg1 .i.xf3 0-1 should play 10...'1Ve71? 11 lbbS lbba6 12
..xdS lidS 13 'ffc 4 1Wxe5 with a messy posi­
Game66 tion) 11 lbb5lbe6 12 c4 a6 13 cxdS axbS 141
Smirin-Aiterman dxe6 ..xe6 15 .i.xbS+ c6 112-lfz Volokitin�
Haifa 1995 Mikhalchishin, Portomz 2001. In the final:
position 16 .i.e2 lZXI7 17 f4 l:lxa2 18 l:lxa.Z
1 e4 e5 2 li)f3 ll)f& 3 d4 li)xe4 4 dxe5 d5 11'xa2 19 0-0 appears to offer White some
5 li)bd2 chances.
c2) 6lbxe4 .i.xe4 7 .i.d3lbc6 8 0-0 .i.e1
9 l:le1 .i.xd3 10 1i'xd3 '1Vc.l7 11 .i.f4 0-0-0 12
a3 Wg4 13 .i.d2 f6 14 .i.c3 d4 15 ..tb4 l:lhe�
(t s....i.xb4?! 16 axb4 lbxb4 17 Wc4 lbc6 t!i
b4! with a clear plus for White, e.g.J
IS...c;ilbS?I 19 bS lbxeS 20 lbxeS fxeS 21
l:lxa7! 1-0 Glek-Mikhalchishin, Zurich 2001)�
16 .i.xe7 :Xe7 17 exf6 gxf6 18 l:lxe7 lbxe"f!
19 fle1 .!Og6 20 h3 'iWd7 21 l:le4 with a sligh�
edge for White (Giek).
d) s....!Ocs 6 .!Ob3

5 . .i.e7
..

Black has many alternatives to this modest

move:
a) S...ltk6 6 .i.bS .i.d7 7 1ie2 a6 8 .i.xc6
.i.xc6 9 tbd4 lbxd2 10 .i.xd2 .i.d7 11 0-0-0
11'h4?! (11...1le7 12 l:lhcl 0-0-0 13 f4 may be
a bit better for White, but the advantage is
not that obvious) 12 e61 11xd4 13 cxd7+
Wxd7 (Mainka-Schmidt, Gennany 1996) and
now after 14 .tgs 1i'a4 15 110 White would
have a very promising attack.
b) 5...15 6 exf6 lbxf6 (6...1i'xf6?! 7 lbxe4 and now:
dxe4 8 .tg5 1l'd6 9 1We2 would be bad for d1) 6....!0xb3 7 axb3 ..te7 8 .i..d3 .!Od7 9
Black) 7 c4!? d4 (7....i.d6!? 8 cxdS 0-0 9 .i.e2 c3 .!Ocs I 0 .i.c2 ..tg4 11 b4 lbe6 t 2 1i'd3
lbxdS looks equal) 8 .i.d3 .i.e7 9 0-0 0-0 10 '1Vd7 13 0-0 .i.hS (the weakening 13...�,>6?{
lte1 cS!? (an improvement over 10. lbc6?! 11
.. was played in Ivanchuk-Kasimd7.hanov
a3 aS 12lbgs .i.d6 13 tbdf3 h6 14 cS! hxgS Elisra 1998; now after 14 .!Od4 0-0 15 .i.. a4 c6
15 cxd6 11xd6 16 lbxgS with a clear edge in 16 h3 .!Oxd4 17 hxg4 .!Oe6 18 f4 .i..d8 19
Sveshnikov-Raetsky, Podolsk 1992) 11 b4!? Wh1 White would have enjoyed a clear edge)
cxb4 12 1i'c2lbc6 13 cS and White has com­ 14 lbd4 .i.g6 15 .!05 .i.d8! 16 :dt c6 17
pensation for the pawn. .!Od6+ ..xci6 18 exd6 .i..xd3 with equal
c) s....i.t5 might not lead to equality, for chances (Kasimdzhanov).
example: d2) 6....i..g4 is also sound: 7 h3 .i..hS S
cl) 6lbd4!? .i.g6 7 h4!? .i.cS 8lb2b3 hS 9 .i.e2 .!Oc6 9 .!Ofd4 .i..xc2 10 ..xc2 .!Oxb3 11
lbxcS lbxcS 10 .i.e3 1Wd7 (maybe Black .!Oxb3 ..d7 12 .i..f4 0-0-0 13 0-0-0 11'c6 14

1 64

lld2 �c7 with level prospects, Kharlov­
Akopian, Moscow 1991. More complex is 8
c4 .!tlxb3 9 'A'xb3 �xO 10 Wxb7 (10 gxO
ltlc6 1 I cxdS .!tlxcS with unclear play is pos­
sible as weU) 10....!tld7 with unclear play, for
example 11 gxO .!tlxeS 12 �e2 c6 13 cxdS
..icS 14 dxc6 0-0 and Black has good com­
pensation for the pawn.
c) s....!tlxd2 6 'A'xd2!? (6 �xd2 �e7 gives
n standanl position that has not been investi­
).,'llted very much; a possible continuation is 7
.idJ .ig4 8 h3 .i.hS 9 0-0 0-0 to .llcl lbd7
11 c3 ltlcs 12 .i.c2 l&6 13 �f5 c6 14 g4
.ig6 15 'A'c2 �gS 16 .llad1 �xd2 17 'A'xd2
fle7 with unclear play, Kotsur-Zulfugarli,
lsL1nbul 2000) 6...�e7 (or 6...c5!? 7 �d3
tlk6 8 cJ h6 9 0-0 �e6 10 :C1 .ic7 1 I h3
1Wd7 with chances for both sides) 7 'A'f4 (1
.i.bS+ c6 8 .i.e2 0-0 9 0-0 f6 10 c4 fxeS 1 1
cxdS cxdS 12ltlxe5 .!tlc6 leads to equality: 13
.!tlxc6 bxc6 14 b3 cS 15 L3 .i.b 7 16 llact
nc8 17 �g4 D.c7 18 D.cet d4 19 1t'd3 ..LI6
20 �c1 Wh8 21 f4 :C7 22 .llxc7 ..xe7 23
..id2 �e4 1/2-'12 Svidler-Kramnik, Dos Her­
manas 1999) 7...0-0 8 'A'g3 �f5 (8 ...f6?! 9
..th6 :n 1o 0-0-0 ltlc6 11 �c4 �e6 12
.i.xdSI �xdS 13 e6! �xc6 14 .llxd8+ :Xd8
15 .i.f4 gives White a clear edge) 9 c3 lieS 10
..ie2 .!tlc6 11 0-0 .d7 and Black seems to be
no \vorse.
6..i.b5+
More critical is 6 .!tlxe4 dxe4 7 'A'xd8+
.i.xd8
3 d4: Fourth Move Alternatives
and now:
a) 8.!tlg5 isn't dangerous: 8.. ..i.f5 (8 ....!tlc6
9 .!tlxe4 ..if'S 1() .ld3 lDxeS 11 c!iXI6+ cxd6
12 �x£5 0-0 13 0-0 �b6 14 .i.f4 �cS 15
.llad1 J:lae8 16 d left White with some
chances in lordachescu-Chi Fengtong, Ulan
Bator 2002) 9 .lc4 ..i.xg5 to .lxgS �7 11
�f4 .!tlb6 12 .lb3 aS and we think Black wiD
equalise.
b) 8 .!tld4 R....id7 9 �f4 .!tlc6 10 0-0-0
gS!? (after 10...0-0 11 .lc4 .!tlxd4 12 .llxd4
.l£5 13 e6!? fxc6 14 lld7 White has obtained
an unplt:asant pressure) 11 .lgJ hS 12 .lbS
h4 13 e6 fxe6 14 Lc6 (14 lDxc6?! hxg3 15
.!tlxd8 .llxd8 16 �xd7+ .llxd7 17 .ll xd7 �d7
18 fxg3 Wd6 and Black's activity gives him a
good endgame) 14...bxc6 15 .leS ltfB 16
.!tlb3 Itt'S (16...:Xf2 17 lDcS .lc8 18 .!tlxe4
llxg2 19 D.hg1 l:lxg1 20 llxg1 looks very
strong for White) 17 .ld4 .le7 18 g4 llf7 19
.i.c3 and White's position was preferable,
Zviagintsev-Stohl, Rethymnon 2003. f-low­
ever, instead of 12...h4, Black should proba­
bly play 12....!tlxd41? 13 .ixd7+ �d7 14
ltxd4+ We6 15 h3 (15 h4!?) 15 ...h4 16 .lh2
c3!? 17 fxe3 .i.e? with compensation for the
pawn through the safe, yet active, king.
6 c6
...
Also possible, although appearing less
logical, is 6...�d7 7 Wc2 c6 8 �d� ltlcs 9
.!tlb3.!tlxd3+ 10 cxd3 cS with unclear play.
7 i.d3&5
7....!tlxd2 8 .lxd2 cS 9 c3.!tlc6 10 0-0 0-0,
165
 The Petroff Defence
with level chances, is simpler.
8 �e2 .i.g4 9 llXI4 .i.xe2 1 0 •xe2 tLlbd7
1 1 0-0 tLle6
11 ...0·0 12 f4 lbe6 13 ib2b3 l0xd4 14
l0xd4l0c5 is 'calmer', although the position
remains terribly double-edged.
1 2 �xe& fxe& 1 3 •g4 tLlxe5
Black must take accept the challenge.
13...Wf7 gives White the initiative after 14
l0f3 1i'c7 15 .i.gS l:lhe8 16 l:lfe1 - Black's
king is not really safe.
14 Wxg7 tng& 1 5 tLlf3 Wd7
After 15...1i'd6 16 lOgs .i.xgS 17 .i.xgS
..d7 18 ..c3 White is slighdy better accord­
ing to Greenfeld.
16 c4
1 6. .••g8?1
Black should probably play 16...Wf8 17
11fc3 llgs 18 :C1 .tb4 19 �3 dxc4 20
:dt+ Wc7 21 Wc2 .i.cS with a very unclear
position (but not 19....i.xe1? 20 ..xb7+ Wd6
21 cS+ WxcS 22 .i.e3+ Wd6 2.1 .i.c5+1 and
mate is imminent).
1 7.d4 llf81?
17....i.d6 18 cxdS exdS 19 :C1 looks to
give White a slight edge.
1 8 Wxa7 .:Xf3 1 9 1i'xb7+ Wd&
After 19 . We8? 20 ..c8+ .i.d8 21 .xc6+
. .
Wfl 22 cxdS White's position is crushing.
1 66
20 lle1 e5
20...l:ld3? would lose to 21 cS+! WxcS 22
.i.e3+ Wd6 23 .i.cS+! WxcS 24 llacl + with
mate to follow- a very nice line.
21 b3 lld3?1
21...l:lf6! is best, although this can be
strongly met by 22 f4!? (22 .i.a3+ We6 23
..xc6+ Wf5 24 'ird7+ We6 25 g4+ Wxg4 26
1i'xe6+ :Xe6 27 cxdS .i.xa3 28 dxe6 Wf5
looks fmc for Black) 22.. .llxf41 (22. e4 is
. .
weaker due to 23 .i.a3+ We6 24 f5+! l:lxf5 25
1i'xc6+ Wfl 26 l:lfl !l0f4 27 l:lxf4! l:lxf4 28
1i'c7 and Black is lost) 23 cxdS 1i'c8 24 �
.i.d8 25 .i.a3+ WxdS with unclear play. Our
computer prefers Black, but that assessment
can change in an instant!
22 .i.a3+
22 .•• c5?
Black is better off playing 22. .We6. After
.
23 cxdS+ l:lxdS 24 Wxc6+ .i.d6 25 l:lad1
l0e7 26 l:lxeS+ WxeS 27 .i.xd6+ Wf5 Black
seems to survive miraculously. However, the
stronger 23 .i.xe7l0xe7 24lLe5+ WxeS 25
1i'xe7+ Wd4 26 l:le1 gives White a very dan­
gerous attack.
23 .i.xc5+ ! wxc5 24 llac1 dxc4
24... Wd6 25 cS+ We6 26 11fa6+ would pick
up the rook.
25 .:Xc4+ 1 -0
 3 d4: Fourth Mo ve A lternatives

Summary
I�xpcrience has proven that in the case of 4 dxe5 d5 5 ltlbc12 (and 5 i.d3) Black maintains
e<.Juilibrium in various ways. However, like in some other systems of the Petroff Defence, it is
difficult for Black to move from a solid equalisation to seizing the initiative. That's why we
recommend trying the sharp 4....tc51? - we can't fmd any advantage for White after either 5
1Wd5 or 5 .tc4.
Murey's brilliant discovery (4...ltlc61?) looks artificial and poor at ftrst sight. In fact, it is un­
clear how White can gain <.-ven a slight advantage. In any case, 5 i.xe4 d5 6 .txh7 (as in Game
64) cannot be recommended for White since it gives Black active counterplay.

1 e4 e5 2 �f3 �f& 3 d4 �xe4 4 dxe5 (D)

4 .td3 ltlc6 (D) - Gan1e 64
4...d5 (D) 5 .i.d3- Game 65
5ltlbd2 - Gan1e 66

4 dxe5 4... &6 4... d5

167
 CHAPTER NINE I
3 d4: Black Plays 3 ...
1 e4 e5 2 lLlf3 ll:Jf6 3 d4 exd4
In this chapter we wiD discuss 3...cxd4,
Black's main alternative to 3...l£lxe4. foirsdy
though, let's briefly go over a couple of other
tries for Black. The move 3...d6 simply trans­
poses to the Philidor Defence, which is out­
side the scope of this book. The symmenical
reaction 3...d5 drew players' attentions as far
back as 1900 (Pillsbury-MarshaiQ, but became
especially noticeable after the Stein-Bronstein
game (Ibilisi 1966). At the same time, this
game contributed to the decline of 3...d5. It
has been proven that White doesn't have to
transpose to the other lines of the Petroff
Defence with either 4 dxeS or 4 l£lxe5 l£lxe4 5
.i.d3. Instead, the continuation 4 exdS! cxd4 5
.i.bS+ breaks the symmetry in White's favour
(see the notes to Game 67).
Moving onto 3 ...exd4, let's leave aside the
Urusov Cr.unbit (4 .i.c4), since this is the
Bishop Opening (1 e4 eS 2 .i.c4 l£lf6 3 d4
cxd4 4lL\f3), or the Two Knights Defence Of
4 ...l£lc6 is played). Instead the games in this
chapter arc devoted to the tabiya after 4 eS
llk4.
White's main choice is the obvious 5
'i6'xd4 (Game 67-69). After S...dS 6 exd6
t£lxd6 Black faces a dreary prospect of fight·
ing for equality with only a slight chance of
sei;dng the initiative. This is why one of the
168
exd4
numerous ideas from Morozevich, S ... f5!?, is
an enticing option (sec the notes to Game
67). Bacrot appears to have chosen a correct
reply, with his position being preferable after
6 .i.c4 .i.cS 7 ..xeS!?lL\xcS 8 .i.gS.
Let's get back to the main line S ... dS 6
exd6 t£lxd6. The specific features of this
position are the two open files in the centre,
the queen on d4 that is vulnerable but puts
pressure on the enemy camp, and the black
knight on d6 that has J.,rained stability but
prevents Black's development. To avoid con­
fusion with the transposition of moves, we
collected all the lines where White declines to
play the most natural move 7 l£lc3 in Game
69. In this case it is interesting to contrast the
typical fianchetto ....i.f8-g7 with the ma­
noeuvre .i.e1-d2-d.
Black has various ways to protect himself
after 7 l£lc3 l£lc6 8 ..f4, three of which arc
studied in Game 67. The arrangement wid1
...JJ..e7 and ....i.e6 is effectively met by the
queenside castling plan. The attack on the c2-
pawn by ....i.f5 is not popular in view of 9
.i.b5, but White doesn't seem to have a sub­
stantial advanl:3b>c. The main part of Game
67 concentrates on the surprising 8. ..t£1f5!?­
the knight moves for the fourth time our of
eight to prepare attacking the enemy queen
with ....i.d6. This idea was rriet.l for the ftrst
 3 d4: Bisek Plsys 3 . . . exd4

time during the Kasparov-Karpov match
(New York, 1990). A fourth option, the very
fashionable fianchetto of dte black bishop
with 8 . g6, is covered in Game 68. Here cas­
. .
ding long looks more risky for White than
against 8....i.e7, but White still stands better.
The paradoxical S .i.b5 attracted our at­
tention (fal employed it a few times) and is
c3 dxc3 9 1t'xd8+ �xd8 10 l0xc3 �e7 11 0-0
lld8 (Stein-Bronstein, Thilisi 1966) and here
Stein could have obtained a sb.eable advan­
tage with 12ll)a4 .i.b6 13 b3 .i.c7 14 .i.e3.
4 e5
4 .i.c4 transposes to the Bishop's Open­
ing (2 .i.c4ll)f6 3 d4 exd4 4ll)o).
4 �
...

covered in the annotations to Game 70. The only sensible move. If 4...ll)d5?! then
Black can turn the game into the Berlin De­ 5 '1Vxd4 c6 6 .i.c4lt!b4 7 0-0 lt!xc2 8 .i.xf7+!
fence of the Ruy Lopez by means of 5 .lik6,.. �f7 9 ..c4+ �e8 10 .xc2 and White's
transposing to a line that is harmless for advantage is clear-cut.
Black (1 e4 e5 2 ll)f3 ll)c6 3 .i.b5 ll)f6 4 d4 5 1Vxd4
exd4 5 e5 ll)c4). A wormy alternative is 5 ...c6
6 1Wd4 'tWaS 7 c3 ll)xf2 (Keres) 8 0-0 ll)h3!
(F.uwe) - anomer in-between move that de­
stroys White's pawn chain.
The main part of Game 70 concentrat<..'S
on me interesting move 5 1t'e2. Steinitz dis­
cussed dtis move in his book Modmt Chess
lflslnlctor (1889). White attacks the centralised
knight from the e-ftle in order to prevent
Black from strengthening the knight with
either the d- or the f-pawn. However, after
s....i.b4+ 6 �dl !? (Steinitz) it is possible to
improve support the knight with 6...d5 7
exd6 f5, when Black has the initiative. That's 5 d5
•..

why White prefers the calmer 6 ll)bd2, while Morozevich has risked the remarkable
Black often retreats with 5 ...ll)c5 after 5 1t'e2. 5 ... f5!? here:
Black's play in the opening in Game 70 is still
considered to be a perfect model.

Game67
Cabrera-Collas
Malaga2003

1 e4 e5 2 �f3 l0f6 3 d4 exd4

3...d5?! is inferior:. 4 exdS! (4 dxcS ll)xe4
transposes to 3...ll)xe4 4 dxe5 dS) 4...exd4
(after 4...e4 5 ll)e5 ll)xd5 6 .i.c4 .i.e6 7 0-0
.i.d6 8 .i.xd5 .i.xd5 9ll)c3 c6 10 '1Ve2 White
has a clear advantage- Yusupov) 5 .i.bS+ c6
6 dxc6 bxc6 (6...1t'a5+ is critical but White is a) After 6 cxf6 ll)xf6 7 .i.g5 lt!c6 8 Wh4
much better after 7 ll)c3 bxc6 8 ll)xd4! cxb5 1t'e7+ 9 .i.c2 �+ 10 'irxb4 .i.xb4+ 11
1J 'IVO 1Wc7 10 ltldxb5!? 1Wc7+ 11 .i.e3 .i.b7 lt!bd2 0-0 it is doubtful that White has any
12 'irf4 - Znosko-Borovsky) 7 .i.e2 .i.c5 8 advantage.

1 69
Th e Pe troff De fen ce
b) 6 .i.c4 .i.cS 7 'ifxcSI? (the queen sacri­
fice is, of course, only temporary; instead 7
1i'd5 .i.xt2+ 8 We2 1i'e7 9 lDbd2 lDf6 10
1i'd3 dS 1 1 .i.bS+ c6 1 2 Wxf2 /0g4+ 13 We2
cxbS is very complicated) 7 lDxcS 8 .i.gS
..•
1i'xgS 9 lDxgS lDc6 10 0-0 lDe6 1 1 ltlxe6 (1 1
ll)f3 b6 1 2 .ids .i.b7 13 c4 0-0-0 14 lDc3 h6
1S l:tad1 gS gave Black good countetplay in
Bacrot-Moro7..evich, Biel 2003) 1 l ...dxe6 12
f4 a6 1 3 lDd2 and Black's position is slightly
cramped.
6 exd6 lLlxd6 7 lL!c3 �6 8 11f4 lLlf5
Other than 8 ...g6 (sec Game 68), Black
has two inferior options:
a} 8 ....i.e7 9 .i.d3 (also good enough is 9
.i.c3 .i.e6 to 0-0-0 0-0 1 1 lDgS .i.xgS 1 2
1i'xgS 1i'e7 1 3 9xe7 tDxe7 1 4 .i.cS l:tfd8 1 5
�e2, when White's bishop pair will be use­
ful) 9 ...�e6 10 .i.e3 �f6 1 1 0-0-0!? .i.xc3 12
bxc3 11'£6 (after 1 2...�xa2 13 c4 'if£6 14
:bet White keeps the initiative) 1 3 ...xf6
gxf6 1 4 ltld4 �d4 1 5 .i.xd4 Wc7 16 l:thc1
b6 17 f4 c5 1 8 .tf2 f5 1 9 g41? and Black was
under severe pressure, Parkanyi-Krivolapov,
Gyongyos 1 998.
b) 8. .. �f5 9 .i.bS
and now:
bl) lf 9 ...�xc2? then 1 0 lDcs wins in­
stantly.
b2} 9 ...lDxbS? 10 lDxbS 1i'e7+ 1 1 �fl l
and c7 drops.
b3) 9..i.e7 10 .i.xc6+ bxc6 1 1 llks 0-0
1 2 lDxc6 1i'e8 1 3 lDxe7+ 1i'xe7+ 14 .i.e3
1 70
�xc2 1 5 :tel .i.d3 1 6 tnds 'ird8 17 1i'd4
�a6 (after 17....tg6 Sax gave 1 8 0-0 lDfS 19
'ireS lDxe3 20 lDxe3 as clearly better for
White) 18 lDxc7 l:tb8 (Sax-Yusupov, Rotter­
dam t 988). Now White should play 1 9
lDxa6! WaS+ 20 ltlb4 ltlf5 (not 20...l:txb4? 21
'ifxd6! lld4+ 22 b4, winning immediately) 21
l:tcSI lDxd4 22 LaS l:txb4 23 0-0 with good
winning chances.
b4) 9...9e7+
with a further split:
b41) 10 Wf1 .te4 1 1 .i.xc6+ .i.xc6 1 2
lDe s 'ire6 1 3 lDxc6 bxc6 1 4 'ir f3 'ifc4+ 1S
Wg1 .i.e7 1 6 b3 1i'a6 17 �b2 0-0 and White
has a slight edge. Instead Klovans-Harman,
correspondence 1 967 continued 1 2... 0-0-0?!
1 3 lDxc6 bxc6 14 'ifa4 lDbS 1 S 1i'a6+ Wb8
1 6 .tc3 9b4 17 'ifxc6 tnd4 1 8 11fa6 .i.cS 19
a3 'ifb7 20 1i'xb7+ Wxb7 21 l:tct �b6 22 g3
when White was simply a pawn up.
b42) The simple 10 .i.e3 is also promising:
l O...lDxbS 1 1 lDxbS 9b4+ 1 2 'iVxb4 �xb4+
1 3 c3 .td6 (not 1 3 .. i.aS?! 14 b4 �d3 1 S a4
a6 16 lDbd4! .tb6 1 7 lDxc6 bxc6 1 8 .txb6
cxb6 19 lDeS 0-0-0 20 lDxc6 l:the8+ 21 Wdl
lldS 22 tnd4 and White is a pawn up) 1 4
lDxd6+ cxd6 1 S 0-0-0 .i.e6 1 6 :Xd6 .txa2 1 7
.tcS! 0-0 1 8 l:txc6 bxc6 19 .i.xf8 Wxf8 20
tnd2 .tdS 21 8 l:te8 22 Wc2 and White's
better structure gave him the edge in Matu­
lovic-Kholmov, Sochi 1 968.
9 .i.d2
White has a range of options, but Black

should be fine in all cases:
a) 9 lDbS .i.b4+ 10 c3 .laS 1 1 1i'e4+ ..e7
12 1i'xe7+ lDcxe7 1 3 .lf4 lDdS 1 4 .ics f6 1 S
0-0-0 a6 1 6 llxd5 axbS 1 7 .lf4 c6 and the
JX)Sition is level.
b) 9 .i.c4 .id6 1 0 ..e4+ .le7 (1 0.....e7 is
what White wants: 1 1 .id2 ..xt.-4+ 12 �xe4
0-0 1 3 0-0-0 J.e6 1 4 .i.xe6 fxt.-6 1 5 l:hel
l:r.ae8 1 6 �fgS eS 1 7 .lc3 a5 1 8 a4 b6 1 9 f3
h6 20 lDxd6 lDxd6 2t lt)c4 lDf5 22 b3 with a
pleasant advantage, Bc:rzinsh-Nciksans, Riga
2003) 1 1 .ld2 0-0 1 2 .idS .if6 13 0-0-0
lDcd4 1 4 g4 lle8 1 5 ..f4 lDxf3 1 6 'Wxf3 lDd4
17 1i'g2 c5 (the more passive 17...c6 leaves
White with the initiative after 1 8 .le4 .le6
1 9 gS .leS 20 .i.e3) 1 8 .i.e3 'Wb6 1 9 lDe4
i..e7 20 c3 .le6 21 .ixe6 lDxe6 was unclear
in Tiviakov-Ye Rongguang, Groningen 1 997.
c) 9 .ibS .i.d6 1 0 ..e4+ ..e7 1 1 .ig5 f6
12 .id2 .id7 1 3 0-0-0 1i'xe4 1 4 lDxe4 ole7
1 5 g4 a6?! (more accurate is 1 5...lDd6 16
lDxd6+ .lxd6 1 7 :del+ �fB 1 8 l:btg1 lle8
1 9 llxe8+ ¢>xe8 with equality - A7mai­
parashvili) 16 .lxc6! Qnstead 1 6 .i.c4 lDd6 17
lDxd6+ J.xd6 1 8 l:lde 1+ was agreed drawn in
Kasparov-Karpov, World Championship
!Game 1 01, New York 1 990) 1 6 ....i.xc6 1 7
llhe1 .lxc4 1 8 llxe4 lDd6 1 9 lle2 � f7 20
�4 lDc4 21 .if4 and White has an edge.
9....td6 1 0 1t'a4 •e7+
10....i.d7 1 1 0-0-0 lDcd4 1 2 llet+ �£8 13
'ifc4 with equality is also acceptable.
1 1 .te2 .td7 1 2 0-0·0 �cd4 1 3 1i:lxd4!?
3 d4 : Black Pla ys 3 . . . exd4
A bold and unclear <:JUeen sacrifice. l11e
simple 1 3 .i.bS leads to equality after
1 3 ... lDxb5 1 4 lDxbS 0-0 1 5 llhe1 1i'f6 1 6
�3 .i.c6.
1 3....i.xa4 1 4 �xf5 1t'd7 1 5 �xg7 + �dB
Instead 1 S ...We7?1 16 llhet Wf6?1 walks
into trouble: 1 7 �hS+ �g6 1 8 g4 and Black
will lose too much material in escaping from
the mating net.
1 6 .tg5+ �c8 1 7 ZZ.d41 ZZ.g8?1
Now White regains the sacrificed material
with interest. Black should play 17 .....c6 and
now:
a) Not 1 8 l:txa4? ..xg2 1 9 .ig4+ fSI and
Black wins more material.
b) Also weak is 18 �xa4?! 1i'xg2 1 9 .i.g4+
Wb8 (19... 5?! 20 lDxfS ..xht+ 21 Jldt
..xh2 22 lDxd6+ �b8 23 lDfl is unclear) 20
:Ct hS! 21 lDxh5 bSI 22 lDc3 llxh5! and
Black is much better.
c) Best is 1 8 :hd ll 1 8..ltg8 1 9 .i.g4+
Wb8 20 .i.h6 'Wxg2 with a massively compli­
cated position.
18 .tg4 ZZ.xg7 1 9 .txd7+ .txd7 20 .tf4
ZZ.xg2 21 �d6 cxd6 22 �d6 �7 23
ZZ.hd1
The game has simplified leaving White
with a useful extra pawn.
23 .tc6 24 ZZ.f6 ZZ.xh2
••.
Or 24...:m 2S h4 and White keeps his
material advantage.
25 lbf7+ �6 26 1i:ld5+ �.5
Now the king is in danger. Black had to
171
The Petroff Defenc e

take the knight. even though the rook ending
is probably lost.
27 a3 :h8 28 b4+ Wb5 29 :d4
29 ••• 85
The king is also caught after 29...a6 30 c4+
Wa4 31 �b2 .i.xd5 32 llxdS b6 33 .J:.ff5 b5
34 1lf6 liaS 35 cxb5 axb5 36 .J:.dd6.
30 a4+ Wxa4 31 �c3+ ¢'a3 32 b5
:h1 + 33 ¢'d2 1 -0
r------.
Game 68
Stefansson-Yusupov
Eupen 1994
1 e4 e5 2 �f3 �f6 3 d4 exd4 4 e5 �
5 11xd4 d5 6 exd6 �xd6 7 .!003 .!006 8
1rf4 g& 9 ..ie3
Preparing to castle long without worrying
about the prospect of doubled pawns. Alter-
natively:
a) Getting n."3dy for kingside castling with
9 J..bS is also reasonable: 9 ...J..g7 10 0-0 0-0
1 1 J..xc6 bxc6 12 J..e3 l:.b8 1 3 l:lab1 aS 1 4 a."
(it's unclear after 1 4 J..cs llc8 1 5 llkl4 J..e S
16 'irf3 J..b7 17 .J:lfd1 •gS) 14...l:.e8 15 J..a7
l:.b7 1 6 J..cS ( 1 6 J..d4 gives Black the option
of 1 6...g5!? 17 .cl f6 HI h3 .i.f5 with a
sharp position) 1 6...J..f5 1 7 1i'a4 ll:k4 1 8
l:.lxlt ti'c8 19 J..d4 li)xc3 20 .i.xc3 J.. xc3 21
bxc3 l:.bS and White may have a tiny advan­
tage, Romanishin-Smyslov, Leningrad 1 977.
b) 9 .i.d2, like 9 J..e3, prepares castling
queenside, but this time preventing the pos­
sibility of doubled c-pawns, although of
course the bishop is slightly less active:
9...'..c7+ 10 J..e2 .i.e6 1 1 0-0-0 J..g7 1 2 h4
h6 1 3 l:.he1 0-0-0 14 .i.d3 ti'f6 1 5 1i'xf6
.i.xf6 1 6 .i.xg6 .J:.dg8 (this is the right rook;
instead 1 6...l:.hg8 17 hS J..g4 18 J..h7 l:.h8 19
..td3 J..xh5 20 .i.e2 J..g6 2·1 lbd5 ..tg7 22
.i.c3 .i.xc3 23 li)xc3 l:.he8 24 li)h4 gave
White an edge in Geller-Smyslov, Moscow
1 991) 17 hS ll:k7 1 8 .i.f4 J.. xc3 19 bxc3
li)xg6 20 hxg6 l:.xg6 21 li)h4 llhr4 22 g3 h5
and the position is level.
9 ....i.g7 1 0 0-0-0
1 0 .i.d3 is a worthwhile alternative. For
example, lO....i.e6 1 1 0-0-0 1i'f6 1 2 �
ti'xf4 1 3 J..x f4 0-0-0 (Yusupov assessed
1 3....i.xc3 1 4 bxc3 0-0-0 1 5 li)xe6 fxe6 1 6
l:.het l:.hfB 1 7 .i.g3 .J:.de8 1 8 f3 as a bit bet­
ter for White) 1 4 l:.he1 .i.xc3 1 5 bxc3 .i.xa2
16 c4 (Black has an edge after the weaker 16
J..xd6 l:.xd6 1 7 c4 .J:.d4 1 8 lDxfl llg8 19 �5
lDxcS 20 .J:lxeS .J:.gu8) 16 ...h6 1 7 lbe4 lDxe4
(less accurate is 1 7 ...lba5 1 8lbxd6+ cxd6 19
cS! d5 20 .J:.e7 .J:.d7 21 llxd7 �xd7 22 J..bS+,
as in Raetsky-Rodionov, correspondence
1 982; now 22...�c8 23 lt:t .i.c4 24 l:.c8+
.l:lxe8 25 .i.xe8 would have been marginally
in White's fitvour) 1 8 .l:lxe4 l:r.he8 1 9 llde 1
:Xc4 20 llxe4 g5 21 .i.d2 ..ib3 22 h4 and
White has reasonable compensation for the
pawn.
1 0 . . ...ie6
This is more flexible than 1 (J...0-0, which

1 72

perhaps commits the king a move too early:
1 1 h4 h6 12 .i.cS!? .i.e6 13 .i.bS a6 14 .i.xc6
bxc6 1 5 .id4! (the greedy 1 5 .ixd6 cxd6 1 6
l:lxd6 allows Black dangerous counterplay
after 1 6 ...'iWb6 l7 lDd4 ltab8) 1 5...f6 1 6 .icS
l%f7 1 7 llhel ..Ll7 1 8 �4 WeB 19 'ifg3
�b7 20 �a4 g5 21 'ifb3 �aS 22 'ifd3 and
White had a powerful bind in lvanchuk­
Akopian, Lucerne 1997.
1 1 lbg5
A promising alternative is 1 1 .ibS!? .i.xc3
12 bxc3 0-0 13 �gS fle7 14 .icS when
White's activity compensates for his weak­
ened strucrure.
11 0-0 1 2 h4 �e5 1 3 •f3 h5
•..
Grabbing a pawn with 1 3....ixc3 14 bxc3
.ixa2 leads to great danger after 1 5 hS tbcs
1 6 Wf4.
14 lbxe6 fxe6 1 5 �3 •f6 1 6 �g5
�f4+
Or 16 ...flxf2 1 7 .id3 .i.xc3 1 8 bxc3 tlJf5
19 .i.c4 and White's initiative continues.
1 7 �xf4 •xf4+ 18 �b1 Aae8
If 1 8...lClf5 then 19 f3 consolidates
White's edge.
1 9 �e2 lbe5
Taking the f-pawn is still risky: 19 ... Wxf2
20 .llhft 1fc5 21 llxf8+ �xf8 22 g4 with a
powerful attack.
20 lZhe1 .J:lf5 21 f3 1ib4 22 a3 �6 23
g4 l:lf4 24 gxh5 lbdc4 25 �xc4 lbxc4 26
b3 lbxa3+ 27 �c1 •f2
If 27...Wc6 then 28 'ifg2 llf7 29 /.De4 with
3 d4: Bisek Pla ys 3 . . . exd4
a winning attack for White.
28 lle2 •xf3 29 1i'xf3 llxf3 30 �2 c6?
Now White wins by force. 30...gxh5 31
�13 l:.xd3 32 cxd3 �18 woukl have given
Black some drawing chances.
31 lbe4 lbb5 32 c4 lbc7 33 lld7 lba6 34
hxg6 l:lef8 35 g7 Jl8f7 36 lld8+ �xg7
37 llg2+ �h7 38 lbg5+ 1 -0
Game 69
Bonch-Osmolovsky - Baranov
Moscow 1954
1 e4 e5 2 lL'lf3 lbf6 3 d4 exd4 4 e5 lbe4
5 •xd4 d5 6 exd6 lt!xd6 7 j.d3
White's other alternatives to the usual 7
�c3 are:
a) 7 .if4 ltlc6 8 'ilfd2 'ii'e7+ 9 .ie2 tlJe4
1 0 'ilfe3 ltlb4 1 1 Wet .if5 (chasing material
with 1 t ...1Wc5 is risky: 12 0-0 �xc2 - or
1 2...'ilfxc2 1 3 'ii'c: t !? with an unclear position
- 13 tlJc3 tlJxc3 1 4 1Wxc2 ltldS 1 5 We4+
.ie6 1 6 ltac1 and White has a dangerous
initiative) 1 2 0-0 0-0-0 with a double-edged
position.
b) 7 .igS
and now:
bl) 7 ...ltlc6 8 Wc3+ (or 8 'Wc3 f6 9 .if4
1fc7+ 10 .i.e2 .ie6 1 1 0-0 0-0-0 and Black
has equalised) 8....i.e7 9 lDc3 lDf5 10 .ixe7
lDcxe7 1 1 WeS 0-0 1 2 lld1 ltld6 1 3 .i.d3
ltlg6 1 4 flg3 We7+ 1 5 'it>J2 ltlf5 1 6 .i.xf5
.i.xf5 17 ..t>ct lladB 18 .!lkl4 .i.c8 19 llhe1
1 73
The Petroff Defence
� was level in Konstantinopolsky­
Smyslov, SvenUovsk 1 993.
b2) Black should seriously consider the di­
rect 7 ... f6. For example, 8 .i.f4 c!lk6 9 11t"d2
�fS 10 .i.e2 11t"e7 1 1 0-0 0-0-0 1 2 lte1 lfr4
1 3 11t"ct gS 14 .i.d3 �h6!? 1 5 llk3 (the wild
t S li)h4!? leads to unclear play after 1 5...gxh4
16 �xh6 llhg8; note that White must avoid
1 7 1i'f4? ltxg2+! 1 8 Wxg2 :gS+ 1 9 �1
/bg3+) 15 ...gxf4 16 .i.xc4 .i.xe4 17 l:lxc4
flg7 1 8 11'f1 .l:thg8 1 9 .1Le1 1i'g4 20 �h1
�f8 was unclear in Stcinitz-PiUsbury, St Pe­
tersburg 1895. However, after 12 ..'iWf7!? 1 3
.i.d3 .i.g4 Black has a slight plus.
7 fke7+
•••
Also interesting is 7 ...c!lk6 8 11'f4 f!!> 9 0-0
�7 and now:
a) 10 llc!t+ leads nowhere after 1 0....i.e6
1 1 c!bgS 0-0 1 2 c!bxe6 fxe6 1 3 .g3 (Hiibncr­
Scgal, Dresden 1969 continued 1 3 11t"g4?!
•f6 14 11Vg3?! c!bb4 1 5 ltfl c!bfS 16 11'xc7?
.1Lc8 1 7 1i'xb7 c!bxd3 1 8 cxd3 c!bg3! 1 9 hxg3
ltxcl! 20 ltxct 11'xf2+ 21 �h1 .xg3 22 1l'e7
1i'c3! and Black won) 1 3....i.d4 14 :C2 (not
1 4 ltfl? c!be4! 1 5 .i.xc4 .i.xf2+! and Black
wins immediately) 14 ...c!bf5 1 5 flh3 c!bcs and
Black has excellent counterplay.
b) This sacrificial to .i.d2!? is the only war
to cause Black trouble: 10...11'f6 (White has
good comp<.>nsation if Black tries 1 0....i.xb2
1 1 .i.c3 .i.xa 1 , for example 1 2 .i.xa I O-Il 1 3
.i. f6 'iWd7 14 1ih6 c!bes 1 5 .i.g7! c!bxg7 16
c!bgs :es 1 7 1Wxh7+ �f8 18 flhs+ �e7 19
1 74
..xg7 1i'd5 20 .i.xg6 and the king is in real
danger) 1 1 11xf6 .i.xf6 12 .i.c3 �xc3 1 3
c!bxc3 and White may have a tiny advantage.
8 i..e3 .ltlf5?1
1bis wastes too much time - instead
Black should develop with 8.. ..i.f5. For ex­
ample, 9 ill c!bc6 1 0 11'f4 .i.xd3 1 1 cxd3
'iWe6 12 0-0-0 .i.e7?1 (instead 12 ...1fg6 1 3
lthe 1 0-0-0 is unclear) 1 3 d4 11'f5 14 d 5 c!bb8
1 5 Wxf5 c!bxf5 16 .i.f4 c!ba6 1 7 g4! c!bh4 18
c!bxh4 Lh4 1 9 d6! 0-0-0 20 c!bbs c6 21 d7+!
:Xd7 22 c!bd6+ l:lxd6 23 .i.xd6 and White
was much better in Spassky-Kholmov,
Rostov on Don 1 960.
9 i..xf517 .i.xf5 1 0&3 Wb4
White's lead in development is significant
whichever way Black plays:
a) 10....i.xc2 1 1 :ct llk6 1 2 'ff£4 c!bb4 13
0-0 and the initiative continues. For example,
1 3...c!bd3?! 14 'ti'c4 c!bxct 1 5 Let and, with
c!bd5 also threatened, White is winning.
b) 1 0...c!bc6 1 1 .f4 1lt"b4 1 2 0-0-0 and c7
is again impossible to defend.
1 1 fke5+ i..e6 1 2 0-0-0 cltlc6 1 3 fkxc7
Ac8 1 4 fkf4 1ra5
14 ...•xf4 is met by 15 ..i.x£4 .i.b4 1 6 c!bc4
and White is a pawn up.
1 5 ..g5
Simple is best. Instead 1 5 lOgS .i.xa2 1 6
c!bxh7?1 leads to trouble after 1 6. ..1Lh71 1 7
We4+ .i.e6 1 8 Wxh7 1Wat+ 1 9 c!bbt c!bb4
when Black's attack is very dangt:rous.
1 5 11ra6 1 6 .l:he1 10b47
•••
 3 d4: Black Pla ys 3 . . . exd4

Now Black is losing. 1 6...h6 1 7 1Ph5 i.e7 5 tDc5

.•.

1 8 i.gS! 0-0 1 9 i.xe7 lhxc7 20 lhd4 is un­ This simple move is more reliable th:m
pleasant but at least Bl.1ck can play on. 5...j.b4+
1 7 c!i)d4 Zlxc3

and now:
Allowing a beautiful finish. 17 ...f6 1 8 a) Steinitz suggested the strange 6 Wd1,
'irh5+ f!!l 19 11b5+ 11'xb5 20 �lxbS would for example 6...d5 7 exd6 f5 8 tbgs (after 8
have lost slowly and painfully. dxc7 11'xc7 9 lhxd4 lhc6 10 c3 lhxd4 1 1
1 8 1Vd8+1! cxd4 i.d7 Black has a tremendous initiative)
A mating combination d1at exploits the 8. . .0-0 9 lhxe4 fxe4 10 11'c4+ Wh8 1 1 dxc7
power of discovered checks. (Lipschutz-Showalter, USA 1 896 continued
1 8 Wxd8 1 9 lDxe6+ We7
..• 1 1 11'xb4 .i.g4+ 1 2 i.e2 i.xe2+ 1 3 'itxc2
Or 1 9 ...We8 20 lhxg7+! i.xg7 21 i.g5+ lhc6 14 ..el?! Wfxd6 1 5 �d1 L8 1 6 b3 c3
�fB 22 :ld8 mate. 1 7 .ta3 'lrf4 and Black had a wonderful posi-
20 �g5+ f6 21 lDd8+1 1 -0 tion) 1 1 ...'ifc7 1 2 cxb811' .l:txb8 and Black
.------., obviously has a powerful initiative - Porreca.
Game 70 b) 6 lhlx12 lhxd2 7 i.xd2 'il'e7 8 0-0-0
Tal-Kholmov lhc6 9 i.xb4 Wxb4 10 11'c4 b6 11 lhxd4
AlmaAla 1968 i.b7 1 2 i.bS 0-0-0 1 .3 a3 lhxd4 (or 1 3 .. ."ti'c5
1 4 lhxc6 i.xc6 1 5 i.xc6 dxc6 1 6 f4 and
1 e4 e5 2 tDf3 tDf& 3 d4 exd4 4 e5 tDe4 White has a defmite cdbrc) 1 4 axb4 lhb3+ 1 5
5 1i'e2 cxb3 i.xc4 1 6 .l:td4 i. £5 1 7 f4 and despite his
A rare alternative to 5 ..xd4. White has bi?.arrc quecnside structure White was a litde
another unusual try in 5 .tbS c6 (5...lhc6 better in Rodrique?. Andrcs-Ginzburg, San
transposes to the Ruy l .opez: 1 e4 c5 2 lht'3 Martin 1 995.
�c6 3 i.b5 lhf6 4 d4 ed4 5 e5 lhc4) 6 'ifxd4 6 c!i)xd4 lDc6 7 �e3
'ifa5+ Qess active is 6...lhc5 7 i.c4 lhe6 8 This is more testing than 7 lbxc6 dxc6 (the
'1Ve4 dS 9 exd6 i.xd6 1 0 0-0 when White has com.'Ct recapture; instead 7 ... bxc6 8 .i.c3 'irh4
a tiny edge) 7 c3 lhxf2!? 8 0-0 (White should 9 1i'c4 'lrxc4 1 0 i.xc4 lha4 1 1 i.b3 lhb6 12
consider 8 'ifxf2 1fxb5 9 a4 11'd5 1 0 i.e3 lhd2 aS 1.3 a3 a4 1 4 .ta2 i.n6 15 .txb6 cxb6
.te7 1 1 0-0 0-0 1 2 lt:la3 with good play - 1 6 llk4 left White a touch better in Spassky­
Yusupov) 8...lhh3+! 9 b'Xh3 'lrxb5 10 11'f4 Vistiniet7.ki, Tallinn 1 959) 8 lhc3 i.£5 9 i.e3
11'd3 t l lhd4 'ifg6+ 12 �hl d5 with a com­ h51? 1 0 f4 l&.:4 1 1 lhxe4 i.xe4 1 2 1if2 ..d5
plex posicion, Zapata-Castro, Colombia 1 999. and Black has good counterplay.

1 75
The Petroff Defence
7 �xd4
.•.
Safer than 7...l£lxeS?I 8 f4 � 9 l£JbS l£la6
to .i.d4+ 'ile7 1 1 f5 1i'xc2+ 12 .i.xc2 l£lh4
13 f6 g6 14 0-0 when, despite the exchange of
queens, White has a dangerous initiative.
8 hd4 1ih4
A more active approach than 8...l£le6 9
.i.c3 dS 10 exd6 'Wxd6 1 1 l£ld2 .i.d? 1 2 l£lc4
'ife7 (or 12 ...Wcs 1 3 l£les with a promising
attack, e.g. t3..J:d8? 1 4 1i'f3 'We? 1 S 0-0-0 c6
1 6 .i.c4 hS 1 7 l:the1 l:th6 1 8 ..i.b4 was already
winning for White in Rac::tsky-Varlamov,
correspondence 1 983) 1 3 l£leS 0-0-0 1 4 W'e3
and White has a pleasant edge.
9 ..te3 itb4+ 1 0 c3
1 0...1re4
Not 1 0 ...l£ld3+? 1 1 1i'xd3 1Wxb2 1 2 WbSI
9xa1 1 3 .i.c4 L7 14 0-0 0-0 1 5 'ifb3 when
the queen is trapped.
1 1 f4 d5 1 2 �d2
Or 1 2 exd6 Ld6 13 l'Od2 We6 1 4 f5 1i'e7
1 5 f6 gxf6 with an unbalanced position.
1 2...1Vg6 1 3 lDf3 c6 1 4 0-0-0 ..te7 1 5
llg1 h5 1 6 1rf2 �e4
Black can also consider 1 6 . b6, for exam­
. .
ple t 7 llXI4 � 1 8 1i'c2 0-0 19 .i.d3 cS 20
l£1£3 .i.f5 21 llXI2 l:tfd8 22 c4 d4 23 l£lxe4
(not 23 .i.xe4?1 d31 24 1i'c3? .i.xe4 25 l£lxe4
1i'xe4 26 lldel b5! when Black was much
lx:tter in Ractsky-Matsukcvich, correspon­
dence 198S) 23... dxc3 24 l£lf6+ gxf6 25 .LfS
1i'h6 with a murJ..l' position.
1 7 1rc2 b5!?
1 76
Instead the slow 1 7...0-0 is marginaUy in
White's favour after 18 .i.d3 .i.g4 1 9 .:dfl
h4 20 l£ld2.
1 8 ..td3
1 8 ...f6?!
Black should continue the qucensidc pawn
storm with 1 8...a61? 1 9 l:tdf1 c5 when he has
a strong attack.
1 9 exf6 ..txf6 20 ..tc5
20 l£ld41? .i.d7 21 f5 with the initiative is
also good.
20-....tf5 21 llge1 <i>t7 22 �2
And here the simple 22 l:lc3 aS 23 l:ldc1 is
promising.
22 . . �xc5
.
Black had to be very careful: 22 .. l[hc8?1
allows a nasty trick after 23 l£lxc4 .i.xc4 (not
23 ...dxe4? 24 .i.xe4 ..Le4 25 Wb3+ .i-dS 26
l:txd5 l:txet+ 27 .lld t+ l:lc6 28 l:td7+ and
White wins) 24 l:lxe4 and White has an edge
because 24... dxe4? loses to 25 Wb3+ l:te6 26
.i.xb5 .i.e5 27 l:tft !.
23 ..txf5 1rxg2 24 b4
TI1e aggressive 24 l:lg1 allows Black to es­
cape with a safe edge after 24...'Wxh2 25
.i.g6+ We7 26 l:tdet+ �dB 27 11f5 Wc7 -
the king is no longer in danger.
24 ...lDb7 25 1i'd3 llae8 26 llg1 1te2
27 ..tg6+ <i>t8 28 llge1 •xe1 29 llxe1
llxe1 + 30 Wc2 �6 31 lDf3 lle7 32
� llh6 %-%
Tal gave the explanatory variation 33 l£lc6
l:lg6 34 'ifg6 :CZ 35 Wd3 l:th2.
 3 d4 : Bla ck Pla ys 3 . . . exd4

Summary
Statistically 3...l0xe4 has not scored overwhelmingly better than 3 ... cxd4. However, the �"Xam­
ples shown here after 3...cxd4 were mainly played more than a decade ago.
The continuations 4 e5 lbe4 5 .ib5 and 5 '1Ve2 do not worry Black any longer, but with 5
11xd4 the situation is more serious. It appears that after 5 ...d5 6 exd6 l0xd6 7 c!lk3 c!lk6 8 1i'f4
itlf5!? Black maintains equilibrium. However, with the symmetrical pawn structure and easier
development, White has many aggressive possibilities that Black needs to be ready for. Conse­
quently, the general interest in 3 ...cxd4 has dropped, but Morozevich's experiment with 5 ..f5 is
.

interesting.

1 e4 e5 2 l0f3 l0f6 3 d4 exd4 4 e5 l0e4 (D) 5 •xd4

5 1i'e2 - Ca1ne 70
5 ...d5 6 exd6 l0xd6 (D) 7 l0c3
7 .id3 Game 69
-

7 lOc& 8 ..f4 (D) l0f5

... - Ca/lle 67
8.. g6 Ca1ne 68
. -

4. . . l0e4 6. . . l0xd6

1 77
 CHAPTER TEN I
Third Move Alternatives
For White
1 e4 e5 2 lbf3 l0f6
After 2. ..lt)f6 White has two equally im­
ponant continuations: 3 d4 and 3 lbxe5.
Theory pays much less attention to other
white possibilities, but statistics prove that
White often chooses 3 .t.c4, 3 d3 and espe­
cially 3 lbc3. In this case the positions
reached are not always in the domain of the
Petroff Defence, as other openings can also
arise.
After 3 lbc3 (Games 71 -72) Black is of
course able to transpose into me Four
Knighrs (with 3...lbc6). However, Pillsbury
employed the move 3....i.b4, Alekhine calling
this continuation the 'Ruy Lopez for Black'.
The extra tempo lbgt-f3 plays a significant
role here, but it has to be used properly. 4
.i.c4 (Game 7Z) is best met by 4....i.xc3 5
dxc3 d6 with comfortable development for
Black's pieces. More promising is 4 lbxe5
(Game 71), after which it is easy for Black to
win his pawn back but it is more difficult to
achieve an equal position. After 4...0-0 5
lt:k-13 .t.xc3 6 dxc3 lbxe4 7 .i.e2 the modest
7...d6 can put Black under long-term posi­
tional pressure. Black should play 7...d5 and
... c7-c6 with 1,•ood prospects of equalising.
The continuation 3 .i.c4 (Game 73) offers
Black the interesting opportunity to play
3 ... lbf6 with the transposition into the Two
1 78
Knights Defence, but 3...lbxc4 seems to be
strongest move. A fter 4 lbc3 Black can re­
turn to the Two Knights again with 4...lbc6
(by the way, this variation is favourable for
Black). However, it is even more promising
to hold onto the extra pawn with 4 ...lbxc3 5
dxc3 f6. After this Black's pawn chain be­
comes a formidable force. I t is especially
useful to activate the pawn chain with ...c7-c6
and ...d7-d5, neutralising the bishop on c4.
White should aim to undermine the chain
tactically by taking advantagt: of the insecure
position of Black's king (especially with the
push f2-f4). Objectively speaking, Black's
chances are preferable in these sharp posi­
tions. However, he must be very careful:
there are numerous examples of Black facing
mpid defeats in this line.
After the modest 3 d3 (Game 74) we have
a position that also arises from the Alekhine
Defence (1 e4 lbf6 2 d3 e5 3 lbf3) - in effect
White is playing the Philidor Defence with an
extra tempo. In this manoeuvring battle
White is rarely able to gain an advantage.
Black can maintain the symmetry with 3...d6,
but 3...lbc6 followed by ...d7-d5 is more
promising. Black can develop the dark­
squared bishop to c5, e7 or g7, although he
normally delays the choice until move six
(after 4 c3 d5 5 lbixl2 aS 6 .i.e2). ln brcneral

Black docs not have any difficulties and often
maintains a small space advantage.
Game 71
Svidler-Akopian
World Team Ch., Lucerne 1997
1 e4 e5 2 ll:lf3 ll:lf& 3 ll:lc3
White is trying to escape to the l'our
Knights, but Black will not allow it!
3 ....tb4
4 ll:lxe5
The critical move. 4 .i.c4 is discussed in
the next b>ame, while in Ragozin-Kan, Lenin­
grad 1 936 White: played 4 �S lllids S exdS
e4 6 tlJd4 0-0 7 c3 .i.cS 8 d3 .i.xd4 9 cxd4
1Vc7 10 .i.e3. Now after 10...d6!? 1 1 .i.e2
.i.f5 1 2 llct c6 Black should be okay.
4 0-0
...
Also possible is 4...'tle7 S �3 (S tlJg4!?)
S ... .i.xc3 6 dxc3 tlJxe4 7 .i.e2 dS 8 0-0 0-0 9
lLlf4 c6 (but not 9...lld8? 10 lLlxdS 'it'es 1 1
c4 c6 1 2 .i.f4 1i'e6 1 3 .i.g4! f5 14 .i.xf5 and
White wins - Alckhine) 1 0 c4 dxc4 1 1 .i.xc4
.i.f5 1 2 'it'e2 1le8 1 3 llet 1i'd7 14 .i.c3 and
White was slightly berter in Alapin-Aiekhinc,
Carlsbad 191 1 .
5 ll:ld3 .txc3 6 dxc3 ll:lxe4 7 .te2
This is the critical line (although there is a
limit to how critical such a position can be).
Instead Galdunts-Ractsky, Aachen 1994 con­
tinued 7 .i.c3 lieS 8 1Wf3 d6 9 0-0-0 lLld7 1 0
h3 (or 1 0 lLlf4 lL!df6 I t h3 .i.d7 1 2 .i.c4 .i.c6
Third Mo ve A lterna tives for White
and Black is fmc) 1 0...a6 1 1 .i.e2 b5 12 1if4
.i.b7 1 3 .i.f3 lLldf6 14 tlJcs lLlxcS I S .i.xcS
.i.xf3 16 9xf3 lL!e4 1 7 .i.e3 lle6 1 8 9f5
lieS 1 9 .g4 'it'cS 20 .i.d4 'lfxg4 21 hxg4
lle6 and a draw was agreed. [f anyone should
feel relieved in tlte final position, it is White.
7 d5
...
Very natural. Black has also played 7...l:.eS
8 0-0 d6 and now:
a) 9 llcl lL!c6 to .i.fl .i.f5 1 1 f3 lLlf6 1 2
..t.gs lL!c:s 13 'iVdz h6 t4 .i.h4 b..s 1 5 .tf2
lLldS?! (1 S ...lL!c4 16 .c l l:xc1 17 lLlxe1 lL!b6
1 8 a4 would have left White only slightly
better) 16 lL!xeS dxeS 17 .l:lad 1 tlJf4 1 8 l%xc:SI
'iVxd2 1 9 llxe8+ l%xe8 20 .llxd2 and Black
didn't have any compensation for dte pawn
deficit, Lau-Ractsky, Bad Ragaz, 1 994.
b) 9 lLlf4 tiJc6 (or 9 ...lL!d7 1 U c4 lL!ffi 1 1
a4 aS 1 2 .i.e3 .i.d7 1 3 lL!ds .i.c6 1 4 l%e 1
lLlg6 1 5 .i.ft h6 16 f3 lL!f6 11 Wd2 b6 1 s b3
and White has an edge in SoUeveld-Pikct,
Bundcsliga 2002) 1U c4 ..tf5 1 1 lL!ds h6 12
b3 lL!e7 1 3 .i.b2 lL!xd5 1 4 1i'xdS 'ifg5 lS
.i.d3 (1 S f4 1Vg6 16 'iVxb7 lL!cS 1 7 'iVf3
.i.xc2 leaves Black in a solid condition)
·
1 5...c6 1 6 Wd4 l%e6 17 h4 'ifg6 1 8 l:ae1
llaeS 1 9 lle3 and White is slightly better,
Chudinovskikh-Raetsky, Orel 1 992.
8 0-0 c& 9 ll:lf4 Ae8
Black is also doing okay after 9 ... .i.f5, for
example l 0 c4 d4 1 1 .i.d3 lieS 1 2 f3 lL!d6 1 3
c 5 i.xd3 1 4 Wxd3 tiJbs 1 5 a4 lL!c7 1 6 c 3 (1 6
b4 aS 17 bS cxbS 1 8 axbS lL!ct7 gives Black
179
The Petroff D e fence
l."XccUcnt countcrplay) 16... dxc3 17 'irxc3 a5
18 �h5 f6 19 i.h6 liXI5 20 'irc4 g6 21 �3
�7 22 J:tfet f5 23 .i.d2 :Xet+ 24 :Xel
..fB and Black has equalised, AJams­
Rozentalis, Copenhagen 1997.
1 0 c4
to i.c3 is best met by 10...�16 (10...�d7
11 c4 dxc4 12 i.xc4 �5 13 'irxd8 :Xd8 1 4
.i.c2 i.fS 15 g4 i.d7 1 6 f3 �6 17 l:tad1
li:XIc4 18 i.c1 i.e8 19 b3 �b6 20 .lb2 gave
White a bit of pressure in Biro-Chetvt:rik,
Nagykanizsa 1 995) 1 1 i.d3 i.fS 12 'ifh5 g6
13 'ifh6 .lxd3 14 cxd3 �fS 15 'iFh3 � 16
:ae1 'irf6 17 .i.d2 �5 18 d4 llk4 19 Act
Ae4 with level chances, Benjamin-Yusupov,
Munich, 1994.
10 d4
.••
1 1 .id3
After 11 i.f3 li)g5 12 i.g4 �6 13 Ac t
�6 (13...li)xf4 14 llxe8+ 1i'xc8 1 5 i.xf4
180
i.xg4 16 'irxg4 c5 17 i.h6 rfl 1 8 Wft may
be slightly better for White) 14 i.xe6 i.xe6
15 �xe6 llxc6 16 Axe6 fxc6 17 i.d2 'irf6
the game is level, Adams-Hubner, Dortmund
1996.
1 1 . ..1tla6 12 c3
12 0 �c5 13 lle l li)xd3 14 cxd3 i.fS
might even preferable for Black.
1 2....if5!?
An interesting pawn sacrifice. After 12..c5
13 ..c2 �f6 14 i.d2 White has a pull.
1 3 g4 .tg& 14 f3 �5 1 5 .txg& hxg& 1 6
cxd4
Black can answer the alternative capture
16 ..xJ4 with 16...'irc71?.
1 6... b5
The alternative 16...1if6!? 17 d5 llad8 18
h4 �h7 19 g5 1ie5, with play for the pawn,
is also strong.
1 7 cxb5
17 ..d3 is powerfully met by 17...'1fd6 18
cxb5 �b4 19 1Wb3 'ifxd# 20 Wht cxb5 21
a3 �a6 22 1ixb5 llk5 when Black is very
active.
1 7...cxb5 1 8 d5
Or 18 h4 �6 19 �xe6 Axe6 20 i.g5
1id6 21 ..d2 llaeB 22 i.f4 1id5 with suffi­
cient play for the pawn.
1 8 llc8 1 9 d&
.•.
19 i.d2 :C4 20 �g2 �c7 gives Black
good play, while after 19 h4 �h7 20 g5 �b4
Black's active pieces also promise him a
pleasant game.

1 9 .Ac6 20 d7 lle7 21 •d5 llc5!
.•
This is stronger than 21 ......xd7?1 22 ...xg5
...d4+ 23 Wh1 lles 24 Wh4 g5 25 Wg3 gxf4
26 i.xf4, after which Black looks a bit worse.
22 ..d6?
22 'il'b3 llxd7 23 i.e3 lle5, with a com­
plex position, was the right path.
22 ...llxd7
Black misses a direct win with 22...1lc2! 23
1Wd3 (23 Wxa6 l0xf3+! 24 llxf3 l:let+ 25 J:lft
Wxd7! and Black wins; or 23 h4 l:lxd7 24
1We5 l0xf3+1 25 llx£3 llc7 26 lt.I3 Wc7 and
White )uses his queen) 23...l0b4 24 Wxb5
llxd71 25 i.c3 (25 Wxb4 lld l ! and White has
no defence) 25...1ldd2!!
(maybe it was this brilliant move the play­
ct"S overlooked) 26 Wxb4 l0h3+ 27 l0xh3
J:lg2+ 28 Wh t llxh2+ 29 �gl llcg2 mare.
23 •xa6 �xf3 + ?
Black could stiU have gained the advantage
Third Mo ve A lterna tives for White
with 23...1lc21 24 i.e3 (24 .xbS lOx£3+1 25
J:lx£3 lldt+ and Black wins) 24 .. .lldd21 25
l:lf2 1ldt+ 26 Wg2 llxf2+ 27 .Lf2 llxa1 and
Black has all the chances.
24 llxf3 l:ld1 + 25 �g2 11h4. 26 .!i)e2
llc2 27 llf2 •xg4+ 28 /t)g3 l:lxf2+ 29
�xf2 •d4+ 30 �2 %-%
Game 72
Lev-Alterman
Ramal-Can 1992
1 e4 e5 2 /t)f3 ll)f6 3 lbc3 .ib4!? 4 .ic4
4 . 0-0
. .
Black has reasonable alternatives here:
a) 4..ixc3 5 dxc3 d6 6 ...c2 l0bd7 7 h3
lOcS 8 i.d3 0-0 9 0-0 lle8 10 c4 i.d7 1 t b4
l0c6 12 llet aS 13 bS c6 14 c3 1Vc7 with
e'Jual prospects, Gaponenko-R.1etsl-y, Kras­
nodar 1995.
b) 4...d6 5 0-0 i.g4 6 h3 i.hS 7 g4 i.g6 8
d3 c6 (safer is 8.,.i.xc3 9 bxc3 l0c6 10 a4 0-0
1 1 l:lbt l:lb8 when Black is close to equality)
9 l0c2 l0bd7 10 c3 .ia5 1 1 b4 i.c7 12 i.b3
lOfB 1 3 ltlg3 l0c6 14 g5 lOhS 1 5 i.xe6 fxe6
1 6 l0xe5 dxeS 1 7 lLlxhS 0-0 and Black had
good compensation in Chemyshov­
Yusupov, Ohrid 2001 .
5 0-0
5 tOds is rather unexciting: 5 ...t0xd5 6
.i.xdS c6 7 i.b3 dS 8 0-0 ..i.g4 (or 8...dxe4 9
lOxeS tlX17 10 d4 lOxeS 1 1 dxeS Wxd1 1 2
l:r.xdt .i.g4 13 l:ld4 llfd8 14 l:lxd8+ l:[xd8 15
181
The Petroff Defence
.i.e3 with level chances) 9 h3 .i.hS 10 d3
dxe4 11 dxe4 .!LXI7 12 We2 'flc7 with an
equal position, Spiclm:mn-Marshall, Buda­
pest 1 928.
5 ....i.xc3
Black usually uses this chance to inflict
double pawns on White, but he can also try
s ... d6!?, for example 6 d3 .i.g4 7 h3 .i.e6 8
lbc!S .i.xdS 9 exdS h6 10 c3 .i.aS 1 1 d4 exd4
(our improvement over 1 1 ...e4 1 2 �h4 c6 1 3
dxc6 �xc6 1 4 � f5 d S 1 S .i.b3 .i.c7 1 6 f3
when we prefer White, Priehoda-Chetvcrik,
Martin 1 996) 12 �xd4 �bd7 1 3 �f5 �S
14 .i.b3 J..b6 with more or less level chances.
6 dxc3
Black has nothing to fear after 6 bxc3
lDxe4 7 �xeS dS R .i.b3 aS 9 a4 l:leR 1 0 .!LXIJ
lDc6.
6 ...d6
6...�xc4?! 7 �xeS L-. better for White.
182
One line continues 7...c6 8 llc 1 dS 9 l:lxe4!
dxc4 (9...dxe4? to J..xf7+ �h8 1 1 "ifhs J.. f5
1 2 lL\g6+! J..xg6 1 3 J..xg6 leads to mate) 10
:04 1ie7 t 1 lDxc4 and White was a pawn up
in LGuliev-Smougalev, Moscow 199S.
7 i.g5
Practice has also seen 7 1ie2 lDbd7 8 .i.gS
h6 9 J..h4 lDcs to l:ad1 We7 I 1 lbd2 gS 12
J..g3 Wh8 13 h4 (dlis docs not seem logical)
1 3...J..d7 1 4 hxgS lucgS I S f3 Wg7 16 Wf2
llh8 1 7 l:l.h1 lDhS 1 8 �fl 1if6 and Black
had the initiative in Kofidis-Alterm:m, Ko­
motini 1992.
7 . . . h6 8 i.h4 i.g4
8. ..g5 c:m be met by 9 �xgS!? hxgS 10
.i.xgS with the idea of meeting 10 ... J..e6 with
1 1 .i.d3 �bd7 12 f4!, after which White has
a very strong attack - the absence of Black's
dark-squared bishop is really felt.
9 h3 .i.h5 1 0 'ird3 �7 1 1 b4 'ireS
Also interesting is 1 l ...gS 1 2 J..g3 �b6 1 3
.i.b3 .i.g6 1 4 :re 1 aS when Black has good
counterplay.
1 2 'ire3 i.g6 1 3 lL\d2 lbh5
Or 1 3...c6 1 4 l:tfc1 dS 1 S exdS �xdS
(tS...cxdS 16 .i.bS a6 1 7 .i.xf6 axbS 18 .i.xeS
�xeS 1 9 1WxeS 1ixcS 20 l:txeS would leave
Black with a tedious endgame in which he is
slightly worse) 16 .i.xdS cxdS 1 7 c4 d4 1 8
1W f3 and the position i s rather unclear.
14 a4 Wtl8 1 5 a5 f5 1 6 exf5 i.xf5 1 7 a6
b6 1 8 g4 •us 1 9 i.d5 lbf4 20 .i.xa8
llxa8 21 Wh2 i.xc2

Black has escaped from the opening with
adequate play for the exchange.
22 c4 l:lf8 23 .ig3 h5 24 f3 Wh7 25
.i.xf4 llxf4 26 �3 1ih6 27 Aae1 c5 28
b5 ltlf8?1
28...lbf6!?, with good play, would have
been better.
29 ltle4 .ixe4 30 fxe4 ltle6 31 gxh5
'lrxh5 32 Wg2 'lrf7 33 1i'd3 'lrg6+ 34
'lrg3 1i'h6 35 llxf4 exf4 36 1i'g4 t£!d4 37
h4?!
37 �h2! would have kept an edge for
White.
37 . . .f3+?
Black misses his chance. After 37...lbc2! Wd7 1 1 g7 .i.xg7 12 ..g4+ Wd6! (12...Wc7?!
the game would most likely have ended in a
draw.
38 Wh3 Wh7 39 llg1 f2 40 llf1 1i'e3+
41 Wg2 l002 42 llxf2 ltle1 + 43 Wt1
ltld3 44 llf7 1i'e1 + 45 Wg2 ...d2+ 46
Wg1 •c1 + 47 Wh2 'lrb2+ 48 1tg2
•e5+ 49 1tg3 1tb2+ 50 Wh3 Wg8 51
1tf3 1 -0
,.-------------.. 6 . . .l006
GaHJe 73
Morphy-Barnes
umdon 1858
_______________. tack.
1 e4 e5 2 ltlf3 ltlf6 3 .ic4!?
A romantic gambit.
3 ...ltlxe4 4 t£Jc3 ltlxc3 5 dxc3 f61
This should lead to an advanta�;,>e for
Black. Altc�tivcly:
Third Mo ve A l ternatives for White
a) S...d6? loses to 6 lbgS .i.e6 7 .i.xc6 fxc6
8 11ff3 (Bilgucr).
b) S ....i.e7 is met strongly by 6 lbxe5 0-0 7
•n i.. f6 8 tDg-4 i..e7 9 lbh6+l? gxh6 10
.i.xh6 with a powerful attack.
c) S ...c6!? leads to equality after 6 lbxeS dS
7 Wc2 i..e6 8 .i.d3 llkl7 9 f4 lbxeS 10 ..xeS
'ild6 1 1 1Wc2 .i.c7 1 2 0-0, as in San Claudio­
Bonari, Mislata 2001 .
6 0-0
6 lbh4 g6 7 f4 is probably less dangerous
(I 0-0 transposes to 6 0-0 g6 7 lbh4). Play
continues 7 ... c6! 8 f5 dS 9 fxg6 (9 .i.b3 �f7
to c4 d4 is better for Black) 9 . ..dxc4 1 0 'ifhs
1 3 11'xg7+ lbd7 1 4 .i.h6 'iVe8! I S 0-0-0 l:r.g8
1 6 ...xh7 %lh8 is 'only' equal) 1 3 'iVxg7 ...f8
14 ...g3 (or 1 4 ..xfB+ :Xf8 1 S .i.e3 .i.c6 1 6
0-0-0+ .i.dS and Black dominates th e board)
14....i.e6 1 S .i.c3 lbd7 16 0-0-0+ Wc7 17
llhfl llg8 1 8 •rz b6 t 9 lbf5 J:ld8 and Black
has a small plus, Tribushevsky-Raetsky, cor­
respondence 1982.
Or:
a) 6...c6? is now strongly met by 7 lbxeSI
d5 8 'ifhs+ We7 9 .i.d3 with a winning at-
b) 6...g6 7 lbh4 ..e7 8 �h I c6 9 f4 dS 1 0
.i.b3 e4 i s worth thinking about - we prefer
BL1ck here.
c) 6 ...'ile7 7 lle1 d6! (l...c6?! is met by 8
lbxcS!? fxcS 9 'fibS+ g6 10 .xeS dS I 1
183
Th e Petroff Defence

1txe7+ .i.xe7 12 .i.gS 0-0 13 il..xe7 dxc4 14 1 1 ... .i.e6 1 2 11h5+ g&
.i.xf8 �fB t 5 llad1 and the endbrame looks 12 ...i.f7 13 'li'g4 g6, with a slight edge,
good for White) 8 �4 c6 9 f4 g6 1 0 P.i dS was also poNSible.
1 1 .i.d3 'ilg7 also looks better for Black.
d) 6...d6 7 l0h4 g6 8 f4 'ile7 9 f5 'flg7
should be better for Black. Now 10 .i.e3 c6
1 1 1te2 dS 12 .i.b3 gS 13 c4!? gxh4?1 14 cxdS
cS 1 5 d6! .i.xd6?! 1 6 lladt gave White tre­
mendous compensation in Crepan-Rezonja,
Bled 2000. However, stronger is 1 3...d4! 1 4
�xd4 gx.h4 when Black has a ck"al' plus.
7 � '11fe7
Or:
a) 7.../8:7?! is very weU met by 8 .i.d3 g6 9
f4, when White has a powerful initiative. One
game continued 9._.i.g7 10 fxeS fxcS 1 1 �gS
c6? 1 2 lLlf5! gxf5 1 3 1Whs+ WfB 1 4 :X f5+ 1 -0 1 3 ltlxg6?!
Saburov-Lutze, correspondence 1 906. Risky. After 13 'ile2 0-0-0 14 f4 .i.cS 1 5
b) 7...g6 8 f4 f5 9 l0f.3 e4 10 lOgs .i.cS+ 1tf2 .i.xe3 1 6 'li'xe3 White i s maybe only
1 1 Wh1 'ti'f6 12 'ti'dS (12 .i.£7+ We7 13 .i.ds slighdy worse.
d6 14 :C1 WEB 1 5 g4 lDe71? 1 6 gxf5 lLlxdS 1 3. . ..i.f7 1 4 11h4 hg& 1 5 •xf& llgB 1 6
17 1txd5 gxf5 is l.'VCO worse, Black having a %lad 1 .i.e7
clear plus in Schlcchter-Marco, Berlin 1 897) Not 1 6...l0e7? 17 l:lxdS! lLlxdS?l 1 8
12...d6 13 lLlxh7 1te7 1 4 ltlg5 lDds and we 1txe5+ .i.e7 1 9 .i.xdS when White wins be­
prefer Black even if it is not entirely clear. cause of the threat of .i.£7+.
1 7 .e6 .i.f7 1 8 11h3 �

8 �f5?!
8 lthS+ Wd8 9 lLlf5 (9 c!l\g6? 1te8 and
Black wins) 9...g6 10 l0xe7 hrxhS 1 1 lLlxc8
l:lxc8 1 2 .i.e3 is better, giving White good
compensation fi>r the pawn.
8 ......c5 9 .i.b3 d5 1 0 .i.e3 1t'a5 1 1 �h4
Black's position is also more pleasant after
1 1 l0g3 .i.e6 1 2 1th5+ g6 1 3 1th4 .i.e7.
1 8...l:td8!? 19 'ilxh7 D.d6 looks good,
White not having <.:nough for the piece. After
1 8...ltld8 the trend starts to chaf1bre.
1 9 f4 e4 20 llxd5
20 .i.xdS!? is an interesting option:
20 ....i.xd5 21 1ibS+ l0£7 22 :XdS ...xa2 23
:Cs and White has an auack.

184

20 .. �xd5 21 'lh15+
21 .•. �8?1
Much stronger is 21 . ..l:g6 22 .ixd5 .lc5
23 ..thl 'ifa6 (23 ...ixe3 24 .if7+ ltlxf7 25
.
11fxa5 leaves White slightly better) 24 l:le I
.ixe3 25 f5 ..f6 26 fxg6 hxg6 27 1i'g4 and
the game is still very unclear.
22 .ixd5 :g7
22.. c6? loses to 23 'irf5+ �g7 24 ..id4+.
.
23 b4 ...86 24 f5!
24. . .tnf7?
24...'irf6 25 ..id4 'ifg5 26 .i.xg7+ ��7 27
•es was necessary, when White's attack is
fear.;omc. but the game goes on.
25 f6! .ixf6 26 b51 1Wd6 27 .ixf7 + b6
p
27 ...l:lxn s .icS is the end.
28 .ih6 W87 29 .i.xg7 .txg7 30 ..tb3 :m
31 :n + :xn 32 1Wxf7 + �dB 33 1Wxg7
•d1 + 34 �2 1Wd2 + 35 Wg3 e3 36
irf6 + lilc8 37 i..e6 + �7 38 1Wf3 + 1 .0
Third Mo ve A lternatives for White
Game 74
Svetushkin-Miles
Alushta 1999
1 e4 tnf6 2 d3 e5 3 tnf3
Por our purposes, the Petroff move order
would have been 1 e4 e5 2 ltlf3 ltlf6 3 cB.
3 . . .tnc6
4 c3
4 g3 is also harmless: 4...d5 5 exd5 ltlxd 5 6
.ig2 .i.cS 7 0-0 0-0 8 :Ct %le8 9 h3?! (9 ltlc3
ltlxc3 I 0 bxc3 .i.g4 would transpose to the
so-called Glek Variation of the Pour Knights
Opening - the chances are equal) 9 .. .i.f5 10
.
ltlh4?! (1 0 ltlbd2 is better, although we pre­
fer Black after 10 .. f6 1 1 ltle4 i..b6) t o...i..e6
.
1 1 ltld2 ltlf4! (this surprising move leaves
Black with a substantial plus) 1 2 gxf4 9xh4
1 3 ltle4 .ixh3! (a nice sacrifice) 14 ltlxc5
i..xg2 1 5 �xg2 exf4 t6 ltle4 ltld4 17 f3 1le6
I 8 ltlf2 ..g3+ 19 �ft Dg6 0-1 Manik-Oral,
Trcncin 1 995.
4 d5 5 li:lbd2 aS 6 .i.e2 .i.c5
.•.
Black has also tried 6...g6 7 b3 (7 0-0 i..g7
8 .l:lcl 0-0 9 .ift J:leR 1 0 b3 b6 I I .la3 .la6
12 'it'c2 'ii'd7 1 3 llad t llad� looks pretty
standard and pretty level) 7...i..g7 8 .la3
il)h5 9 0-0 ltlf4 10 J:lc 1 ltlxe2+ I I 'it'xe2
.ie6 (or 1 1 ...d4!? 1 2 cxd4 exd4 13 .l:lacl ltlb4
with equality - Makarychev) 1 2 exd5 'ii'xd5
13 c4!? tlfd7 14 .i.b2 .ig4 15 d4 0-0-0!? 16
dxeS llhe8 17 'We3 £6 and Black had good
185
The Petroff Defence
compensation in Morozevich-Makarychev,
Moscow 1992.
0-0 8 a4
7 ()..()
Other moves to consider are:
a) 8 "irc2 a4!? 9 llb 1 lle8 to b4 axb3 1 1
axb3 liJh5 1 2 llet liJf4 1 3 .i.fl wiili equal
chances.
b) 8 b3 lle8 9 a3 .i.a7 (or 9...d4 1 0 .i.b2
dxc3 1 1 .i.xc3 .i.g4 with equality) to llbt
dxe4 1 1 dxe4 liJh5 1 2 c!2k4 "irf6 13 b4 axb4
1 4 axb4 tl'lf4 1 5 .i.xf4 1fxf4 1 6 b5 with un­
ck'ar play, Gclashvili-Hellsten, Korinthos
2002.
8. . .1le8
8...h6 9 "irc2 .i.c6 is also a very sensible
way to play. The following continuation
looks promising for Black: 1 0 exd5 liJxd5 1 1
lle1 1Wd7 12 .i.fl f6 1 3 lbc4 .i.a7 14 b3 .i.g4
(14 ... 1lad8 1 5 .i.a3 llfe8 is a sound alterna­
tive; we prefer Black) 1 5 .i.e2 f5 1 6 ltJg3
186
llae8 17 h3 e4!? 18 dxe4 fxe4 1 9 tl'lxe4 .i.f5
20 .i.c4 �h8 21 .i.a3 .i.xhJ! 22 .i.xf8 1Wg4
23 liJM .i.xg2 24 liJg3 .i.e4! (Lcin­
Kupreichik, Hastings 1 982/83). Now after
25 llxe4 'iVxg3+ 26 �h 1 1Wh3+ the game
would have ended with a draw.
9 exd5
White has also played 9 .c2 .i.g4 to lle1
h6 1 1 tl'ln 'iVd7 1 2 .i.e..'� .i.f8 (or 1 2...d4 1 3
cxd4 .i.xO 14 d5!? .i.xc2 I S dxc6 "irxc6 1 6
llxe2 .i.xe3 1 7 tl'lxe3 1Wd7 with level pms­
pects) 1 3 llad1 llad8 1 4 liJg3 d4 1 5 .i.cl
(after 15 cxd4 .i.xf3 16 .i.xf3 liJxd4 17 .i.xd4
"irxd4 1 8 "irxc7 .i.c5 1 9 llf1 b6 Black has
fantastic play for the pawn) 1 5....i.c5 16 liJh4
.i.e6 17 liJhf5 �h7 18 .i.f3 liJg8 1 9 llk2
.i.f8 and Black was a bit better, Vt.>sclovsky­
Scrgeev, Ceske Bude�>Vice 1997.
9 �xd5 1 0 lDe4
•.•
to tl'lc4 .i.f5 1 1 "irh3 "ird7 12 tngs llab8
also gives a fair share of chances to both
players.
1 0...i.f8
Or 1 O....i.b6!? 1 1 tl'lfd2 f5 t 2 l;jg3 .i.e6 1 3
c!ik4 .i.cS with unclear play.
1 1 lle1 h& 1 2 1ib3?!
This is the path to destruction. White was
better off playing 1 2 liJfd2 liJf4 1 3 tl'lc4
ltJxe2+ 14 'irxe2 .i.f5, though Black has easy
equality.
1 2 . ...te6! 1 3 1i'xb7?!
.
13 'iVbS was better, though after 13 ...'iVc8
14 .i.ft lla6 White is still struggling.
 Third Move A lternatives for Wh ite

29 .i.d1 tne6 30 g3 lidS

1 3...li)b6!?
Maybe even stronger is 1 3. ..d71? 14 �5
..
:Cb8 1 5 li)c5 1t'e8 1 6 li)xe6 l:txb5 17 axb5
llkb4! (the hidden point) 18 cxb4 ..txb4 with
a clear plus for Black.
1 4 1Wxc6
Forced. After 1 4 ..ie3? ..id71 1 5 ..ixb6
cxb6 followed by Jla7 Black wins the queen
..
for a rook.
1 4... -*.d5 1 5 'ttxeS
The queen is trapped, and after 1 5 'irb5 c6
it would be sold more cheaply.
1 5 1Wxe8;1'6 .i.d1 c5 1 7 c4
.••
3 1 f4?1
The only chance was 31 l:te3 f5 32 llk3
�5 although Black still has all the fun.
31 ...1lxd3
Simpler was 3t .....txc4! 32 l:txe4 l:txd3 33
fxe5 li)g5 34 :C1 Wb7 and Black wins.
32 lDf2 lidS 33 fxe5 f5! 34 .*.c1 .*.aS!
After this there is no defence.
35 lla2 1Wb7 36 Wf1 1fg2+ 37 ¢>e2
.i.f3+ 3S �3 .*.xd1 0-1

1 7 d4 also leads to a black edge after

1 7... cxd4 1 8 cxd4 ..tb4 19 ..id2 Ld2 20
�xd2 e4.
1 7 ...-*.c6 1 S lDc3 f6 1 9 b3 'tfdS 20 .*.e2
g5!? 21 lDd2 lDcS 22 .i.b2 lDe7 23 I.Ode4
ll'if5 24 lDd5 .*.e7
After the inferior 24... ..ixd5?! 25 cxd5
lDd4 26 ..txd4 cxd4 27 ..tg4 White has excel­
lent countcrplay on the light squares.
25 .*.g4 lDg7 26 .*.f3 g4 27 tl'ixe7+
1txe7 2S .*.e2 h5
It is safe to say dmt White is far from hav­
ing sufficient compensation.

187
The Petroff D e fence

Summary
In practice White's deviations from the main continuations on move three quite often trans­
pose to other openings. The least justifiable is 3 d3, since White gives away his advantage of
the first move. Those who enjoy attacking the king should study 3 .i.c4 l£lxe4 4 l£lc3 l£lxc3 5
dxc3, which is especially recommendable fi>r rapid and blitz tournaments. Unfortunately,
4...l£lf6 cools down White's ardour.
3 l£lc3 .i.b4 leads to l)uite another type of position with a slow, manoeuvring fight and
White having the advantage of two bishops (often in the endgame). Like in many other basic
variations of the Petroff Defence, Black has to neutralise White's pressure with careful play.

1 e4 e5 2 .!Df3 .!Llf6 3 l003

3 .i.c4 (D) - Game 73
3 d3 - Gm11t 74
3 .•• .i.b4 4 c!Dxe5 (D) - Game 71
4 .i.c4 (D) - Game 72

3 -*.c4 4 c!Dxe5 4 .i.c4

188
I INDEX OF COMPLETE GAMES I

Aagaard-I..egky, RHdapest 1996 ................................................................................ 1 54

Alekhine-Rabinovich.A, Moscow 1918..................................................................... 81
Anand-Gelfand, Mo.rcoll' 2004..................................................................................... 4 3
Anand-Hiibner, Dortn11111d 1992 .............................................................................. 126
Anand-Ivanchuk, UtJttrrs 1993 ................................................................................ 1 16
Anand-Mishra, Itulia 1988 ........................................................................................ 1 52
Anand-Shirov, U11arr.r 2000 ........................................................................................ 66
Anand-Shirov, 1\tfoscoliJ'-200 1 ....................................................................................... 12
Anand-Sokolov.1, J)orlnllmd 1999 ............................................................................. 24
Areshchenko-Mista, Ctrppelle Ia Gra11de 200 3 .......................................................... 88
Aronian-Akopian, Ohrid 2001 .................................................................................... 86
Bonch-Osmolovsky - Baranov, Moscon' 1954 ..................................................... 1 73
Burkov-Raetsky, Con"Cspo11dmce 1985........................................................................ 4 7
Cabrera-Collas, Malaga 2003.................................................................................... 169
Dolrnatov-Mamedyarov, MoscoJI' 2002.................................................................. 1 36
Dolrnatov-Raetsky, Podolsk VoroneifJ 1992............................................................. 92
Elizarov-Raetsky, Relorrcbmsk 1989........................................................................ 128
Firman-Bick, .flmllo/1 M01mtaifl 2003 ........................................................................ 59
Grischuk-Adarns, Halkidikj 2002............................................................................ 1 1
Grischuk-Gelfand, Wyk atlfl 'Lee 2002 ..................................................................... 52
Grischuk-Pavasovic, lstaii!JIII 2003 ........................................................................... 23
Gurevich.V-Meijers, GerffiCI'!)' 1999 ...... . ................ . . . . .�
... .. . ... .......... . . .
... .. .............. . . 62
.

Hamdouchi-l.e Roux, Be!Jort 2003 .......................................................................... 60

Herrera- Perez.Ro, f/aradero 2000............................................................................ 72
Howell.J-Makarychev, Fnmze 1989....................................................................... 120
lvanchuk-Shirov, DorlnJIIfld 1998 .............................................................................. 4 1

1 89
The Pet roff Defence

Kasparov-Anand, I i11ares 2000 .................................................................................. 21

Kasparov-Ivanchuk, Debrece11 1992 ........................................................................ 122
Kasparov-Karpov, Moscm11 1981 ................................................................................. 69
Kasparov-Shirov, Wyk aa11 Zee 2001 ........................................................................ 50
Khalifman-Atalik, Ha/kidiki 2002............................................................................. 83
Kholmov-Belousov, Gorf?y 1974 ............................................................................. 101
Kotronias-Motylev, Moscow 2004 ............................................................................. 18
Kovalev-Rozentalis, G/o._�OJI' 200 / .. . . . . . . . . . . . 57
. . . . . .. . . . ........ ...... .... ...... ...... ... ... ... ... .... .... .....

Kupreichik-Yusupov, Mi11sk 1987 . . . . . .. . .. . ...... ..... .

..... . . . . ... 74
.... ....... ......... ............. .. .. .. .

Lasker.Em-Marshall, 51 Petersbu'l!, 1912.................................................................. 85

Lastin-Najer, Elisltt 2000 124
................. . . . . . . . . . . ............................. . . . . . . ............................

Leko-Grischuk, Wijk aa11 Zee 2002 .......................................................................... 14

Leko-Kramnik, World Ch. (GatHe 1), Brissago 2004 .................................................. 8
Leko-Yusupov, JJortflll/1/d 1998.................................................................................. 39
Lev-Alterman, H.atnai-Gan 1992............................................................................... 181
Macieja-Rozentalis, Re11o 1999 ............................................................................... 1 12
Morphy-Barnes, Londo11 1858 .................................................................................. 183
Naiditsch-Timman, Dortt111111d 2002 ........................................................................ 90
Nevednichy-Collas, Montpellier 2003...................................................................... 161
Novozhilov-Raetsky, Comsponde��ce 1983 ............................................................... 98
Nyysti-Rychagov.M, Gothenburg 2003 ..................................................................... 26
Pavlovic-Raetsky, Bie/ 1999 ....................................................................................... 63
Psakhis-Cooper, Port Erin 2003 ................................................................................ 38
Rublevsky-Vallejo, Ohrid 2001 ................................................................................ 107
Sax-Yusupov, Tbessaloniki Ofympiad 1984 ................................................................. 37
Shirov-Gelfand, Aslalla 200 1 ...................................................................................... 28
Shirov-Gelfand, Leon 2001 ........................................................................................ 1 7
Shirov-Kramnik, Belgrade 1997................................................................................. 1 14
Shirov-Piket, Wijk aa11 Zee 2001 ................................................................................. 54
Shirov-Yusupov, Fiuropea11 Tetllll Ch., Bat11mi 1999 ................................................. 35
Shirov-Yusupov, Ter Ape/ 1997 ............................................................................... 147
Smagin-Makarychev, Moscmv 1987 .......................................................................... 94
Srnirin-Alterman, Haija 1995................................................................................... 164
Sorokin-Raetsky, K.ra.r11odar 1984 ............................................................................ 13 3
Stefansson-Yusupov, Eupe11 1994 .......................................................................... 172
Svetushkin-Miles, Alushta 1999 .............................................................................. 185
Svidler-Akopian, IPorld Tet1111 Ch., l_J1ceme 1997 ................................................... 179
Svidler-Ponomariov, Mosco/1' 2001 .......................................................................... 105
Svidler-Yusupov, Bad HonJbHIJ!. 1998 ...................................................................... 109

190
 Index o f Complete Games

Tal-Kholrnov,Abna Ata 1968 1 75

..........................................•.......................................

Timofeev-Raetsky, Correspondence 1982 ................................................................. 139

Timoscenko-Yusupov, F"rnn� 1979...................................................................... 150
Tiviakov-Forintos, Porto San Giorgio 1994 ............................................................. 158
Tiviakov-Miles, Unaros 1998 ................................................................................... 131
Topalov-Anand, Wijk aan Zee 2003 .......................................................................... 30
Tukmakov-Bronstein, Mosmw 1971 ........................................................................ 76
Van Der Wiel-MeUado, E/goibar 1998 ................................................................... 14 3
Vitolinsh-Anikaev, Riga 1982 96
....................................................................................

19 1