Lahore University of Management Sciences

CMPE 501: Applied Probability (Fall 2010) Homework 2: Solution

1. The Chess Problem: This years Belmont chess champion is to be selected by the following procedure. Bo and Ci, the leading challengers, rst play a two-game match. If one of them wins both games, he gets to play a two-game second round with Al, the current champion. Al retains his championship unless a second round is required and the challenger beats Al in both games. If Al wins the initial game of the second round, no more games are played. Further, we know the following: The probability that Bo will beat Ci in any particular game is 0.6. The probability that Al will beat Bo in any particular game is 0.5. The probability that Al will beat Ci in any particular game is 0.7. Assume no tie games are possible and all games are independent. Solution: For this question we dene the dierent possible outcomes by the sequence of the winners of the dierent games. So e.g. BBA means that the rst two games are won by Bo and the third game (i.e. rst game of round two) is won by Al. (a) Determine the a priori probabilities that i. the second round will be required. Solution: P(Rnd 2 will be required) = P(BB) + P(CC) = (0.6)(0.6) + (0.4)(0.4) = 0.52 ii. Bo will win the rst round. Solution: P(Bo will win Rnd 1) = P(BB) = (0.6)(0.6) = 0.36 iii. Al will retain his championship this year. Solution: P(Al champ) = 1 P(Bo champ) P(Ci champ) = 1 P(BBBB) P(CCCC) = 1 (0.6)2 (0.5)2 (0.4)2 (0.3)2 = 0.8956 Page 1

Alternately you can use the total probability theorem. P(Al champ) = P(Rnd 2 not required) P(Al champ | Rnd 2 not required) + P(Bo wins Rnd 1) P(Al champ | Bo wins Rnd 1) + P(Ci wins Rnd 1) P(Al champ | Ci wins Rnd 1) = (1 0.52)(1) + (0.36)(1 0.52 ) + (0.16)(1 0.32 ) = 0.8956 (b) Given that the second round is required, determine the conditional probabilities that i. Bo is the surviving challenger Solution: P(BB) P(Rnd 2 required) 0.36 = 0.52 9 = 13

P(Bo wins round 1 | Rnd 2 required) =

ii. Al retains his championship Solution: P(BBA) + P(BBBA) + P(CCA) + P(CCCA) P(Rnd 2 required) 2 (0.5 + 0.52 ) + 0.42 (0.7 + 0.3 0.7) 0.6 = 0.52 = 0.7992

P(Al champ | Rnd 2 required) =

Alternately you can use the conditional version of the total probability theorem (see Question 2 below). P(Al champ | Rnd 2) = P(Bo challenger | Rnd 2) P(Al champ | Rnd 2, Bo challenger) + P(Ci challenger | Rnd 2) P(Al champ | Rnd 2, Ci challenger) = 0.6923 (1 0.52 ) + 0.3077 (1 0.32 ) = 0.7992 (c) Given that the second round was required and that it comprised only one game, what is the conditional probability that it was Bo who won the rst round? Solution: P(BBA) P(BBA) + P(CCA) 0.62 (0.5) = 0.62 (0.5) + 0.42 (0.7) = 0.6164

P(Bo wins Rnd 1 | Rnd 2, one game) =

Page 2

2. (Chapter 1 Problem 19): Conditional version of the total probability theorem. Show the identity P(A | B) = P(C | B) P(A | B C) + P(C c | B) P(A | B C c ) assuming all the conditioning events have positive probability. Solution: See the text book. 3. Hira, a LUMS freshman, makes one to ve new friends every week, with equal probability. The number of friends she makes each week is independent from all other weeks. We are concerned with two consecutive weeks. Let event A be Hira made a total of 10 friends during the two weeks. Let event B be Hira made more than 5 friends during the two weeks. (a) Are events A and B independent? Solution: Since A B, therefore P(A B) = P(A). This is equal to P(A)P(B) only if P(A) = 0 or P(B) = 1. In this case, clearly P(A) = 0 and P(B) < 1. Therefore P(A B) = P(A) P(B) and A and B are not independent. (b) Let C be the event Hira made exactly 5 friends during the rst week. Are events A and B independent, conditioned on C? Solution: Given event C, event A can only occur if Hira made exactly 5 friends in the second week and event B will always happen no matter what happens in the second week. So P(A | C) = and P(B | C) = 1. Moreover 1 = P(A | C) P(B | C) . 5 Therefore conditioned on C, events A and B are independent. P(A B | C) = 1 5

(c) Is A independent of C? Solution: 1 From part (b) above we have P(A | C) = . 5 Also 1 1 1 = P(A | C) P(A) = = 5 5 25 and therefore A and C are not independent.

Page 3

(d) Is B independent of C? Solution: From part (b) above we have P(B | C) = 1. Also P(B) < 1 meaning P(B) = P(B | C) and therefore B and C are not independent. (e) (Optional) Given that Hira made a total of 6 friends in two weeks, what is the probability that she made exactly 2 friends in the rst week? How about 3 friends in the rst week? Solution: We dene the following events. Fi as the event that i friends were made in the rst week (i = 1, 2, , 5). Si as the event that i friends were made in the second week (i = 1, 2, , 5). Tj as the event that a total of j friends were made in the rst two weeks (j = 2, 3, , 10). Now P(F2 | T6 ) = = = = = = P(F2 T6 ) P(T6 ) P(F2 S4 ) 5 i=1 P(Fi S6i ) P(F2 ) P(S4 ) 5 i=1 (P(Fi ) P(S6i )) 1/5 1/5 5 i=1 (1/5 1/5) 1/25 1/5 1 5

where step 2 follows from the total probability theorem and step 3 follows from the independence of Fi and Sj . Similarly P(F3 | T6 ) can be calculated to be 1/5. (f) (Optional) Given that Hira made a total of 6 friends in two weeks, what is the probability that she made exactly 2 friends in at least one the weeks? How about exactly 3 friends in at least one the weeks? Solution: P(F2 S2 | T6 ) = P(F2 | T6 ) + P(S2 | T6 ) P(F2 S2 | T6 ) 1 1 = + 0 5 5 2 = 5 where step 1 follows from the fact that conditional probability denes a valid probability law. P(F2 | T6 ) was calculated to be 1/5 in part (b) above and since the weeks are identically distributed, P(S2 | T6 ) is also 1/5. Page 4

Also P(F3 S3 | T6 ) = P(F3 | T6 ) + P(S3 | T6 ) P(F3 S3 | T6 ) 1 1 1 = + 5 5 5 1 = 5 The two results make sense as there is only one way of making 6 friends in two weeks by making exactly 3 friends in at least one of the two weeks. Compared to this, there are two ways of making 6 friends in two weeks by making exactly 2 friends in at least one week. 4. (Chapter 1 Problem 35): Let A and B be independent events. Use the denition of independence to prove the following: (a) The events A and B c are independent. (b) The events Ac and B c are independent. Solution: See the text book. 5. Matching Socks: Ijaz, a grad student, has a basket that contains 5 blue socks, 4 red socks and 2 green socks. One day he is getting late for the CMPE 501 class and in his hurry, picks two socks at random from the basket. What is the probability that both socks are of the same color? Solution: There are a total of 11 ways we can select 2 socks from a total of 11 socks. Moreover there 2 are a total of 5 ways we can select two blue socks, 4 ways we can select two red socks and 2 2 2 ways we can select two green socks. So 2 P(Matching socks) = = = # of ways we can select two socks of same color # of was we can select any two socks 2 4 5 2 + 2 + 2
11 2

10 + 6 + 1 17 = 55 55

Alternately, you can consider a sequential experiment (using the obvious notation): P(Matching socks) = P(BB) + P(RR) + P(GG) 5 4 4 3 2 1 = + + 11 10 11 10 11 10 17 = 55

Page 5

6. The Hat Problem: The hats of n persons are thrown into a box. The persons then pick up their hats at random (i.e. every assignment of the hats to the persons is equally likely). What is the probability that exactly n 2 persons pick their own hat? Solution: Exactly (n 2) persons picking their own hats means that the other 2 persons picked each others hats. The number of ways n persons can pick n hats is n!. The number of ways we can choose 2 persons to pick each others hats is n . Therefore 2 P(Exactly n 2 persons pick their own hats) =
n 2

n!

1 2 (n 2)!

7. (Chapter 1 Problem 50): Correcting the number of permutations for indistinguishable objects. When permuting n objects, some of which are indistinguishable, dierent permutations may lead to indistinguishable object sequences, so the number of distinguishable object sequences is less than n!. For example, there are six permutations of the letters A, B and C: ABC, ACB, BAC, BCA, CAB, CBA, but only three distinguishable sequences that can be formed using the letters A, D and D: ADD, DAD, DDA. (a) Suppose that k out of the n objects are indistinguishable. Show that the number of distinguishable object sequences is n!/k!. (b) Suppose that we have r types of indistinguishable objects, and for each i, ki objects of type i. Show that the number of distinguishable object sequences is n! k1 !k2 ! kr ! Solution: See the text book.

Page 6

Optional Questions
1. Non-transitive Dice: Consider four die labeled A, B, C and D. These are normal dice, in that each has six sides and the probability any given die lands on any particular side is 1/6 (regardless of the outcome of the other tosses for that die or any of the tosses for all the other dice). They are unusual only in that the numbers on their six sides are dierent from standard dice. The numbers for the four dice are: A: 0,0,4,4,4,4 B: 3,3,3,3,3,3 C: 2,2,2,2,6,6 D: 1,1,1,5,5,5 In a single game, you choose a die and toss it and I choose a dierent die and toss it. The highest number wins. (For these dice, a tie is not possible). (a) Find the probability that i. die A beats die B. Solution: A beats B when A lands on a 4. So P(A beats B) = ii. die B beats die C. Solution: B beats C when C lands on a 2. So P(B beats C) = 2 3 2 3

iii. die C beats die D. Solution: C beats D when either C lands on a 6 or when C lands on a 2 and D lands on a 1. Using the total probability theorem, the required probability is given by P(C beats D) = 1 + 3 2 3 1 2 = 2 3

iv. die D beats die A. Solution: Again using the total probability theorem gives the required probability as P(D beats A) = 1 + 2 1 2 1 3 = 2 3

(b) Is there anything unusual about your answers above? If so, explain. Solution: The results seem a little un-intuitive at rst. This is because if A is more likely to beat B (than the opposite) and B is more likely to beat C (than the opposite) and so on, then one would expect it more likely that A would beat D (than the opposite). However this is not the case because of the way the game has been constructed and the probabilities calculated are correct.

Page 7

(c) Suppose I get to pick my die rst, and you then get to choose any one of the remaining three. Show that by making a clever choice, you will beat me in the game with a probability of 2/3, regardless of which die I choose! Which die would you choose if I choose die A? What if I choose die B, or die C or die D? Solution: Pick the die that is one alphabet before the one already chosen. So if B has been selected then select A, if C has been selected then select B, if D has been selected then select C and if A has been selected then select D. 2. Alice plays with Bob the following game. First Alice randomly chooses 4 cards out of a 52-card deck, memorizes them, and places them back into the deck. Then Bob randomly chooses 8 cards out of the same deck. Alice wins if Bobs cards include all cards selected by her. What is the probability of this happening? Solution: The total number of ways Bob can choose 8 cards from a deck of 52 cards is 52 . The number 8 of ways Bob can select 4 cards from Alices four and 4 cards from the remaining 48 cards is
4 4 48 4

. So the required probability is

4 4

48 4 52 8

3. Conditional Independence & Independence of a Collection of Events: Please have a look at the worked examples 1.18, 1.19, 1.20 and 1.21 from the course text book (Bertsekas & Tsitsiklis). Solution: See the text book.

Page 8