6 3.

11

Arguing as in (3.43), we get the corresponding iteration equation

Pn = pPn+ + qPn; Pn = Pn+ + qPn;

and proceed as in Example 3.15. 3.12 Suppose one best on k = 1 2 Then

6. 5 6 5 6
2

1 p1 = P (k appears on one dice) = 3 1 6 1 2 p2 = P (k appear on two dice) = 3 2 6 5 p0 = P (k appear none) = 6 Thus, we get
3

1 p3 = P (k appear on all the tree dice) = 6

Net gain = 2p1 + 3p2 + 4p3 p0 = 0:343:

;

5 which gives for i = a

8 > > > < Na = > > > : 8 > > > < => > > :

)a a a + b 1 (q=pa p q 1 (q=p) +b p q p = q ab p=q )b p = q: a + b 1 (p=qa b 2p 1 2p 1 1 (p=q) +b ab p=q

; ; ; ; ; 6 ; ; ; ; ; 6

4 3.10 Using the hint, we obtain

p (Nk+1 Nk ) = q (Nk Nk;1) 1

; ; ;

Let

Mk+1 = Nk+1 Nk
;

so that the above iteration gives

Mk+1 = q p Mk 1 p 8 n o k q q 1 > k > M 1 ( ) p=q > p 1 p q p < => > M1 k p=q > : p
; ; ; ; 6 ;

This gives

Ni =

i;1 X k=0

Mk+1 M1 + p 1 q
;

8 > > > > > < => > > > > :

i;1 X k=0

q p
;

i p q p=q
; 6

iM1 i(i2p 1)
;

p=q

where we have used No = 0. Similarly Na+b = 0 gives

a + b 1 q=p : M1 + p 1 q = p q 1 (q=p)a+b
; ; ; ;

Thus

8 > > > < Ni = > > > :

i a + b 1 (q=pa)+ i p q 1 (q=p) b p q p = q i(a + b i) p=q

; ; ; ; ; 6 ;

3 3.8 De ne the events A=\ r successes in n Bernoulli trials" B =\success at the ith Bernoulli trial" C =\r 1 successes in the remaining n 1 Bernoulli trials excluding the ith trial" P (A) = n r pr qn;r P (B ) = p 1 P (C ) = n r 1 pr;1 qn;r We need ) = P (BC ) = P (B ) P (C ) = r : P (B A) = PP(AB (A) P (A) P (A) n
; ; ; ; j

3.9 There are 52 13 ways of selecting 13 cards out of 52 cards. The number of ways to select 13 cards of any suit (out of 13 cards) equals 13 = 1. Four such (mutually exclusive) suits give the total number 13 of favorable outcomes to be 4. Thus the desired probability is given by 4 ! = 6:3 10;12 52 13

2 3.7 (a) Let n represent the number of wins required in 50 games so that the net gain or loss does not exceed $1. This gives the net gain to be 1 < n 50 4 n < 1 16 < n < 17:3 n = 17
; ; ;

1 17 3 33 = 0:432 P (net gain does not exceed $1) = 50 17 4 4 P (net gain or loss exceeds $1) = 1 0:432 = 0:568
;

(b) Let n represent the number of wins required so that the net gain or loss does not exceed $5. This gives 5 < n (50 2 n) < 5 13:3 < n < 20
; ; ;

X 50 1 P (net gain does not exceed $5) = 19 n = 14 n 4 P (net gain or loss exceeds $5) = 1 0:349 = 0:651
;

3 4

50;n

= 0:349

Problem Solutions for Chapter 3

3.1 (a) P (A occurs atleast twice in n trials) = 1 P (A never occurs in n trials) P (A occurs once in n trials) = 1 (1 p)n np(1 p)n;1
; ; ; ; ; ;

(b) P (A occurs atleast thrice in n trials) = 1 P (A never occurs in n trials) P (A occurs once in n trials) P (A occurs twice in n trials) = 1 (1 p)n np(1 p)n;1 n(n2 1) p2(1 p)n;2
; ; ; ; ; ; ; ; ; ;

3.2

1= 1 P (doublesix) = 1 6 6 36

= 1 50 0 = 0:162
;

P (\double six atleast three times in n trials00)

1 36
0

35 36

50

50 1

1 36

35 36

49

50 2

1 36

35 36

48

3.6 (a) (b) 1 (c) 1

; ;

p1 = 1
5 6 18 1
!

5 6 = 0:665 6
!

12
;

12 1 5 6

1 6
;

5 6 18 2
!

11

= 0:619
2

5 6

18
;

1 6

17

1 6

5 6

16

= 0:597