Math 5010-1, Spring 2005 Assignment 4 Problems #1, p. 171. Write O, W, B respectively for “orange,” “white,” and “black.” The possible values of X are: 0, 1, ±2, 4. The respective probabilities are as follows: 8 2 14 , 2 P {X = −2} = P {2 W’s} = P {X = −1} = P {1 W and 1 O} = 8 1 2 14 2 1 , 2 2 14 , 2 P {X = 0} = P {2 O’s} = P {X = 1} = P {1 B and 1 W} = P {X = 2} = P {1 B and 1 O} = 8 1 4 1 14 , 2 4 2 1 1 14 , 2 4 2 14 . 2 P {X = 4} = P {2 B’s} = #7, p. 171. (a) (b) (c) (d) 1, 2, 3, 4, 5, 6. 1, 2, 3, 4, 5, 6. 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12. 0, ±1, ±2, ±3, ±4, ±5. #14, p. 172. Let Nk denote the number that Player k draws. It follows that P {X = 0} = P {N1 < N2 }. By relabeling the Players you find that this is the same as P {N2 < N1 }. But P {N1 < N2 }+P {N2 < N1 } = 1 because one of the two events must occur. Therefore, P {X = 0} = 12 . Similiarly, P {X = 1} = P {N2 < N1 < N3 }. Relabel the players to find that this is the same as the probability of any given win order for three players. Because there are 3! 1 such relabelings, we have shown that P {X = 1} = 3! = 16 . If you have successfully this far then you probably have noticed that in general P {X = i} = i!1 for i = 0, . . . , 5. #17, p. 172. The pmf p satisfies: p(x) = 0 unless F jumps at x. In that case, p(x) = F (x)−F (x−). 1 (a) The pmf is: p(x) = 1 4 16 if x = 1, 1 12 0 if x = 3, if x = 2, otherwise. 1 For instance, p(3) = F (3) − F (3−) = 1 − 11 12 = 12 . Note also that p(0) = F (0) − F (0−) = 0 because F is continuous at zero. (b) P { 21 < X < 32 } is the probability that X = 1. This is p(1) = 14 . #18, p. 173. The possible values of X −2 are: 0, ±1, ±2. Moreover, X is a binomial with parameters n = 4 and p = 12 . Therefore, p(x) = 4 0 2−4 = 1 16 if x = −2, 4 1 2−4 = 1 4 if x = −1, 4 2 2−4 = 3 8 if x = 0, 4 3 2−4 = 1 4 if x = 1, 4 4 2−4 = 1 16 if x = 2. For instance, p(−2) = P {X − 2 = −2} = P {X = 0} = 4 0 ( 21 )0 (1 − 12 )4−0 = 1 16 . Theoretical Exercises #1, p. 180. As in Example 1e (p. 125), we compute P {T > n} first. Let Ai be as in that example, and note that N [ P {T > n} = P ! Ai i=1 = N X P (Ai ) − i=1 + XX P (Ai ∩ Aj ) + · · · 1≤i<j≤N X ··· X (−1)k+1 P (Ai1 ∩ · · · ∩ Aik ) + · · · + 1≤i1 <···<ik ≤N + (−1)N +1 P (A1 ∩ · · · ∩ AN ). 2 So far, this is the same as before. But now note that P (Ai ) = (1 − Pi )n , P (Ai ∩ Aj ) = (1 − Pi − Pj )n , etc. Plug to find that N [ P {T > n} = P ! Ai i=1 = N X i=1 + XX n (1 − Pi ) − n (1 − Pi − Pj ) + · · · 1≤i<j≤N X ··· X n (−1)k+1 (1 − Pi1 − · · · − Pik ) + · · · + 1≤i1 <···<ik ≤N + (−1)N +1 0 = N −1 X X ··· X n (−1)k+1 (1 − Pi1 − · · · − Pik ) . k=1 1≤i1 <···<ik ≤N Therefore, P {T = n} = P {T > n − 1} − P {T > n} = N −1 X X ··· X n (−1)k+1 (1 − Pi1 − · · · − Pik ) k=1 1≤i1 <···<ik ≤N − N −1 X X ··· X n−1 (−1)k+1 (1 − Pi1 − · · · − Pik ) . k=1 1≤i1 <···<ik ≤N But (1 − α)n−1 − (1 − α)n = (1 − α)n−1 α. Apply this with α := Pi1 + · · · + Pik to find that !n−1 k k N −1 X X X X X k+1 Pi` Piν . ··· (−1) 1− P {T = n} = k=1 1≤i1 <···<ik ≤N `=1 ν=1 #2, p. 180. Note that if b ≤ c then {X ≤ b} ⊆ {X ≤ c}. That is, as b decreases to −∞, the events {X ≤ b} decrease to ∩−∞<b<∞ {X ≤ b} = ∩∞ n=−∞ {X ≤ n} (why?). The point is that the latter intersection is taken over the integers, which is a countable set. Therefore, by the continuity properties of P , limb→−∞ P {X ≤ b} = P (∩∞ n=−∞ {X ≤ n}). But this is the probability that X = −∞ (why?). The latter probability is zero because X is a bona fide random variable. 3