这是数理统计中的数学问题，关系几何型随机变量的分布问题，希望大神们多多帮忙哇~~~~

我也用英文写好了- -
let Y be a bernoulli distributed variable such that P(Y=1)=p (success with prob p). Let X be a geometric distributed variable such that P(X=k)=(1-p)^(k-1)*p, this can be interprated as the probability that for a sequence of independent bernoulli distributed varialbes, the first k-1 are unsuccessful and the k th is successful.

Therefore P(X>k) stands for the probability that all of the first k bernoulli variables are 0 (unsuccessful).
P(X>n+k-1|X>n-1)
=P{the first n+k-1 Bernoulli variables are 0 give the first n-1 of them are 0}
(b/c of independence)
=P{all of the n th to the n+k-1 th variables are 0}
(there are k Bernoulli variables in this range)
=P{the first k bernoulli variables are 0)
=P(X>k)

不知道满不满意 有问题再问我吧