# I.i.d. Geometric Random Variables

Let $X_{1},X_{2},X_{3},\ldots$ be i.i.d. geometric random variables, where $\mathbf{Pr}(X_{i} = k) = p(1 - p)^{k - 1}$.

Define $\displaystyle Z_{r} \stackrel{\text{def}}{=} \sum_{i=1}^{r} X_{i}$.

What is $\displaystyle \lim_{k \to \infty} \sum_{r=1}^{\infty} \mathbf{Pr}(Z_{r} = k)$?

Also, define $\displaystyle F(k) \stackrel{\text{def}}{=} \min_{r} (Z_{r} ~|~ Z_{r} \geq k)$.

What is $\displaystyle \lim_{k \to \infty} \mathbf{Pr}(F(k)-k=5)$?

Thanks.

-
Maybe it would help if you can give some context or your thoughts on the given problem – Belgi Feb 25 '13 at 12:19
@Solver: Are you OK with my answer? – Bravo Feb 27 '13 at 16:16

Part 1: $\{Z_r=k\}$ is the event of the $r^{th}$ success arriving in the $k^{th}$ trial. Thus the distribution of $Z_r$ is as follows.
$$P(Z_r=k)={{k-1} \choose {r-1}} p^{r-1} (1-p)^{k-r} p,\ k=r,r+1,\ldots\\ \sum_{r=1}^{\infty}P(Z_r=k)=\sum_{r=1}^{k}P(Z_r=k)=p\sum_{r=1}^{k}{{k-1} \choose {r-1}} p^{r-1} (1-p)^{k-r}=p\\ \lim_{k\to \infty}\sum_{r=1}^{\infty}P(Z_r=k)=p$$
Part 2: $$\lim_{k\to \infty}P(\min_r (Z_r|Z_r\ge k)-k=5) \\=\lim_{k\to \infty}P(X_1+X_2+\ldots+X_{r^*}=k+5|X_1+X_2+\ldots+X_{r^*-1}\le k)\\ \lim_{k\to \infty}\frac{P(X_1+X_2+\ldots+X_{r^*}=k+5,X_1+X_2+\ldots+X_{r^*-1}\le k)}{P(X_1+X_2+\ldots+X_{r^*-1}\le k)}\\ \lim_{k\to \infty}\frac{P(k+5-X_{r^*}\le k)}{P(X_1+X_2+\ldots+X_{r^*-1}\le k)}\\ \lim_{k\to \infty}\frac{P(X_{r^*}\ge 5)}{P(X_1+X_2+\ldots+X_{r^*-1}\le k)}\\ P(X_{r^*}\ge 5)=1-q^4$$