$X_1, X_2, \ldots \sim \operatorname{Ber}(p)$, those variables are i.i.d.

Suppose $S_n = X_1 + X_2 + \cdots + X_n$

$$N = \min\{ n \geq 1: S_n = 1\}$$

1) Show that $X_{N+1}, X_{N+2}$ also follow i.i.d $\operatorname{Ber}(p)$;

2) show that $\{X_1, X_2, \ldots , X_N\}$ and $\{X_{N+1}\}$ are independent each other.

I'm a bit of a dummy so a thorough explanation would be much appreciated. Thanks in advance!

  • $\begingroup$ Probably the reason for the votes to close this question is that it's phrased like a homework problem without expressing any thoughts about it beyond the statement of the problem. $\endgroup$ – Michael Hardy Apr 16 '18 at 17:08

$\newcommand{\e}{\operatorname{E}}$ The law of total probability says the conditional probability of an event $A,$ given a random variable $N,$ is a random variable whose expected value is the marginal (or "unconditional") probability of $A{:}$ $$ \Pr(A) = \e(\Pr(A \mid N)). $$ We have \begin{align} & \Pr(X_{N+1} = x_1\ \&\ \cdots\ \&\ X_{N+k}= x_k \mid N=n) \\[10pt] = {} & \Pr(X_{n+1} = x_1\ \&\ \cdots\ \&\ X_{n+k} = x_k) \\[10pt] = {} & p^{x_1+\cdots+x_k} (1-p)^{k - (x_1+\cdots+x_k)} \tag 1 \\[10pt] & \text{But that last expression does not depend} \\ & \text{on the value of $n$, so we conclude that} \\[10pt] & \Pr(X_{N+1} = x_1\ \&\ \cdots\ \&\ X_{N+k}= x_k \mid N=n) \\[10pt] = {} & \Pr(X_{N+1} = x_1\ \&\ \cdots\ \&\ X_{N+k}= x_k) \tag 2 \end{align} Since the expressions on lines $(1)$ and $(2)$ are equal to each other, regardless of the value of $k$, we have $$ X_{N+1},X_{N+2},X_{N+3},\ldots\sim\operatorname{i.i.d.} \operatorname{Bernoulli}(p). $$ Maybe I'll be back to address the second question later.

  • 1
    $\begingroup$ Thank you so much Michael. I really appreciate it. This helps me understand the problem now. Thanks! $\endgroup$ – Olympia111 Apr 16 '18 at 16:09
  • $\begingroup$ @Olympia111 If this answer meets your needs then it is good thing to accept it. It always feels good for the one who answered. $\endgroup$ – drhab Apr 16 '18 at 17:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.