Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

We've got probability space $(\Omega, F,P)$ and Bernoulli scheme with $n$ trials and with success probability equal to $p$. $A_k$ means exactly $k$ successes in $n$ trials. Prove that for any $B \in F$ and for any $k$, $P(B | A_k)$ doesn't depend on $p$. So I thought of writing it this way: $P(B | A_k) = \frac{P(A_k \cap B)}{P(A_k)}$ and $P(A_k) =$$ {n}\choose{k}$$p^k q^{n-k}$, but what next?

share|improve this question

2 Answers 2

up vote 1 down vote accepted

Call $X=(X_i)_{1\leqslant i\leqslant n}$ the $n$ results and $S_n=X_1+\cdots+X_n$, then $A_k=[S_n=k]$. It seems that, by $B\in F$, you mean that $B$ is measurable with respect to $X$, that is, that there exists some $C\subseteq\{0,1\}^n$ such that $B=[X\in C]$.

Call $C_k=\{c\in C\mid c_1+\cdots+c_n=k\}$, then $B\cap A_k=[X\in C_k]$ and $\mathbb P(X=c)=p^k(1-p)^{n-k}$ for every $c$ in $C_k$ hence $\mathbb P(B\cap A_k)=|C_k|\cdot p^k(1-p)^{n-k}$.

Likewise, $\mathbb P(A_k)=\displaystyle{n\choose k}\cdot p^k(1-p)^{n-k}$ hence $\mathbb P(B\mid A_k)=\dfrac{|C_k|}{{n\choose k}}$, which does not depend on $p$.

share|improve this answer

This property is often referred to as the memorylessness property in probability and statistics. Let's rewrite $B$ in the same fashion as we have written $A_k$. If $A_k$ means exactly k successes in n trials, we can say $B_j$ means exactly $j$ successes in $m$ trials, where $j\geq k$, $m\geq n$. What we want to show is that:

$P(B_j |A_k)=C_{j-k}$, where $C_{j-k}$ means exactly $j-k$ successes in $m-n$ trials, using

$P(B_j | A_k) = \frac{P(A_k \cap B_j)}{P(A_k)}$, as you stated, where

$P(A_k) =$$ {n}\choose{k}$$p^k q^{n-k}$

$P(B_j) =$$ {m}\choose{j}$$p^k q^{n-k}$ and

$P(C_{j-k}) =$$ {m-n}\choose{j-k}$$p^{j-k}q^{m-n-(j-k)}$.

The next step is to apply our set theory definitions along with properties of $ {n}\choose{r}$ in order to reduce $P(B_j | A_k) = \frac{P(A_k \cap B_j)}{P(A_k)}$ to the form of $P(C_{j-k})$.

share|improve this answer
    
Why B means j successes in m trials? I mean, we have probability space for Bernoulli scheme for n trials, then why do we take event for m trials? Is $B \in F$? –  Anne Mar 10 '13 at 11:51

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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