We roll a six-symmetric dice until we get a roll of 6. What is the average number of throws (including the final roll of six) assuming all rolls are even numbers?
I do
$P_n\left(A|B\right)=\left(\frac{2}{6}\right)^{n-1}\cdot\frac{1}{6}=\frac{1}{2\cdot3^n}$,
$EN=\sum_{n=1}^\infty{\frac{n}{2\cdot3^n}}=\frac{1}{2}\cdot\frac{3}{4}=\frac{3}{8}$.
The answer given to this exercise is 1.5.
What's wrong?
Of course, the probability $P_n\left(A|B\right)$ above is not a right conditional probability. It should be like this:
$A$ — the event of receiving a sequence of twos and fours ending in a six,
$A_n$ — the event of receiving a sequence of twos and fours ending in a six of length $n$;
then
$\mathbb EN=\sum_{n=1}^\infty n\mathbb P\left(A_n\vert A\right)=\sum_{n=1}^\infty n\frac{\mathbb P\left(A_n\cap A\right)}{\mathbb P\left(A\right)}$,
$\mathbb P\left(A\right)=\sum_{n=1}^\infty \left(\frac{1}{6}\right)^{n-1}\cdot 2^{n-1}\cdot \frac{1}{6}=\frac{1}{4}$,
$\mathbb P\left(A_n\cap A\right)=\left(\frac{1}{6}\right)^{n-1}\cdot 2^{n-1}\cdot \frac{1}{6}=\frac{1}{2\cdot 3^n}$,
$\mathbb EN=1.5$.
Thank you.