I am trying to solve this problem but my answer is coming out to be 1.5 where the answer is 3. I even ran a simulation to verify it and its correct. I am not really sure whats wrong in my method and how can I correct it.
I considered for getting my first 6 on Nth trials, without any 5's till now, I have probability of $$P(A^N) = \left(\frac{4}{6}\right)^{N-1} \times \frac{1}{6}$$
Now to get the expected number of rolls,
$$\sum_{N=1}^{\infty} N \times \left(\frac{4}{6}\right)^{N-1} \times \frac{1}{6}$$
So, If we bring 1/6 outside, and consider x = 4/6 then the above equation is the 1st derivative of x/(1-x)
$$\frac{d}{dx} \left( \frac{x}{1 - x} \right) = \frac{1}{{(1 - x)}^2}$$
Putting back the values and multiplying by 1/6 we get 1.5 (9/6). Where am I going wrong?
Let $N$ be the number of rolls until you stop, and $\mathcal C$ be the condition.
6(which is not a5, of course).5nor a6.5and the first6being on rolln. IE: $\mathsf P(N{\,=\,}n\cap\mathcal C)$6and we never roll a5before stopping.6when given that we never rolled a5before stopping.$$\begin{align}\mathsf E(N\mid\mathcal C) &= \dfrac{\mathsf E(N\mathbf 1_\mathcal C)}{\mathsf P(\mathcal C)} \\[1ex]&= \dfrac{\sum\limits_{n=1}^\infty n (4/6)^{n-1}(1/6)}{1/2}\\[1ex] &= 3\end{align}$$