# How to compute conditional expectation in Markov Chain?

There are $$d$$ spheres numbered $$1,2,3,...,d$$ distributed in two marked boxes. Let $$X_0$$ be the number of spheres in box $$1$$ at time $$0$$ and at each time $$n= 1,2,...,d$$ number is randomly selected from the set $$\{1,2,...,d\}$$ and the sphere marked with that number is taken out of the box it is in and placed in the other box. We denote by $$X_n$$ the number of spheres in box $$1$$ at time $$n$$; prove that $$\{X_n,n≥0\}$$ is a Markov chain with state space $$S=\{1,2,...,d\}$$ and determine the transition probabilities, finally compute $$\mathbb{E}(X_{n+1}|X_n)$$.
I've tried this:
The transition matrix is $$\mathbb{P_{ij}}=\left\lbrace\begin{array}{c} 1-\frac{i}{d}~~~~~~if~j=i+1 \\ \frac{i}{d}~~~~~~~~~~~~if~j=i-1\\0~~~~~~~~~~~otherwise \end{array}\right.$$

• saying "number is randomly selected from the set $\{1,2,...\}$" doesn't make sense Mar 8 at 21:27
• I already corrected that part Mar 9 at 18:15

I would use the following:

$$\mathbb{E}​[X_{n+1}​|X_n=i]=\sum_{k=0}^d k\cdot \mathbb{P}(X_{n+1}​=k|X_n=i)​=(i+1)(1-\frac{i}{d})+(i-1)\frac{i}{d}=1+i(1-\frac{2}{d}),$$ Where in the second inequality I used the transition probabilities you derived.

Thus I think you can conclude $$\mathbb{E}​[X_{n+1}​|X_n]=1+X_n(1-\frac{2}{d})$$.