Let $(X_k)$ a sequence i.i.d. of random variables such that $\mathbb E[|X_1|]<\infty $ and let fix $N\in\mathbb N$. We set, \begin{cases}Z_n=X_1+...+X_n\\ Y_n=\frac{1}{n}Z_n\\ R_n=Z_{N-n}\\ S_n=Y_{N-n}\end{cases} with $n\in\{0,...,N-1\}$. Show that $(S_n)$ is a martingale refer to $(R_n)$.

I have to prove that $$\mathbb E[S_{n+1}\mid R_0,...,R_n]=S_n.$$ I have that $$\mathbb E[S_{n+1}\mid R_0,...,R_n]=\mathbb E[Y_{N-n}\mid Z_N,...,Z_{N-n}]=\frac{1}{N-n-1}\mathbb E[Z_{N-n-1}\mid Z_N,...,Z_{N-n}]\underset{m=N-n}{=}\frac{1}{m-1}\mathbb E[Z_{m-1}\mid Z_m,...,Z_N]=\frac{1}{m-1}\sum_{i=1}^{m-1}\mathbb E[X_i\mid Z_m,X_{m+1},...,X_N].$$

This is what I don't understand:

Why do we get $\mathbb E[X_i\mid Z_m,X_{m+1},...,X_N]$ ? to me we should get $\mathbb E[X_i\mid X_1,...,X_m,...,X_{N}]$ (because $Z_m$ depend of $X_1,...,X_m$ and all $Z_i$ depend of $(X_1,...,X_i)$).

Finally, like $(X_n)$ is a sequence of i.i.d. shouldn't we have that $$\mathbb E[X_i\mid X_1,...,X_m,...,X_{N}]=E[X_i]$$

and if we really have $$\mathbb E[X_i\mid Z_m,X_{m+1},...,X_{N}]$$ that $$\mathbb E[X_i\mid Z_m,X_{m+1},...,X_{N}]=\mathbb E[X_i\mid Z_m]\ \ ?$$


Why do we get $\mathbb E[X_i\mid Z_m,X_{m+1},...,X_N]$ ?

This is due to the fact that the $\sigma$-algebra generated by the random variables $Z_m,Z_{m +1},\dots,Z_N$ is the same as the $\sigma$-algebra generated by $Z_m,X_{m+1},\dots,X_N$. To see this, note that $(Z_m,X_{m+1},\dots,X_N)=L(Z_m,Z_{m +1},\dots,Z_N)$ where $L$ is a bijective bi-continuous function given by $(x_m,\dots,x_N)\mapsto (x_m,x_{m+1}-x_m,\dots,x_N-x_{N-1})$.

The next step is to show that for each $i\in\{1,\dots,m-1\}$, $$\mathbb E\left[X_i\mid Z_m,X_{m +1},\dots,X_N\right]=\frac 1mZ_m.$$


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.