1
$\begingroup$

Let $X_1, X_2, X_3, \dots$ be a sequence of random variables that converges almost surely $$(X_n) \rightarrow X$$ to a number $X \in \mathbb R$ (or more precisely the delta dirac distribution centered at some real number). I am interested in the convergence of the average sequence $(Y_n)$ where $$Y_n = \frac 1 n \sum_{i=1}^n X_i$$ Can we make any statement about the convergence of $(Y_n)$? Intuitively, I would guess it also has to converge to $X$, but I am not sure about the mode and how to prove it.

$\endgroup$
0
$\begingroup$

For numerical sequences $(x_n)$, it is a standard exercise that $\lim x_n=L$ implies $\lim \frac{1}{n}\sum_{k=1}^n x_k=L$.

By assumption, there is a set $E$ of full measure such that for every $\omega\in E$, $X_n(\omega)\to X(\omega)$. By the above, this implies $Y_n(\omega)\to X(\omega)$.

Thus, $Y_n \to X$ almost surely. (And consequently, converges in probability and in distribution.) This is pretty much it; since there is no reason for $X$ or $Y_n$ to be integrable, we cannot expect, say, convergence in $L^1$.

$\endgroup$

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.