$S_n$ is a symmetric random walk. How do I show that the event $A_M=[\limsup\limits_{n\to\infty}S_n\leqslant M]$ is in the tail sigma algebra $\cap_{m>n}\sigma(X_m,X_{m+1},...)$? I would be really grateful for any help.


For a sequence $(S_n)_{n \in \mathbb{N}}$ of random variables given by $$S_n := \sum_{j=1}^n X_j,$$ it does (in general) not hold that

$$A_M := \left\{ \limsup_{n \to \infty} S_n \leq M \right\}$$

is a tail event with respect to $(X_n)_{n \in \mathbb{N}}$ - instead, it is a tail event with respect to $(S_n)_{n \in \mathbb{N}}$. Since the $\sigma$-algebras $\sigma(S_n)$, $n \in \mathbb{N}$, are (in general) not independent, this means in particular that Kolmogorov's 0-1 law is not applicable.

Remark: It is not difficult to see that $A_M$ is an element of the exchangable $\sigma$-algebra and therefore

$$\mathbb{P}(A_M) \in \{0,1\}$$

by Hewitt-Savage's 0-1-law.


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.