Let $\left( X_n \right)$ be a sequence of independent random variables on the measure space $(\Omega, \xi,\mathbb{P})$ with $$ \mathbb{P} \left( X_n=1 \right)= p_n \text{ and } \ \mathbb{P} \left( X_n=0 \right)= 1- p_n. $$ I have proved that $$ X_n \to 0 \text{ in probability}\Leftrightarrow p_n \to 0 , \text{ as } \ n\to \infty $$ and $$ X_n \to 0 \text{ in the $L^p$ meaning of convergence} \Leftrightarrow p_n \to 0 , \text{ as } n \to \infty$$ My problem is to show the following implication: $$ X_n \to 0 \text{ almost surely} \Rightarrow \sum_{n=1}^{\infty} p_n < \infty. $$ I have some problem to make out what is the set $$ \{ w\in \Omega \ s.t. \ X_n(w) \nrightarrow 0 \} $$

Thanks in advance.

Notation: I'm using the notation and the definitions about convergence as in the book Probability by A.N. Shiryaev


By the Borel Cantelli lemma, if $\sum_{n=1}^{\infty}p_n = \infty$ then

$$P(X_n = 1 \ \ \textrm{infinitely often}) = 1.$$

But clearly

$$\{X_n = 1 \ \ \textrm{infinitely often}\} \subseteq \{X_n \nrightarrow 0\}$$

So the latter set also has probability $1$. Hence $P(X_n \rightarrow 0) = 0$.

I have shown the contrapositive statement to your required result; the result is therefore proved.


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.