Let $X_n$ be a sequence of non-negative iid random variables.

Is it true that the condition,

$$\limsup_{n\rightarrow\infty} \frac{X_n}{n} = \infty \text{ almost surely}$$

is equivalent to the condition,

$$\mathbb{P} \Bigg( \limsup_{n\rightarrow\infty} \Big\{ {\frac{X_n}{n} \geq x} \Big\} \Bigg) = 1 \ \text{ for all } \ x > 0$$

I am wondering because I would like to rephrase the initial condition so as to make use of the Borel Cantelli lemma.


EDIT (thanks, Learner):

For Borel-Cantelli, note that $\omega\in \{\text{limsup}_n \frac{X_n}{n}=\infty\}$ iff for every $k> 0$ there is $n\geq 1$ such that $\frac{X_n(\omega)}{n}>k$. So, if $A_k:= \{\frac{X_n}{n}>k\}$, then $\omega \in \cap_{k=1}^\infty \cup_{n=1}^\infty A_k = \text{limsup}_n A_n$ . Thus your initial condition is equivalent to $P(\text{limsup}_n A_n)=1$.


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.