2
$\begingroup$

Given a standard Brownian motion $\{B_t;0 \leq t < \infty \}$, we know that $\lim_{t \to \infty}\frac{B_t}{t} = 0$ a.s. I am interested to know if we can prove Strong Law of Large Numbers for any finite second moment IID variables using the above result. One of the ways I was looking at it is to use Skorokhod embedding. If $X_1,X_2,X_3,...$ are IID random variables with zero mean and unit variance we know that we can find stopping times $T_1, T_2, ...$ such that $B_{T_1}, B_{T_2} - B_{T_1},... $ are IID with same distribution as $X_i$. If we can prove that $T_n \to \infty$ then using this we can establish SLLN. Can some one help me with proving $T_n \to \infty$. Thanks

$\endgroup$
1
$\begingroup$

$(T_{k+1}-T_k)_{k=1,2,\cdots}$ is an IID sequence. Let $X=T_2-T_1$. Since $\mathbb{P}[X=0]<1$, there is a positive number $\epsilon$ such that $\mathbb{P}[X>\epsilon]>0$. So, by independence and the second Borel-Cantelli lemma, $\mathbb{P}[T_{k+1}-T_k >\epsilon \text{ infinitely often}] =1$. If $T_{k+1}-T_k >\epsilon$ infinitely often, then $T_n \to \infty$ as $n\to\infty$. Thus $T_n\to\infty$, almost surely.

$\endgroup$

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.