0

I recently came across this question and reckon it should be a direct application of Doob's inequality (correct me if I am wrong). But I struggle to write formal proof.

For each $n \in \mathbb{N}$, let $(X^{(n)}_{t})_{t\geq0}$ be a martingale on $(\Omega,F,P)$ with respect to F. Assume that for all $n \in \mathbb{N}$ we have that $\mathbb{E}[|X^{(n)}_1|^2]=0$ < $\infty$ and assume further that $lim_{n \rightarrow \infty} \mathbb{E}[|X^{(n)}_1|^2]=0$. Show that for any $\epsilon > 0 $, we have $$ \lim_{x\to\infty} \mathbb{P}(\sup_{t \in [0,1]}|X^{(n)}_t| > \epsilon) =0$$

I really appreciate your help :)

Cathy
  • 11
  • 2
  • 1
    Welcome to math.SE: since you are new, I wanted to let you know a few things about the site. In order to get the best possible answers, it is helpful if you say in what context you encountered the problem, and what your thoughts on it are; this will prevent people from telling you things you already know, and help them give their answers at the right level. – Jose Avilez May 23 '22 at 02:15

1 Answers1

0

I guess in the limit you mean $n \to \infty$ not $x$. Moreover, your assumptions are redundants : if $\forall n \in \mathbb{N}, \hspace{0.3cm} \mathbb{E}(|X_1^{(n)}|^2) = 0$ this implies the limit is also 0.

Let us assume that it is n in the limit and just $\lim_{n\to\infty} \mathbb{E}(|X_1^{(n)}|^2) = 0 $. Using Doob’s maximal inequality :

$$\mathbb{P}( \hspace{0.1cm} \underset{t \in [0,1]}{\sup} \hspace{0.1cm} |X_t^{(n)}| > \epsilon) \le \frac{\mathbb{E}(|X_1^{(n)}|^2)}{\epsilon^2}$$

And the result is straightforward by the assumption