2
$\begingroup$

If $Y_n$ is limited in probability and $X_n \to 0$ in probability then $X_nY_n \to 0$ in probability.

Attempt: I would go like:

for every $\epsilon$, $\mathbb P(|X_nY_n|\leq \epsilon) = \mathbb P(|Y_n|\leq \frac{ \epsilon}{ |X_n|})\geq\mathbb P(|Y_n|\leq C_{\delta}) \geq 1-\delta \to 1$ but the truth is that $\frac{ \epsilon}{ |X_n|}\geq C_{\delta}$ is not always true but almost certainly, so this strategy should be wrong!

Thanks!

$\endgroup$
2
$\begingroup$

For any fixed positive $\epsilon$:

$\mathbb{P}[|X_nY_n| \ge \epsilon] \leq$

$\mathbb{P}[|X_n| \ge \epsilon^2]$ $+$ $\mathbb{P}[|Y_n| \ge \frac{1}{\epsilon}]$

But (for any such $\epsilon$) the above 2 terms on the rhs of the above inequality go to 0 as $n$ goes to $\infty$.

$\endgroup$
  • $\begingroup$ Why is the first inequality correct? Imo $\mathbb P(|X_n| \ge \epsilon^2,|Y_n| \ge \frac{1}{\epsilon}) \leq \mathbb P(|X_nY_n| \ge \epsilon)$ and not the other way around. $\endgroup$ – Maffred Jun 28 '18 at 20:29
  • $\begingroup$ The only way $|X_nY_n|$ can be at least $\epsilon$ is if either $|X_n|$ is at least $\epsilon^2$ OR $|Y_n|$ is at least $\frac{1}{\epsilon}$. If both inequalities $|X_n| < \epsilon^2$ and $|Y_n| \le \frac{1}{\epsilon}$ hold then the inequality $|X_nY_n| < \epsilon$ also holds. $\endgroup$ – Mike Jun 28 '18 at 20:35
  • $\begingroup$ So, as you wrote above, $\mathbb{P}[|X_n| \le \epsilon^2$ AND $|Y_n| \le \frac{1}{\epsilon}]$ $\le$ $\mathbb{P}[|X_nY_n| \le \epsilon]$. But we also have $\mathbb{P}[|X_n| \le \epsilon^2$ OR $|Y_n| \le \frac{1}{\epsilon}]$ goes to 0, and this is at least $\mathbb{P}[|X_nY_n| \le \epsilon]$. $\endgroup$ – Mike Jun 29 '18 at 3:18

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.