Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

This is a homework problem. The problem is proving that $\lim_{n\to\infty}P\,(X_n=X) = 1$ when the sequence of random variables $X_n$ converges to $X$ almost surely.

I think the problem is essentially about the order of the probability sign and limit sign. I know that the order cannot be interchanged in general. However, if $X_n$ converges a.s. and each $X_n$ is dominated by an integrable random variable $Y$, in the light of the dominated convergence theorem, $$ \begin{eqnarray} 1 = P\,(\lim_{x\to\infty}X_n=X) &=& \int_{\{\omega\,:\,\lim_{n\to\infty}X_n(\omega)=X(\omega)\}} d\mu \\ &=& \int\chi_{\{\omega\,:\,\lim_{n\to\infty}X_n(\omega)=X(\omega)\}} d\mu \\ &=& \int\lim_{n\to\infty}\chi_{\{\omega\,:\,X_n(\omega)=X(\omega)\}} d\mu \\ &=& \lim_{n\to\infty}\int\chi_{\{\omega\,:\,X_n(\omega)=X(\omega)\}} d\mu \\ &=& \lim_{n\to\infty}P\,(X_n=X) \end{eqnarray} $$

I this derivation right? But, even if it is right, I have no idea how to prove it without the dominance condition. How can I prove it generally? Is it possible actually?

share|improve this question
    
what is your definition of almost sure convergence –  K.Ghosh Dec 25 '12 at 9:07
    
@K.Ghosh, it is the usual definition. $P\,(\lim_{n \to \infty} X_n = X) = 1$. (en.wikipedia.org/wiki/Convergence_of_random_variables) –  JDL Dec 25 '12 at 9:10
1  
This is not true. Take $X_{n}=\frac{1}{n}$ for all $n\in\mathbb{N}$. Then $X_{n}\to 0=:X$‚ but $P(X_{n}=X)=P(\frac{1}{n}=0)=0$ for all $n\in\mathbb{N}$. Hence $\lim_{n\to\infty}P(X_{n}=X)=0$. –  Thomas E. Dec 25 '12 at 9:11
add comment

1 Answer

up vote 1 down vote accepted

The claim is not true. Consider the probability space $([0,1],\mathcal{B}[0,1],\lambda|_{[0,1]})$ (where $\mathcal{B}$ denotes the borel-$\sigma$-algebra, $\lambda|_{[0,1]}$ the lebesgue measure on $[0,1]$) and

$$X_n(\omega) := \frac{1}{n} \qquad (\omega \in [0,1])$$

Then $X_n \to 0=:X$ almost surely, but $$\mathbb{P}[X_n=X] = \mathbb{P}[X_n=0]=0$$

thus clearly $\mathbb{P}[X_n = X] \not \to 1$ as $n \to \infty$.

share|improve this answer
    
Yes, I think I have misundersood the problem. Anyway, it seems that my derivation with the dominated convergence theorem is also not true. Your example is a counter example for the derivation. Could you point out which line of the equation is wrong? –  JDL Dec 25 '12 at 9:32
    
Yes. $\chi_{\{\omega; \lim_n X_n(\omega)=X_n(\omega)\}} \not= \lim \chi_{\{\omega; X_n(\omega)=X(\omega)\}}$... The example above shows that even if $X_n(\omega) \to X(\omega)$ there does not necessarily exist $N \in \mathbb{N}$ such that $X_N(\omega)=X(\omega)$. –  saz Dec 25 '12 at 9:37
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.