Random variables $(X_{n})$ are independent and $X_{n}$ has an uniform distribution on $[0,n]$ for n=1,2,... Prove that:

$$\frac{X_{1}+X_{2}+\dots+X_{n}}{n}\rightarrow \infty$$ almost sure.

We can write $X_{n}=nU_{n}$ where $U_{n}$ are iid and have an uniform distribution on [0,1].

Then from LLN we have:

$$\overline{U_{n}} \rightarrow \text{E}U_{1}=\frac{1}{2}$$.

Is it a good approach? Can I use this result to solve the main problem?



  1. By the strong law of large numbers, $$\frac{X_1+ \frac{X_2}{2}+\ldots+ \frac{X_n}{n}}{n} \to \frac{1}{2}$$ almost surely. Show that this implies $$\frac{\frac{X_k}{k}+ \ldots+ \frac{X_n}{n}}{n} \to \frac{1}{2}$$ almost surely for any (fixed) $k \in \mathbb{N}$.
  2. Using the the non-negativity of the random variables, show that $$\frac{1}{k} \frac{X_1+\ldots+X_n}{n} \geq \frac{1}{k} \frac{X_k+\ldots+X_n}{n} \geq \frac{\frac{X_k}{k}+ \ldots+ \frac{X_n}{n}}{n}.$$
  3. Combining step 1 and 2 conclude that $$\liminf_{n \to \infty} \frac{X_1+\ldots+X_n}{n} \geq \frac{k}{2}.$$
  • $\begingroup$ $$\liminf_{n \to \infty} \frac{1}{k}\frac{X_{1}+X_{2}+\dots+X_{n}}{n} \ge \liminf_{n \to \infty} \frac{\frac{X_k}{k}+ \ldots+ \frac{X_n}{n}}{n}= \frac{1}{2}$$ $\endgroup$ – mrnobody Jun 13 '15 at 20:14
  • $\begingroup$ thanks a lot @saz ;) Your hints are really helpful.. $\endgroup$ – mrnobody Jun 13 '15 at 20:15
  • $\begingroup$ @sebuss007 Thanks, you are welcome. $\endgroup$ – saz Jun 14 '15 at 5:27

Does this hint help: $$var(\bar{X})=\frac{1}{n^2}\sum_{j=1}^{n}\frac{1}{12}j^2$$ hence $var(\bar{X})\to \infty$ and $n \to \infty$.

  • $\begingroup$ Sorry, I don't know how I can use this hint... $\endgroup$ – mrnobody Jun 13 '15 at 19:08

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.