Difference of random variable and random sample

29 Views Asked by At

Let $X$ be a real random variable and $\{X_n\}_n$ be a random sample of $X$. My question is whether or not $$ \cfrac{1}{n}\sum_{i=1}^n \text{E}[|X-X_i|]\rightarrow 0 \text{ a.e.} $$

It is a result of the type of Strong Law of Large Numbers, but with the sum in other form. I'm stuck in this point.

Thank you in advance.

1

There are 1 best solutions below

0
On

The claim is false. Take $\{X\}\cup(X_n)$ i.i.d. $N(0,1)$. If your claim is true then it would also hold with $X$ replaced by $-X$. This would imply that $ \cfrac{1}{n}\sum_{i=1}^n \text{E}[|X-(-X)|]\rightarrow 0 \text{ a.e.} $ by triangle inequality, but this is clearly false.

More simply, $E|X-X_i|$ is independent of $i$!