Understanding a precondition for the law of large numbers

54 Views Asked by At

Discussions of the law of large numbers frequently begin like this:

Let $X$ be a real-valued random variable, and let $X_1$, $X_2$, $\ldots$ be an infinite sequence of i.i.d. copies of $X$. Let $\bar{X}_n := \frac{1}{n}(X_1 + \ldots + X_n)$ be the empirical averages of this sequence.

But this is confusing to me; in particular, this is because I cannot imagine a scenario where (given the constraints above) it is not the case that

$$ X = \frac{1}{n}(X_1 + \ldots + X_n) $$

But it must be possible for the left-hand side of this equation to diverge from the right-hand side (otherwise the law of large numbers wouldn't be interesting).

So does there exist a simple example where

$$ X \ne \frac{1}{n}(X_1 + \ldots + X_n)? $$