# Central Limit Theorem
CLT states that the distribution of $\bar X_n$ after standardization approaches
a ==standard Normal distribution==.
> [!tip] Central Limit Theorem
>
> As $n\to\infty$,
>
> $
> \sqrt n \left(\frac{\bar X_n - \mu}\sigma\right) \to \mathcal N(0, 1)
> $
>
> in distribution.
Regardless the distribution of $X_j$, _averaging_ these r.v.s will always lead
to _normality_.
Consider the sum of the i.i.d. r.v.s, we also have
$
W_n \sim \mathcal N(n\mu, n\sigma^2)
$
Which is useful for approximating sum of i.i.d. r.v.s.
- Poisson: $Y \sim \mathcal N(n, n)$
- Binomial: $Y \sim \mathcal N(np, np(1 - p))$
Cauchy distribution has no true mean and variance, and CLT and [[lln|LLN]]
doesn't hold.