# Skorokhod's representation theorem

Let ${\displaystyle \mu _{n}}$, ${\displaystyle n\in \mathbb {N} }$ be a sequence of probability measures on a metric space ${\displaystyle S}$ such that ${\displaystyle \mu _{n}}$ converges weakly to some probability measure ${\displaystyle \mu _{\infty }}$ on ${\displaystyle S}$ as ${\displaystyle n\to \infty }$. Suppose also that the support of ${\displaystyle \mu _{\infty }}$ is separable. Then there exist random variables ${\displaystyle X_{n}}$ defined on a common probability space ${\displaystyle (\Omega ,{\mathcal {F}},\mathbf {P} )}$ such that the law of ${\displaystyle X_{n}}$ is ${\displaystyle \mu _{n}}$ for all ${\displaystyle n}$ (including ${\displaystyle n=\infty }$) and such that ${\displaystyle X_{n}}$ converges to ${\displaystyle X_{\infty }}$, ${\displaystyle \mathbf {P} }$-almost surely.