De Finetti's Theorem

Definition (Exchangeable)

Let (Yi)i1(Y_{i})_{i\ge{1}} be a sequence of Random Variables, we say (Yi)(Y_{i}) is exchangeable if every finite dimensional distribution of (Yi)(Y_{i}) is the same as the corresponding finite dimensional distribution of (Yπ(i))i1(Y_{\pi(i)})_{i\ge 1} where (π(i))i1(\pi(i))_{i\ge 1} is any permutation of 1,2,1,2,\dots. i.e. One can reorder the RVs in any way and it won’t change the distribution.

Theorem (De Finetti)

Any sequence of random variables Y1,Y2,Y_{1},Y_{2},\dots is if and only if there exists some quantity θ\theta, s.t. conditional on θ\theta, Y1,Y2,Y_{1},Y_{2},\dots are Independent and identically distributed.