A simple proof for the continuity of infinite convolutions of binary random variables (Q1116522)

Let \(X_ j\), \(j\geq 1\), be independent random variables taking the values \(a_ j\geq b_ j\) with distribution \(p_ j=P(x_ j=a_ j)=1-P(X_ j=b_ j)\). Assume that the infinite series \(U=\Sigma X_ j\) is (a.s.) convergent. We give a simple proof for Lévy's theorem on the continuity of the distribution function F(x) of U.