International Association
for Cryptologic Research

In this very short note we prove that the pseudo-random index j of RC4 is indeed not pseudo-random. This is a simple result that missed our attention for quite a long time. We show that in long term Pr(j = i+1) = 1/N - 1/N^2, instead of the random association 1/N and this happens for the non-existence of the condition S[i] = 1 and j = i+1 that is mandatory for the non-existence of the Finney cycle.