For an irreducible positive recurrent Markov chain with stationary distributionπ, the expected return time to state i is 1/πi. More generally, expected occupation rewards in a return cycle equal their stationary rate times the expected cycle length.