Number of trials required to estimate a free-energy difference, using fluctuation relations.