We study the problem of estimating the probability P(X<Y)P(X<Y) when two random variables XX and YY are observed with additional errors. We derive the convergence rate for a proposed estimator of P(X<Y)P(X<Y) in the case of compactly supported error densities.