For measuring the weight of evidence for an alternative H1 to a hypothesis H0 or the degree of safety with which H0 can be rejected, this paper suggests an estimator of the indicator function of H1, with two separate squared error loss functions, and a bound on the risk under H0.