A procedure is presented for calculating the costs of failure of a dam and for incorporating these risk costs into the evaluation process. Basically, it involves a calculation of expected value. One critical issue is what the probability of failure should be; given that research will not provide accurate values for some time, a default value (of 10-4 per dam year) is proposed to be used until better information becomes available. Another critical question is whether the refinement of incorporating risk costs is worthwhile; here, the risk-benefit factor α is introduced to be used as a screening tool to decide the question. The importance of implicit value ascribed to lives lost through failure is discussed.