When evaluating public programs with long-term effects, future benefits (or costs) must be discounted to their present value. Modern finance theory recommends pursuing projects with a positive expected Net Present Value (NPV) and rejecting those without. However, a project with a positive expected NPV can still result in an undesirable ex post outcome, creating public regret and reducing long-term societal well-being. In this paper, I introduces a regret-minimization framework for evaluating sustainable policies, incorporating Savage’s min-max regret criterion and the expected NPV approach as limiting cases. I demonstrate that regret-risk aversion and the skewness of a project's contingent valuation significantly influence project evaluation. Additionally, I provide an analytical solution under Savage’s min-max regret criterion, which holds strong normative appeal in the face of deep uncertainty.