We discuss worst-case robustness in the context of macroeconomic policy, finance and process systems engineering. This requires a minimax framework that evaluates the best design, or decision, in view of the worst-case scenario(s). The scenarios can be discrete, such as a set of (rival) models purporting to represent the same system, or continuous, such as uncertain values varying between upper and lower bounds. We consider discrete and continuous minimax formulations and discuss approaches for computing the solution.