Concepts like rational expectations, workable competition, bounded rationa­lity and the second best are often used informally to suggest that imperfectly competitive market outcomes normally are very good and that it is sometimesappropriate to restrict competition further to achieve even better results. Paul Samuelson’s optimal neoclassical analysis adds plausibility to the inference, but the presumption is profoundly misleading because real markets are bounded rational and corrupt. In Herbert Simon’s satisficing framework expanded to include vicious behavior, second bests are more plausibly interpreted as «second worsts» and comparative merit must be determined with Abram Bergson’s social welfare functions rather than by appeal to the Pareto ideal.