Rationality and optimality are the guiding concepts of the probabilistic approach to cognition, but they are not the only reasonable guiding concepts. Two concepts from the other end of the spectrum, simplicity and frugality, have also inspired models of cognition. These fast and frugal models are justified by their psychological plausibility and adaptedness to natural environments. For example, the real world provides only scarce information, the real world forces us to rush when gathering and processing information, and the real world does not cut itself up into variables whose errors are conveniently independently normally distributed, as many optimal models assume.

However, optimal models already address these constraints. There are many methods for dealing with missing information. Optimal models can also be extended to take into account the cost of acquiring information. Finally, variables with unusual distributions can be transformed into nearly normal distributions, and outliers can be excluded. So what's the big deal? Optimal models seem to have met the challenge of adapting to natural environments. If people do not already use these models, then they would want to learn how to use them because they are, after all, optimal.

Thus it would seem that there is no need to turn to fast and frugal heuristics, which appear doomed to be both simplistic and inaccurate. Besides, there is an even stronger reason to shun simplicity and frugality as the basis for human cognition.