Probability Matching, Fast and Slow

A prominent point of contention among researchers regarding the interpretation of probability-matching behavior is whether it represents a cognitively sophisticated, adaptive response to the inherent uncertainty of the tasks or settings in which it is observed, or whether instead it represents a fundamental shortcoming in the heuristics that support and guide human decision making. Put crudely, researchers disagree on whether probability matching is "smart" or "dumb." Here, we consider evidence for both "smart" and "dumb" variants of probability-matching behavior, as well as its alternative, maximizing. We rely on the influential and often-cited distinction between two "systems" of thinking to organize the research and competing interpretations of probability-matching behavior as "smart" or "dumb."