Explaining the flaws in human random generation as local sampling with momentum

In many tasks, human behavior is far noisier than is optimal. Yet when asked to behave randomly, people are typically too predictable. We argue that these apparently contrasting observations have the same origin: the operation of a general-purpose local sampling algorithm for probabilistic inference. This account makes distinctive predictions regarding random sequence generation, not predicted by previous accounts—which suggests that randomness is produced by inhibition of habitual behavior, striving for unpredictability. We verify these predictions in two experiments: people show the same deviations from randomness when randomly generating from non-uniform or recently-learned distributions. In addition, our data show a novel signature behavior, that people’s sequences have too few changes of trajectory, which argues against the specific local sampling algorithms that have been proposed in past work with other tasks. Using computational modeling, we show that local sampling where direction is maintained across trials best explains our data, which suggests it may be used in other tasks too. While local sampling has previously explained why people are unpredictable in standard cognitive tasks, here it also explains why human random sequences are not unpredictable enough.Author summary: When explicitly asked to be random, people are not random enough. Previous accounts of these random generation tasks have argued that people are effortfully trying not to be predictable. In many other tasks, however, people also show random behavior, even when it is unnecessary or outright disadvantageous. Here, we try to bridge this apparent gap. We hypothesize that the randomness people produce when trying to be random and the randomness that they display when trying to make the best choice has the same common mechanism: drawing mental samples to make judgments and decisions. In two experiments, we compare previous random generation accounts, which are task-specific in nature, to the more general account of mental sampling that has been used to explain how people behave in many other domains. We find that the flexibility of human random generation in our data is better explained by the mental sampling account. We also find a novel empirical signature of momentum in random generation, which points to a new kind of mental sampling algorithm. If mental sampling governs behavior in random generation tasks and elsewhere, then this task has great promise in helping to understand wider human behavior.