Fig 1. Convergence on optimal choice behavior.

The performance of matching and maximization as a function of the number of evaluations of the choice options. The assessment was performed over 9000 choice situations, involving up to 10 options, each with distinct contingencies between reward and effort (Materials and Methods). The maximum gain (the value of 100%) was established by allowing maximization to converge. Matching (green) provides high gain following just one sample (one evaluation; see abscissa) of the worth of the choice options. A specific example of how matching achieves high performance following a single evaluation of the choice options is shown in Figure 3B. Here and henceforth, plots show mean±s.e.m. values. Here and elsewhere, the error bars are often smaller than the symbols indicating the mean.