An important problem faced by boundedly rational agents is to identify 'regions of rationality,' i.e., the areas for which simple, boundedly rational models are and are not effective. To map the contours of such regions, we derive probabilities that models identify the best of m alternatives (m ...