Is state-dependent valuation more adaptive than simpler rules?

Abstract

McNamara, Trimmer, and Houston (2012) claim to provide an explanation of certain systematic deviations from rational behavior using a mechanism that could arise through natural selection. We provide an arguably much simpler mechanism in terms of computational limitations, that performs better in the environment described by McNamara, Trimmer, and Houston (2012). To argue convincingly that animals' use of state-dependent valuation is adaptive and is likely to be selected for by natural selection, one must argue that, in some sense, it is a better approach than the simple strategies that we propose.