This paper proposes an aspiration-based model of bounded rationality, which is applied to a large-population context where agents are re-matched every period to play a Prisoner's Dilemma.