An Agent-Based Model of Behavior in “Beauty Contest” Games

Recently, computer simulation, particularly agent-based modeling, has grown in popularity as a method to uncover macro patterns and developments that emerge from simple micro behavior. The present paper combines both techniques by using protocol analysis to uncover player strategies in an experiment and encoding those strategies in an agent-based computer simulation. In particular, Keynes’ (1936) beauty contest analogy is simulated in a number-guessing context. Several researchers have conducted experiments asking subjects to play “p-beauty contest games” in order to compare the experimental results with those predicted by the game-theoretic, deductive reasoning concept of iterated dominance. Our results are compared with those found experimentally in order to demonstrate the usefulness of a combining agent-based modeling with protocol analysis.