Heterogeneity in Lowest Unique Integer Game

Abstract

This study investigates a Lowest Unique Integer Game (LUIG) by agent-based simulation. LUIG is a game in which each player submit a positive integer, and the player who has submitted the unique least integer wins. We consider mixtures of two types of players, naive adaptive learner and more sophisticated quasi fictitious play learner, and investigate how each type of player behaves and learns. Our results show that more sophisticated quasi fictitious play learner often lose against naive adaptive learner. LUIG with small number of players seems belong to one of those situations in which it is better to be naive than to be sophisticated.