The present paper reviews various learning models focusing on the properties of the associated sequence of play whether play converge over time and if it approaches the behavior predicted by traditional equilibrium analysis.