Bridging Level-K to Nash Equilibrium

TL;DR

This paper introduces NLK, a model bridging Nash equilibrium and Level-K reasoning, which better explains experimental game data by incorporating psychological insights into players' beliefs about opponents' sophistication.

Contribution

NLK is a novel model that unifies NE and Level-K reasoning, providing improved predictions for various game experiments with fewer parameters.

Findings

NLK outperforms traditional NE and Level-K models in explaining experimental data.

A simple NLK version fits data better with fewer parameters.

Extensions to multi-player and heterogeneous belief games are discussed.

Abstract

We introduce NLK, a model that connects the Nash equilibrium (NE) and Level-K. It allows a player in a game to believe that her opponent may be either less or as sophisticated as, she is, a view supported in psychology. We apply NLK to data from five published papers on static, dynamic, and auction games. NLK provides different predictions than those of the NE and Level-K; moreover, a simple version of NLK explains the experimental data better in many cases, with the same or lower number of parameters. We discuss extensions to games with more than two players and heterogeneous beliefs.