Yu-Hsiang (Sean) Wang

California Institute of Technology · Social Sciences · 2026–27 job market

Fields

Behavioral & Experimental, Theory

Job market paper

Strategic Complexity in Decomposed Games

A common assumption in game theory is that each player’s scalar-valued payoff is the sum of utilities from different aspects of an outcome. The predictions therefore depend only on the aggregated payoffs and are independent of how these payoffs are distributed across aspects. However, players may evaluate these aspects separately, and some aspects are easier to reason about strategically than others. I experimentally examine whether the strategic complexity of each aspect affects behavior by comparing games presented as a single payoff matrix with the same games presented as two component matrices whose sum equals the actual payoff. I find that game play systematically shifts toward the action recommended by the strategically simpler dominant component, in which each player has a strictly dominant action. To account for this pattern, I develop a behavioral weighting model that allows payoffs from different components to receive different weights. The model captures systematic variation across payoff-equivalent decompositions and significantly outperforms the best-fitting aggregate-payoff benchmark in both in-sample fit and leave-one-game-out prediction.