Date: October 30, 2026
Speaker: Sam Ganzfried, Maxwell Fellow, Center for Data Science for Enterprise and Society, Cornell University
Title: Game Theory Solution Concepts for an Age of High-Stakes Multiagent Decision-Making

Abstract: Nash equilibrium is the central solution concept in non-cooperative game theory; however, it has important limitations as a practical prescriptive framework. We would absolutely not want to deploy an agent in an important multiagent setting that simply followed a Nash equilibrium strategy. I will discuss the two broad classes of limitations of Nash equilibrium and present new solution concepts that address each of them.
Bio: Sam Ganzfried is a Maxwell Fellow at the Center for Data Science for Enterprise and Society at Cornell University. His primary research area is computational game theory. He developed the idea of real-time subgame solving in no-limit Texas hold’em which was the key scientific idea that led to superhuman play. He has worked on algorithms for computing Nash equilibria in complex game classes: games with extremely large state spaces, games with imperfect information, multiplayer games, stochastic games, and continuous games. He has also devised several new practical and robust solution concepts. His research has also addressed modeling and exploitation of suboptimal opponents, computation of human-understandable strategies, and applications to national security. Recently he has been interested in applications of AI to cancer treatment, which is the focus of his project at Cornell. He has a Ph.D. in computer science from Carnegie Mellon University.