[e]

Game Theory

This year's logic tutorial will be given by Amanda Friedenberg (Arizona State University).

Title: The Role of Strategic Uncertainty in Games

Abstract: The traditional approach to game theory takes two steps: First, the description.  This specifies the rules of the game, outcomes, and payoffs.  Second, the prediction.  This comes via a solution concept, which maps each game into a subset of strategies of the game.  The idea is that, if a strategy is chosen by the solution concept, then we can expect players to choose this strategy.  Underlying the statement is an informal story of how players reason which---presumably---leads to the play of the solution concept.

Epistemic Game Theory (EGT) began from this traditional viewpoint.  It sought to provide a justification for standard solution concepts, e.g., iterated dominance, Nash equilibrium, etc. To do so, EGT amended the description of the game, to include strategic uncertainty---i.e., uncertainty about how others play the game.  This broader description allowed the analyst to formally describe players' reasoning and, so, allowed the analyst to take on the question of justifying standard solution concepts.

More recently, EGT has moved beyond the question of justifying standard solution concepts: Precisely because EGT can provide a language to express strategic reasoning, EGT has provided new solution concepts and new predictions for games.  This course will show how to use the language of EGT to express strategic uncertainty.  We will then turn to express conditions on players reasoning, via so-called epistemic conditions.  We will see that reasoning under strategic uncertainty can lead to novel predictions in applications.