Bayesian Games with Intentions
Adam Bjorndahl Carnegie Mellon University, Pittsburgh, PA, USA
Joseph Y. Halpern Cornell University, Ithaca, NY, USA
Rafael Pass Cornell University, Ithaca, NY, USA
Abstract
We define Bayesian games with intentions by introducing a distinction between
“intended” and “actual” actions, generalizing both Bayesian games and (static)
psychological games [1]. We propose a new solution concept for this framework
and prove that Nash equilibria in static psychological games correspond to a
special class of equilibria as defined in our setting. We also show how the
actual/intended divide can be used to implement the distinction between “real”
outcomes and “reference” outcomes so crucial to prospect theory, and how some
of the core insights of prospect theory can thereby be captured using Bayesian
games with intentions.