Title: Measuring Strategic Complexity of Mechanisms with an Application to Matching
Any direct revelation mechanism induces a type revelation game from the perspective of the agents. For mechanisms that operate on finite type spaces, we introduce a new method for assessing the strategic complexity of this game. Loosely speaking, we aim at quantifying the amount of information (about the other agents' type reports) that is needed for any individual agent to determine its best response. We propose an application of this new method to compare the strategic complexity of matching mechanisms, such as variants of the Deferred Acceptance mechanism and the Boston mechanism.