We study a setting where two players prepare for a game and observe what their opponent has played in previous instances of that game. Depending on these observable preferences, we examine how much preparation time is put into the different alternatives and which options are chosen. In a static one-shot game, both mixing and being a predictable expert playing only one option occur in equilibrium. Specialization is always a stable outcome in the dynamic game with adapting experience whereas mixing is sometimes not. To show this, we introduce a new way of applying evolutionary game theory tools to dynamic games.
A consumer asks an expert repeatedly for advice. The expert's incentives are not aligned with the consumer's preferences because he can get a bonus if the consumer takes certain actions. Over time, the expert gets to know the consumer and is therefore able to give better advice (if he wants to do so). In simple equilibria, both - consumer and expert - benefit from the expert's learning if learning is such that the expert's best guess to what is the best advice for the consumer becomes more precise. This gives a natural explanation for why consumers have a preference for personal advice and also for why most internet users do not use anonymization tools.
with Julian Teichgräber
We study congestion problems in decentralized, two-sided matching markets. The main focus is on the impact of market thickness on these congestion problems. We find that it is often not optimal to make an offer to the best observed agent when the likelihood of acceptance is very low. We derive the optimal strategies depending on the market thickness and analyse who benefits the most from the market becoming thicker.