Alex Possajennikov's Research

I am doing research in Game Theory (Mathematical Models of Multi-agent Interactive Decision Making) and its applications to Economics, using theoretical and experimental methods. Most of my research focuses on dynamic processes of learning and evolution in games. I am a member of Centre for Decision Research and Experimental Economics (CeDEx) and of Economic Theory Centre (UNET) at the School of Economics, University of Nottingham.

Current Research

Currently I am working on investigating, theoretically and experimentally, behavior and beliefs in various classes of games, including bargaining games with incomplete information, and designing mechanisms under partial information. The following projects are works in progress.

• "On Unbeatable Strategies in Games"
  The paper considers the definition of unbeatability of a strategy in a game as a strategy that gets a player a payoff that is not lower than (not beaten by) payoff(s) of other player(s). It defines different strengths of unbeatability (unbeatability-for-sure, unbeatability-in-expectation, unbeatability-on-average) and illustrated these concepts in economic games. For two-player finite games, it is shown that an unbeatable-in-expectations strategy always exists for at least one player. [Paper (September 2023)]

• "Minority Protection in Voting Mechanisms - Experimental Evidence" (with Engelmann, D., Gruener, H.P., Hoffmann, T.)
  In an experiment, individuals can suggest one of the threshold voting rules after receiving information about the distribution of possible valuations for the alternatives but their own valuation is not yet realized. Depending on the skewness of a distribution, different threshold are optimal in the sense of expected payoffs. We find that individuals change their proposed rules in the direction predicted by the skewness of the distribution but not perfectly, due to the presence of individuals who react to distribution changes differently. [Paper (March 2023)]

• "Private Value Bargaining with Communication: Does Cheap-Talk Lead to Efficiency?" (with Saran, R.)
  In a setting with a seller and a buyer having private information about their cost or valuation, we run an experiment in which we allow face-to-face communication before traders set bid/ask in a double auction. Even though a fully efficient equilibrium exists in this setting, full efficiency is not always achieved in the experiment. Although many traders use "dyadic" strategies of revealing cost/valuation and agreeing on the middle price, some players try to exploit these strategies, leading sometimes to a breakdown in trade. [Paper (December 2022)] [Experiment Materials]

• "Strategic Sophistication, Lying Costs, and Learning in Private Value Bargaining"
  For a double auction with a seller and a buyer having private information about their cost or valuation, the paper considers learning from low levels of strategic sophistication. Simulation of experience-weighted attraction learning, especially with added costs of deviating from cost or valuation, are able to reproduce many features of experimental results both  with continuous and with discrete distributions of costs/valuations. [Paper (December 2020)]

Publications

See the Publications page for details of publications and links to published or pre-print versions of papers.