A Variety of Players Type: A Data-Driven Approach Applied to the Game of Bridge

Abstract

Games have long been used as models of complex social interactions. One key observation is that not all players behave exactly alike. Can we model this betweenplayers variability in a generalizable framework? We tackle this issue using a new paradigm in cognitive science, the game of Bridge. We use a data-driven approach to reveal individual characteristics of bridge players ( N = 1300) from a 66-items Bridge Inventory. We uncover 5 factors: Aggressiveness, Discipline, Creativity, Emotionality and Experience. Based on them, a gaussian mixture model identifies 3 bridge players’ types: Conventional, Measured and Subversive. We show that personality traits do not convey all the information needed to recover bridge players’ types. These findings provide an ecologically valid, meso-level description of between-players variability, that stand at an intermediate level between broad personality traits and specific patterns of behaviors. We argue that this framework is extendable to other social interactions.

Keywords: personality, players variability, psychometry, bridge game, data mining