A Data-Driven Approach to Deduce Players’ Preferences from In-Game Interactions in Gameful Systems

The way players interact with the system during the gameplay is a valuable source of data. While playing, they encounter a plethora of choices and, when they select an option rather than another, they intrinsically express a preference. This knowledge can be exploited in the analysis phase of the game to fine-tune the game towards each user, either manually or through algorithms. However, misinterpreting the players’ actions could be very harmful. It could lead to misleading choices and could result in players’ abandonment rather than in a better experience. In this preliminary work, we present a methodology to infer player preferences by analyzing the choices made during the gameplay. We gathered data from an on-the-field gameful system and applied the paired comparison technique to compute a preference score for the options available. We, then, tested the reliability of the method by computing players’ consistency in their choices, and, partially, by predicting their last choice based on their previous selections.