GDC Summer 2020 Session Viewer

The talk will showcase a novel method to map player features from out-of-game to player behavior within a game. The out-of-game feature in this specific case is player motivation described by the Ubisoft Perceived Experience Questionnaire (UPEQ). First of all it was necessary to collect gameplay in a very granular manner including all possible activities that players can engage with. Additionally it was necessary to ask the same players to report their levels of competence, autonomy, relatedness and presence using UPEQ. Survey responses were processed in an ordinal fashion. Preference learning methods, based on support vector machines, were used to infer the mapping between gameplay and the 4 motivation factors. Our key findings suggest that gameplay features are strong predictors of player motivation as the obtained models reach accuracy of near certainty, from 93% up to 97% on unseen players.

Takeaway

Attendees will be exposed to a novel method that leverages machine learning to potentially shift the paradigm of user research. First of all, this method allows user researchers and analysts to infer responses to questionnaires for the whole player population and not just the respondents. Additionally, it can map potentially any out-of-game player feature to their behavior in game.

Intended Audience

User researchers, data analysts, data scientists, behavioral researchers, brand and marketing teams will greatly benefit from this talk since it has potential direct impact on their processes. Knowledge of introductory statistics, machine learning and AI algorithms is a plus, but not required.