Mining Controller Inputs to Understand Gameplay

Explore a groundbreaking approach to game analytics presented at the ACM Symposium on User Interface Software and Technology. Dive into the research of Brian A. Smith and Shree K. Nayar, who demonstrate how analyzing players' controller inputs using probabilistic topic models can provide game developers with quantitative insights into gameplay types. Learn how this method allows for the discovery of action types fostered by games and the extent to which each level promotes different styles of play, all without supervision. Understand the progression from latent Dirichlet allocation (LDA) to the more sophisticated player-gameplay action (PGA) model, which identifies gameplay patterns independent of individual play styles. Discover how this innovative approach can be applied to verify level design, recommend similar gameplay experiences, and even recognize players with high accuracy in a short time frame. Gain valuable insights into this 21-minute presentation that challenges traditional event log-based analytics and offers a new perspective on understanding and quantifying gameplay.

Mining Controller Inputs to Understand Gameplay