IAP SIE | Eyes on the Game What Gaze Behavior can Tell us about Player Skill

This lecture explores how gaze behavior reveals underlying cognitive processes that differentiate novice and expert players in dynamic gaming environments. Using Rocket League as a testbed, the study combines eye tracking with computer vision techniques to analyze how players allocate attention across multiple moving targets in real time. By synchronizing gaze data with in-game object detection, the work provides a detailed account of how visual attention strategies evolve with skill. The talk highlights key findings, including differences in spatial distribution of gaze, fixation patterns, and gaze transfer timing between novice and expert players. Experts tend to exhibit more efficient, centralized attention patterns and faster transitions between relevant objects, while novices engage in broader and less focused search behaviors. The lecture also introduces concepts such as target-looking versus center-looking strategies and discusses how these relate to multiple object tracking and decision-making under time pressure. Finally, the session considers applications of gaze analysis for esports training, player modeling, and the development of responsive game systems that adapt to user attention and behavior. Connect with CSAIL Alliances On our site: https://cap.csail.mit.edu/ On LinkedIn:   / mit-csail