S14: Social Interactions

Thu. March 9, 9:36 a.m. – 9:48 a.m. PST

Room 206

Approaches inspired by density functional theory (DFT) have recently found success in modeling a variety of social systems, often through the use of density-functional fluctuation theory (DFFT). In this work, we extend the application to sports by analyzing highly detailed data of basketball player positions collected over the last decade. Not only does DFFT allow us to determine where a player is likely to be given positions of the other players and the ball, but it also enables accurate prediction of whether a particular position is ‘good’ or ‘bad’, as determined by the probability of resulting in a ‘high-quality’ shot. The underlying DFFT description then provides a data-driven approach to determine how players influence the outcome of basketball plays through their positioning and establishes a new method for quantifying a player's value to the team that is substantially more nuanced than traditional metrics which focus on a player’s actions with the ball. Further, as the approach presented here is quite general, DFFT can be expected to benefit sports more broadly as detailed positional data becomes increasingly available in a variety of amateur and professional sports.