Matthew S. Tenan, Ali R. Rezai
Defense is important in basketball, but how much does defender proximity actually CAUSE a player to miss a shot? The goal of this current work is to show within a Causal Inference framework how valid causal conclusions can be made from high-dimensional player tracking data in basketball. We demonstrate and quantify the interactive causal effect of defender proximity and shot distance from the hoop on probability of success. While much focus has recently been on prediction modeling in sport, we postulate that Causal Inference is likely to have a higher return-on-investment when it comes to enhancing athlete and team performance.