Quantile Regression for Human Sports Motion Feasibility
October 28, 2026
A lightweight multilayer perceptron using conditional quantile regression on FIFA World Cup 2022 tracking data outperforms six physics-based player-motion models. The model estimates the lower envelope of time-to-arrival (TTA) to predict the shortest plausible time a player can reach a location.
HOW THIS AFFECTS YOU
●
researcherThis demonstrates that learned quantile regression can better capture kinematic constraints than closed-form physics models.