Learning Human Sports Motion Feasibility via Quantile Regression on Tracking Data
What it is
This paper proposes a method to estimate time-to-arrival (TTA) in sports by treating it as a conditional quantile regression problem rather than using physics-based models. It trains a lightweight multilayer perceptron on optical tracking data from the 2022 FIFA World Cup to predict the lower envelope of player locomotion times, demonstrating superior performance over six existing physics-based baselines.
The problem it solves
Existing physics-based models for player motion often fail to capture the true lower envelope of human locomotion times, leading to less accurate time-to-arrival estimates compared to data-driven approaches.
What you could use it for
- Predicting player reachability for defensive positioning in soccer
- Analyzing player motion dynamics and transferability across different sports like basketball and NFL
- Developing data-driven models for sports strategy and spatial ownership
Why it matters
Provides a more accurate and data-driven alternative to physics-based models for predicting player movement, potentially improving tactical analysis and automated decision-making in sports analytics.
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.