Feature Impact Analysis on Top Long-Jump Performances with Quantile Random Forest and Explainable AI Techniques

Machine learning can reveal hidden performance factors: The study used AI techniques to analyze long-jump biomechanics and discovered that beyond commonly known velocity factors, the knee angle of the supporting leg before take-off is a critical factor for elite male athletes' performance—a finding that traditional expert analysis might have overlooked.

Explainability tools make AI insights trustworthy: By applying SHAP (SHapley Additive exPlanations), Partial Dependence Plots, and Individual Conditional Expectation plots, the researchers made their machine learning model's decisions transparent and interpretable, allowing coaches and athletes to understand why the AI identified certain features as important rather than treating it as a "black box."

AI works best alongside human expertise: The research combined machine learning with expert-proposed biomechanical features rather than replacing expert judgment, demonstrating that responsible AI in sports uses data analytics to enhance—not substitute for—human knowledge and experience.

Focusing on elite performance requires specialized methods: The study used quantile regression to specifically model top-performing jumps (the top 10%) rather than average performance, showing that AI systems should be tailored to the specific question being asked to provide actionable insights for high-level athletes.