Quantitative Analysis of Tennis Ball Momentum Utilizing a Fusion Model of EMA and Bessel Curves

Abstract

Targeting the challenges in quantifying the momentum transmission mechanism during tennis matches and the constraints imposed by traditional models' linear assumptions, this study developed an innovative dynamic model. By integrating logistic regression, exponential moving average (EMA), and Bessel curves, the model captures recent performance trends through EMA weighting while pinpointing momentum turning points using the second derivative of the Bessel curve. When tested with 2023 Wimbledon men's final match data, the model demonstrated a prediction success rate of 78%, surpassing traditional methods by 12%. Moreover, it enables real-time identification of momentum shifts during critical match phases. This advancement not only offers a scientific foundation for enhancing tactical decision-making in tennis but also propels the intelligent evolution of sports analysis, bridging theory and practical application seamlessly. This paper integrates these techniques to provide profound insights into player dynamics and match progression, laying the foundation for more precise predictions and strategic planning in competitive sports.