Momentum Quantification and Prediction of Tennis Match Based on Time Series and Logistic Regression

Abstract

In the highly competitive sport of tennis, it is crucial to understand and capitalize on every small match advantage, and despite the attention paid to the concept of momentum, its quantification, analysis, and prediction remain a challenge. The purpose of this study was to examine the role of momentum quantification, momentum influencing factors, and accurate prediction of momentum shifts in enhancing athletes' performance in tennis. By analyzing the competitive data of tennis players, a tennis-based "Big Data 'Momentum' Shift Prediction Model" was constructed, which not only specifically quantifies momentum, but also captures the impact of potential information on the changing situation of the game, and predicts the momentum of a match with an accuracy of around 80% accuracy in predicting how the situation will change during the match. Specifically, the construction of high-quality models helps athletes and coaches make timely adjustments to their strategies during the game. These findings are important guidance for coaches, athletes, and researchers in related fields to advance sports science and improve athletes' performance.