Research on Momentum in Tennis Based on Random Forest

Abstract

In tennis, there is an increasing emphasis on the significance of “momentum,” particularly in comprehending and effectively using this phenomenon in a scientific manner. However, The existence of momentum in competitions is sometimes challenging to scientifically prove. This work strives to develop techniques based on random forest to quantify momentum and predict players’ winning rates. Upon analysis, the paper extracts three features that have a significant impact on the player’s winning rate in a match. Extensive experiments conducted on two real-world datasets from the Wimbledon Championships show the strong performance of our predictive model. Next, other tennis matches like the US Open are selected to make predictions that demonstrate the good ability of generalization. A run test is also conducted to support the model further. Therefore, there is sufficient confidence to assert that the defined momentum is effective. Finally, this work provides a comprehensive overview of the role of “momentum” and offers insightful suggestions for players to enhance their performance on the court.