Momentum is frequently invoked to explain abrupt shifts in professional tennis, yet many quantitative descriptions either depend on opaque machine-learning outputs or leak post-point information into predictive claims. This study develops a transparent, point-level momentum framework using the public 2023 Wimbledon Gentlemen’s singles point-by-point dataset released with the 2024 Mathematical Contest in Modeling Problem C. The dataset contains 7,284 point records from 31 matches after the first two rounds. We transform point events into signed player-relative increments, combine them through an exponentially weighted moving average, and then evaluate whether the resulting momentum score adds information beyond pre-point score context and serving status. The reconstructed framework separates descriptive visualization from prediction: The momentum index is used to describe match flow, while next-point prediction is evaluated only with information available before a point begins. Results show that serving status remains the strongest predictor of the next point. A score-and-serve baseline achieved a group cross-validation area under the curve (AUC) of 0.683; adding exponentially weighted moving average (EWMA) momentum raised AUC to 0.687, and a full contextual model reached 0.690. This indicates that momentum contains limited but non-negligible incremental signal rather than near-perfect determinism. Large momentum swings were concentrated around winners, aces, close-to-line serves, successful net points, and double faults. The proposed framework provides a reproducible alternative to overfitted momentum models and offers coaches a practical dashboard for identifying high-leverage tactical events without claiming causal psychological measurement.
References
[1] Meier, P., Flepp, R., Ruedisser, M. V., Franck, E. (2020) Separating psychological momentum from strategic momentum: evidence from men’s professional tennis. Journal of Economic Psychology, 78, 102269.
[2] Lv, X., Gu, D., Liu, X., Dong, J., Li, Y. (2024) Momentum prediction models of tennis match based on CatBoost regression and random forest algorithms. Scientific Reports, 14, 18834.
[3] Wang, J., Guo, S., Zhou, Y. (2024) A multidimensional momentum chain model for tennis matches based on difference equations. PLOS One, 19(12), e0316542.
[4] Depken, C. A., Gandar, J. M., Shapiro, D. A. (2022) Set-level strategic and psychological momentum in best-of-three-set professional tennis matches. Journal of Sports Economics, 23(5), 598-623.
[5] Friligkos, G., Papaioannou, E., Kaklamanis, C. (2023) A framework for applying the Logistic Regression model to obtain predictive analytics for tennis matches. Technium, 15, 60.
[6] Xia, Y., Li, C., Zhang, T. (2025) Analyzing momentum shifts in tennis: a machine-learning approach to predicting match outcomes. Applied Sciences, 15(4), 2018.
[7] Liu, C., Yang, J., Cui, Y. (2025) Quantifying and predicting momentum in tennis match via machine learning approach. International Journal of Racket Sports Science, 7(1), 46-58.
[8] Honey, C. J., Mahabal, A., Bellana, B. (2023) Psychological momentum. Current Directions in Psychological Science, 32(4), 284-292.
[9] Zang, Q., Li, Z., Hu, J., Zhang, T., Zhang, W. (2024) Prediction of winning and losing in tennis match based on entropy weight-TOPSIS and machine learning model. Transactions on Computer Science and Intelligent Systems Research, 5, 1529-1535.
[10] Gupta, K., Krishnamurthy, V., Deb, S. (2025) What elements of the opening set influence the outcome of a tennis match? An in-depth analysis of Wimbledon data. IIMB Management Review, 37(1), 100519.
[11] Zhong, M., Liu, Z., Liu, P., Zhai, M. (2024) Searching for the effects of momentum in tennis and its applications. Procedia Computer Science, 242, 192-199.
Share and Cite
Huang, J. (2026) A Transparent Point-level Momentum Framework for Professional Tennis: Evidence from Wimbledon 2023. Scientific Research Bulletin, 3(2), 56-65. https://doi.org/10.71052/srb2024/PAZZ7038