Finding the Best Tennis Serves with K-Means and GMM Clusters of Ball Tracking data to Interpret Serve Strategies
Resumo
The serve is a crucial shot in tennis, that dictates a player’s advantage. However, there has been a noticeable gap in recent data analysis focused on player behavior during serves, when compared to data analysis adoption in other sports. With high speeds, precision, and small margins, ball-tracking systems like Hawkeye are essential for capturing serve steps with fidelity. This data is crucial for decision-making improvements, performance enhancement, and knowledge discovery. However, the Full Hawkeye data is not publicly available. In this manner, this article uses scraping techniques to harness Hawk-Eye serve tracking data from the Australian Open (2020-2024) and Roland Garros (2019-2024), consisting of 152.761 serves from 951 matches. K-Means and Gaussian Mixture Model (GMM) clustering models were employed to discover clusters that summarize thousands of servers into interpretable serve strategies. The best serve strategies optimize success percentages, risk of missing the serve (fault), and may vary from first to second serves, or be affected by pressure in breakpoints, thus the best serve is a serve that best fits a situation and matches a desired outcome. The relation between the serve success and best players was checked, by correlating the server ranking with cluster success using serves from these clusters in different context scenarios. We discovered that the success rate in the clusters increases with player ranking points in high-pressure situations, such as breakpoints and tiebreaks, also that, the hard courts at the Australian Open have greater success rates, while the slower clay courts at Roland Garros have lower first and second serve success rates, despite using similar serve strategies, and that rankings had little bearing on serve performance on these slower courts, indicating that in this surface, other factors may matter more for player advantage in the end than just winning points with the serve right away.
Palavras-chave:
Ball-Tracking, Clustering, Data Mining, GMM, K-Means, Serve Strategies, Successful Serves, Tennis Serves
Referências
Aiken, L. S., West, S. G., and Reno, R. R. Multiple regression: Testing and interpreting interactions. sage, 1991.
Akaike, H. A new look at the statistical model identification. IEEE transactions on automatic control 19 (6): 716–723, 1974.
Association of Tennis Professionals. Atp tour singles rankings, 2024.
ATP Tour. ATP Rankings FAQ. [link].
BBC Sport. Hingis praises mauresmo display, 2006.
Dworkin, M., Barker, E., Nechvatal, J., Foti, J., Bassham, L., Roback, E., and Dray, J. Advanced encryption standard (aes). Federal Inf. Process. Stds. (NIST FIPS), 2001. [online].
Fangraphs. Fangraphs. [link], 2024.
Hawk-Eye Innovations. Hawk-eye innovations, 2024.
Infosys. Infosys atp tennis platform, 2024.
Innovations, H.-E. Hawk-eye’s accuracy reliability line calling. Hawk-Eye Innovations: Basingstoke, UK, 2015.
Lewis, M. Moneyball: The Art of Winning an Unfair Game. W. W. Norton & Company, 2003.
Mecheri, S., Rioult, F., Mantel, B., Kauffmann, F., and Benguigui, N. The serve impact in tennis: First large-scale study of big hawk-eye data. Journal of Sports Sciences 34 (1): 29–37, 2016.
O’Donoghue, G. P. and Brown, E. The importance of service in grand slam singles tennis. International Journal of Performance Analysis in Sport 8 (3): 70–78, 2008.
Rousseeuw, P. J. Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. Journal of computational and applied mathematics vol. 20, pp. 53–65, 1987.
Schwarz, G. E. Estimating the dimension of a model. The annals of statistics, 1978.
Society for American Baseball Research. The research collection. [link], 2024.
Tea, P. and Swartz, T. B. The analysis of serve decisions in tennis using bayesian hierarchical models. Annals of Operations Research vol. 325, pp. 633–648, 2023.
Wei, X., Lucey, P., Morgan, S., Carr, P., and Matthews, I. Predicting serves in tennis using style priors. Proceedings of the 21st ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016.
Whiteside, D. and Reid, M. Spatial characteristics of professional tennis serves with implications for serving aces: A machine learning approach. Journal of Sports Sciences 34 (22): 2073–2081, 2016.
Akaike, H. A new look at the statistical model identification. IEEE transactions on automatic control 19 (6): 716–723, 1974.
Association of Tennis Professionals. Atp tour singles rankings, 2024.
ATP Tour. ATP Rankings FAQ. [link].
BBC Sport. Hingis praises mauresmo display, 2006.
Dworkin, M., Barker, E., Nechvatal, J., Foti, J., Bassham, L., Roback, E., and Dray, J. Advanced encryption standard (aes). Federal Inf. Process. Stds. (NIST FIPS), 2001. [online].
Fangraphs. Fangraphs. [link], 2024.
Hawk-Eye Innovations. Hawk-eye innovations, 2024.
Infosys. Infosys atp tennis platform, 2024.
Innovations, H.-E. Hawk-eye’s accuracy reliability line calling. Hawk-Eye Innovations: Basingstoke, UK, 2015.
Lewis, M. Moneyball: The Art of Winning an Unfair Game. W. W. Norton & Company, 2003.
Mecheri, S., Rioult, F., Mantel, B., Kauffmann, F., and Benguigui, N. The serve impact in tennis: First large-scale study of big hawk-eye data. Journal of Sports Sciences 34 (1): 29–37, 2016.
O’Donoghue, G. P. and Brown, E. The importance of service in grand slam singles tennis. International Journal of Performance Analysis in Sport 8 (3): 70–78, 2008.
Rousseeuw, P. J. Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. Journal of computational and applied mathematics vol. 20, pp. 53–65, 1987.
Schwarz, G. E. Estimating the dimension of a model. The annals of statistics, 1978.
Society for American Baseball Research. The research collection. [link], 2024.
Tea, P. and Swartz, T. B. The analysis of serve decisions in tennis using bayesian hierarchical models. Annals of Operations Research vol. 325, pp. 633–648, 2023.
Wei, X., Lucey, P., Morgan, S., Carr, P., and Matthews, I. Predicting serves in tennis using style priors. Proceedings of the 21st ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016.
Whiteside, D. and Reid, M. Spatial characteristics of professional tennis serves with implications for serving aces: A machine learning approach. Journal of Sports Sciences 34 (22): 2073–2081, 2016.
Publicado
17/11/2024
Como Citar
KALIFFE, Kalil Saldanha; SANTOS, Reginaldo.
Finding the Best Tennis Serves with K-Means and GMM Clusters of Ball Tracking data to Interpret Serve Strategies. In: SYMPOSIUM ON KNOWLEDGE DISCOVERY, MINING AND LEARNING (KDMILE), 12. , 2024, Belém/PA.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2024
.
p. 73-80.
ISSN 2763-8944.
DOI: https://doi.org/10.5753/kdmile.2024.244569.