Abstract Professional men’s tennis has transformed significantly since the Open Era began in 1968, with shifts in playing styles, competition levels, and match patterns driven by advances in technology, training, and rules. However, systematic quantification of these changes remains unexplored. This study examines approximately 198,000 ATP matches from 1968 to 2025 through network analysis and machine learning techniques, applying concept drift methods to characterize temporal changes and understand how the sport has evolved across eras and court surfaces. Key findings include greater overall competitiveness, with more frequent upsets and fewer one-sided wins, as top-player dominance has spread more evenly across the field. Matches have grown longer by about 16.5 minutes on average since 1991, serves have become more effective (with 1.7 more aces per match and 2.4% higher win rates on serve points), and return effectiveness has declined by 2.4%. Court surfaces show increasing similarity, with grass courts converging toward the pace of other surfaces. These findings quantify the sport’s shift toward a more powerful, endurance-based game, amid advancements in racket technology and player conditioning. This research demonstrates the power of machine learning to uncover the evolutionary patterns of the tennis game.
Bayram et al. (Thu,) studied this question.