AESOP home


Predicting the outcomes of tennis matches using a low-level point model

Demetris Spanias, William J. Knottenbelt

Journal Article
IMA Journal of Management Mathematics
IOS Press
DOI 10.1093/imaman/dps010

Quantitative analysis in tennis has become more important in recent years with the introduction of online betting exchanges. In this paper, the probability of winning a match in tennis is derived hierarchically starting from the probability of each player winning a point on their serve. Under the assumption that each point is identically and independently distributed, the probability of each player winning a point when serving and when returning serve is derived using player statistics. These probabilities can be used to calculate probabilities of winning a game, a set and consequently the match from any point in a live game. Simulated results are then tested against real-world data and compared against existing models in the literature.

Information from