The probability of winning break points in Grand Slam men's singles tennis

Abstract The probability of winning games of tennis has been modelled in terms of the probability of the server winning an individual point. These models are based on the assumptions of stationarity of the point and interdependence. Stationarity is the assumption that the outcome of tennis points is not influenced by the game score at the beginning of the point. The purpose of the current investigation was to compare the outcomes of tennis points during break points and non-break points using data from 528 men's singles tennis matches from all four Grand Slam tournaments in 2008 and 2009. The retrospective probability of the receiver winning a point was 0.42 ± 0.23 during break points which was significantly greater than the 0.38 ± 0.10 during non-break points (P < 0.001). This, in turn, led to players creating significantly more break points than expected (P < 0.001) and breaking serve more frequently than expected (P < 0.001) for the points won when receiving serve. This is evidence that stationarity cannot be assumed and that models of winning tennis games may need to be enhanced to address improved receiver performance during break points.