The current investigation was composed of two related studies. The aim of the first study was to compare the proportion of break points and non-break points won by players receiving serve in matches of the 2011 US Open men's singles tournament. Match statistics were recorded from the official tournament website for the 92 matches where both players had at least 5 break points. Wilcoxon signed ranks tests revealed that winning players won a significantly greater proportion of break points than non-break points (p = 0.004) while losing players won a similar proportion of break points to non-break points (p = 0.994). The aim of the second study was to compare the proportion of break points and non-break points won by the World's top 4 tennis players. There were 27 to 39 singles matches for each of these players within Grand Slam tournaments between 2008 and 2011 where the players and their opponents had at least 5 break points each. A series of 95% confidence intervals of the mean revealed different scoreline effects for these players. The findings of these two studies challenge the assumption of stationarity used in models of winning games of tennis.
© Copyright 2012 Journal of Quantitative Analysis in Sports. de Gruyter. All rights reserved.