Rod Cross Feb 2008, Sept 2010


The web site for each Grand Slam tennis event includes a mountain of statistics for every match played and for the event as a whole. The data are presented in a manner that is helpful in seeing how each player performed in each match, but lots of other interesting data can be extracted by averaging over the whole event or by taking some other average. For example, the table below shows how the top 16 male players performed in each of the last four events up to the Australian Open in 2008. The data were extracted from each match played in the fourth round (16 players), the quarter finals (8 players), the semi-finals (4 players) and the final (2 players). That is, the data were extracted from 15 matches played at each of the four Grand Slam events, and then averaged over those 15 matches for each event.


A player can win a point in 5 different ways. He can win the point by (1) serving an ace, (2) hitting a clean winner during a rally, (3) his opponent serves a double fault, (4) his opponent makes an unforced error or (5) his opponent makes a forced error. The data on each web site do not include any direct information on methods (2) or (5), but it is easy to figure out from the information supplied. Usually, the winner of each match wins more points in total than his opponent. If we count the total number of points won by the winner, and the total number of points won by the loser, and the number of points won by each player by each of the 5 different methods, then we can work out the percentage of points won by each of the five different methods for both the match winner and the match loser. The table below shows these percentages as an average over all the winners and all the losers at each event. In the table, W = winners and L = losers.





For example, only 2 or 3% of all points won by a player are won by his opponent serving a double fault. On average, match winners serve fewer double faults than match losers, except at the French Open, where the loser wins a slightly greater fraction of his points by the winner serving double faults.


Winning players at Wimbledon win 10.4% of their points by serving an ace. Their opponents win only 7.4% of their points by serving an ace. At the French Open, it is much harder to ace an opponent, with the result that players win only 3 or 4% of their points by serving an ace.


All winners and losers win about 1/4 of their points by hitting clean winners, except at the French Open where clean winners account for only 10.5% or 10.6% of winning points.


About 61% of all winning points are won because the opponent makes an error, except at the French Open where 83% of all points are won because of a mistake made by the opponent. For that reason, the French Open web site does not list Unforced Errors in their overall event summary table. Either the column is not wide enough or it might look bad if all the best players in the world are making so many mistakes on their clay courts. One of the reasons that Nadal keeps winning the French Open is that he makes very few unforced errors compared with his opponents.


The high number of forced errors at Wimbledon is due to the fact that grass courts are fast. The ball bounces fast and low, so the player has trouble reaching it and returning it. Rallies are generally short. The high number of forced errors at the French Open is due to the fact that clay courts are slow. The ball bounces high and at low speed. Many of these shots would be clean winners on grass but because the ball bounces more slowly on clay, players can usually get their racquet to the ball but may not be able to hit a good return shot.


The data show that the hard courts used at the US and Australian Open events are similar in terms of court speed and bounce, at least in 2007 and 2008. Event statistics for previous years show that the US courts used to be slightly faster than the Australian courts.


The data in the table show that the winners and losers win points in essentially the same manner at each individual event, and that the manner varies from one event to another. The information in the table does not show why winning players are better than losing players, but the answer is obvious. The winners generally win more points than the losers, either because they hit more winners or because they make fewer mistakes, or both. For example, consider the semi-final at the Australian Open in 2008 where Tsonga beat Nadal 6-2, 6-3, 6-2. Tsonga served more aces (17 to 2), hit more clean winners (32 to 11), served the same number of double faults (1 each), made fewer forced errors (16 to 27) but played a much more agressive game and made more unforced errors (27 to 12).




Grand Slam events have evolved over the last 10 years in several surprising ways. In 2001, average first serve speeds for men were 163 kph at the French Open, 168 kph at the Australian Open, 178 kph at the US Open and 182 kph at Wimbledon. These serve speeds reflected the actual court speeds. Since clay courts are much slower, players found that it was better to use a lot of topspin when serving, thereby reducing their serve speed.  Serve speeds have increased since then and are now essentially the same at all four events. In 2007, average first serve speeds for men were 183 kph at the French Open, 185 kph at the Australian Open, 186 kph at the US Open and 190 kph at Wimbledon. In 2008 the average first serve speeds for men were 188 kph at the French Open, 185 kph at the Australian Open,  183 kph at the US Open and 191 kph at Wimbledon. The speeds here were averaged over all players who reached the 4th round and beyond (effectively 30 players), ignoring all serve speeds during the first 3 rounds since there were too many to count and since not all serve speeds were recorded by a radar gun on all courts.  The 2012 data here includes 3rd round results and beyond.


During the same time, the number of double faults has dropped,  despite the increase in serve speed. The number of aces has remained fairly constant in recent years, except at the French Open where aces have risen with the increase in serve speed, and so has the number of tie-break sets, as shown by the following graphs. The number of aces at Wimbledon rose dramatically in 2010, suggesting that the courts were much faster, despite media comments that they were slower.


The changes since 2000 are probably due to the change from natural gut to polyester tennis strings which made it easier to generate topspin.  That would account for the drop in number of double faults and the increase in serve speed.



A description of these and other Grand Slam graphs can be found at in issue 49 (December 2009) of  the ITF Coaching journal (Coaching & Sport Science Review) or it can be downloaded here.