Tennis Note #16
Wimbledon’s Maestro: The Numbers
After exploring the King of Clay and his record in Paris, I considered various different topics to explore prior to Wimbledon but there really was no contest in my mind. Regardless of who wins Wimbledon this year, there is only one maestro over the last 15 years: Roger Federer. As a teenager, he notoriously defeated Pete Sampras, the previous maestro with an equally stunning record (63–7, 100% in 7 finals). Side note: I recommend you read his piece in the player’s tribune.
The grass season covers a small patch of the year long season and does not include the same number of tournaments as clay or hard court. For this reason, I will only analyze Wimbledon and for specific details, I will explore certain finals. Unfortunately, there is no data charted for his losses, which would have been more valuable. Another disclaimer: I am trying out chartblocks, which allows for some interactive graphics, so let me know what you think! It does have its limitations and I think you will need to read this article on a computer.
Roger Federer’s serve is very versatile, consistent, and unreadable, mostly due to the service motion itself. The grass surface creates a lower bounce and faster pace so this further supports him. Typically, 14% of his serves are aces when he wins. You can compare his averages for losses (below the graph) to each year.
My questions were a little different. For example, is Roger Federer’s serve placement independent of his opponent? It is hard to tell but I do think there is a strategy built in. One notable difference is the tactic of first serve against Rafael Nadal in the 2006 and 2007 final. If I had 2008 data, I am almost certain it would reveal the development of a pattern. Another observation — serve placement for a great server (2009 Roddick) and great returners (2012 Djokovic, Murray).
With the revival of his net game as a prominent feature, this is another part of the game plan worth further analysis. However, a lot of the actual net data was incomplete. Instead, I answered a different question: how many of those points were won in less than 3 shots? It is really a question of how quickly Federer finishes points. I took the percentage of points won and determined the distribution of total points won and points won on serve based on rally length.
How do you read these charts? Of the percentage of points won, 1–3 shot rallies make up a majority of the distribution. If you consider only points on serve, it further illustrates this, especially if you determine the percentage of winners from 1–3 shots on serve. I included this year’s Halle final, which is completely different from Wimbledon, just to compare past results to his most recent win. Curiously, his 2012 final against Murray differed quite a bit. He only won 47% of the points from 1–3 shot rallies in which, 62% of these points were on serve and 45.5% of these were winners. Compare this to his previous match, against Djokovic, in which he won 62% of the points with 1–3 shots and 80% of these came when he was on serve. While only 26.7% of these were winners, it just means that Djokovic hit a forced or unforced error. I did not include this year’s Halle semifinal match against Ivo Karlovic but 79% of his points were won 1–3 shots with 83% of them were on serve.
Finally, is the timing of his serve actually random or is there a pattern? I mean, despite the brilliant disguise in the service motion itself, it is in our human nature to be somewhat predictable. Is there a higher probability chance Federer will hit a particular serve in a particular situation? I will leave this visual up to your interpretation but it is important to remember this is only for 5 matches against 4 different opponents.
When Federer loses, his average return points decrease quite below the averages. It makes sense, I mean you cannot simply win from your service game alone. A note about the average for losses — if I do not include 2001 or 2002, the average for losses change. Average 1st serve return rises to 25.4% and average 2nd serve return decreases to 40.1%.
Now consider the fraction of total points won, which is the total points Federer won with respect to all the points in the match. Then consider just the return points Federer won and his success rate. With the faster pace, it is typically harder to return serve, especially with big servers. Most of those matches feature plenty of tiebreaks. On the other hand, if you see higher success rates on return points won, these matches were rather quick 3 set matches.
Take a look at the data points with the lowest fraction of return points won. The lowest green dot was the defeat of Soderling with a score of 6–4, 7–6(5), 7–6(5). The next two lowest green dots were in 2009 against Ivo Karlovic and Andy Roddick, respectively. Of the losses, the lowest blue dots were losses to Mario Ancic in 2002, Sergiy Stakhovsky in 2013, and Jo Wilfred Tsonga in 2011, respectively. The highest return rates were typically very quick 3 set matches. The last piece of the puzzle was return placement but the data was too incomplete for me to truly capture any subtle differences over the years and against different players.
Following from the importance of return at Wimbledon, I examined Federer’s ability to save and convert break points. I thought about combining this data into one graph but decided against it because the two visuals tell two different stories. The first explores Federer’s ability to prevent break points faced and generate break point chances. The second explores Federer’s success rate at saving and converting break points at Wimbledon. Of the total points in graph 1, what fraction was saved/converted, which is illustrated in graph 2?
From 2001 to 2006, there is a reduction in break points faced and an increase in fraction of break points won. This probably corresponds to Federer’s development. However, there is an increase in break points faced for some time until last year and of these break points faced, Federer saved a majority of them. I do not see as much of a trend with break points gained except he did exceptionally well in 2012 to gain break points and converted around his average number of break points (49%) for the years he wins the title. In all instances, except for the first two years, he gained more break point opportunities than he faced. I will leave you to compare the rest of these numbers with his averages to his losses (remember he had early losses 2002 and 2013).
This is an overwhelming amount of data. I have tried to present as much as I can in the most appropriate way possible based on the questions that came to my mind. Many questions were dropped because there was not enough data but I hope in the next years, I can satisfy all my curiosities. I did not have the opportunity to truly analyze the data as much as I wanted but I would loved to hear your thoughts on Wimbledon’s maestro and his consistency for the last 12 years.
All data is from Tennis Abstract’s amazing database available on GitHub. I talked about it previously. Special thanks to those who took the time to chart the matches. If you enjoy reading these tennis notes, make sure to follow the publication, ‘Recommend’ and share! Check us out on Facebook! Made a cool observation? Interested in certain topics and writing? Are you a tennis photographer? Comment, add notes, and check out the submission guideline. Let me know which visuals are good and which are not so great. Cheers!
