Tennis players are allowed to take 20 seconds before serving at Wimbledon, as they are in all Grand Slam tournaments. Some players think it’s time for this rule to start being enforced — with a “shot clock,” a phrase borrowed from basketball and other sports.
“I think it’s the only way to go, to be honest, because how are you supposed to know as a player how long 20 seconds is, or 25 seconds, between a point?” asked Andy Murray, last year’s champion, in a post-match press conference Monday. Two other players — Australian Open champion Stanislas Wawrinka and former world No. 1 Caroline Wozniacki — have also endorsed the idea.
The 20-second rule is almost never enforced. Some players complain this gives those who abuse it an unfair advantage, by giving them extra time to recover. Slow play can also turn off fans and disrupt tournament schedules. Yet some players value extra time to gather their thoughts, catch their breath and wipe their brows. The issue starkly divides the men’s game’s two living legends, Roger Federer and Rafael Nadal, who have comparable career accomplishments but very different paces of play.
A first-time violation of the 20-second rule at a Grand Slam match earns merely a warning. The next and all subsequent violations cost a point. But because umpires almost never hand out time violations, players often flout the rule — sometimes dozens of times in a single match.
That last sentence is qualitative, not quantitative, because data isn’t readily available. Umpires are supposed to track time between points from their chairs, but unlike other stats and events they monitor during a match — the score, aces, first serves in — time between points isn’t reported as an official stat. And the Grand Slams, the men’s tour and the women’s tour don’t include time violations among player stats.
Without an official shot clock to refer to, I took an unofficial one — the stopwatch on my smartphone — to two men’s matches on Centre Court at Wimbledon on Saturday. The first pitted Nadal, known for his slow play, against Mikhail Kukushkin. The second featured Federer, famously fast between points, against Santiago Giraldo.
According to my stopwatch, Kukushkin and especially Nadal were in no hurry, while Federer and Giraldo raced through their encounter. Kukushkin averaged 18.7 seconds before serves, and went over the 20-second limit about one-third of the time. Nadal averaged 22.5 seconds before serves, and went over the limit more than two-thirds of the time.
Meanwhile, Federer averaged 15.3 seconds before serves, and went over 20 seconds on just 4 percent of his service points. Giraldo averaged 16.6 seconds, and went over the limit 18 percent of the time.
The pace of these Saturday matches was particularly relevant because two days earlier, Nadal’s defeated opponent in the second round, Lukas Rosol, complained that Nadal wasn’t penalized for slow play. In his post-match press conference Rosol said, “The referee was not going with the rules.” He added, “Always best players, they’re taking much more than the normal players, you know, and nobody is telling them nothing.”
Federer weighed in that same day, saying, “I just think it’s important that we, as players, play up to speed.” He added, “What I don’t want is that we lose viewers because we play too slow” — noting that he recently found himself among such impatient viewers while watching a match on TV. “They were playing so slow I was like, ‘OK, I really — I can’t watch it.'”
Raw averages don’t tell the whole story. Perhaps Nadal and Kukushkin were playing bruising rallies that required longer recovery time, or were playing many pivotal points, at deuce or break point. To see which factors were significant drivers of how long players took between points, I cross-referenced each serve with stats that Wimbledon data-provider IBM offers journalists at the tournament: how many strokes the prior rally had, and whether it ended with a winner, an unforced error or a forced error. I also tagged each point as either significant or not. Significant points included any point in which the server trailed — since breaking serve is so rare on Wimbledon’s fast grass — plus 30-all, 40-30, deuce and any advantage or tiebreaker points. Then I combined all the serve timings I had for each player from Saturday’s matches — 302 timings in all, at least 50 for each man. And I ran a series of ANCOVA regressions.
It turns out long rallies do make a difference: According to the regressions, each extra stroke in a rally adds about four-tenths of a second to the time the server takes before the next point. The four players also added about two seconds to their pre-serve routine before crucial points. But even after controlling for these factors, the identity of the man serving mattered, too. Kukushkin added about two seconds per serve compared to Federer, while Nadal added seven seconds.
Many tennis fans know that Federer plays faster than Nadal. The two all-time greats often have been compared on pace of play — such as in this video showing Federer finishing a service game in the time it takes Nadal to play one service point — as well as on more significant accomplishments, such as number of Grand Slam titles won. This analysis, of just one match for each, won’t settle anything — and I have better things to do than time every match. Without more comprehensive data, the analysis does suggest that Nadal’s pre-serving routine — including toweling off, picking between at least three balls, touching his face and clothing, bouncing the ball and then rocking his body — affects his pace of play more than the punishing rallies he’s known for.
There are other, incomplete or indirect indicators of players’ pace of play, many of which corroborate the finding that Nadal takes his time before serving. I’ve stopwatched players at prior Grand Slams and found there, too, that Federer is fast and Nadal is slow. Occasionally television broadcasts report average time between points during matches. “You can see on TV the stat, sometimes the average time is 28 or 30 seconds,” Wawrinka said. Federer was told in his press conference that a broadcast of Nadal-Rosol showed Nadal averaged 25 seconds between points. These stats are unofficial, however, and not recorded or compiled in a systematic way.
The crudest method to approximate pace of play is to divide the total time a player is on the court by the number of points he or she plays. Nadal ranks first among men’s tour regulars in time per point, at 46.5 seconds over the last year. Federer is among the fastest, at 38.5 seconds.
Simply dividing time by points is crude because it doesn’t account for time elapsed during points themselves. It also bundles together Grand Slams — where players get 20 seconds before serving — and tour events, where men get 25 seconds. One indicator that the Slam rules aren’t enforced and aren’t heeded by players: Nadal, Federer and the average man all take longer per point at Grand Slams than at tour events.
At Wimbledon, Nadal averaged 40.7 seconds per point, while Federer has averaged 7.2 fewer seconds — with Novak Djokovic and Murray, the other two men to win Wimbledon in the last decade, closer to Nadal’s pace. Again, this measure includes actual tennis, not just time in between tennis. Just 23 percent of rallies in Federer matches through the fourth round extended past four shots, compared to 27 percent for Nadal and 39 percent for Murray.
Will Wimbledon take up players’ suggestions for a shot clock? It’s not clear. I saw Pascal Maria, the umpire who presided over the Nadal-Kushkin match, outside Centre Court on Tuesday and asked him about my data and about time violations. He declined to comment. A spokeswoman for the All England Club, which hosts the tournament, told me umpires had handed out 33 time violations as of Tuesday — or fewer than one every six matches. Some of these were mere warnings; she didn’t know how many times players lost points for playing slow. She declined to comment about slow play and the possibility of a shot clock.
CORRECTION (July 2, 6:14 p.m.): An earlier version of this article said the analysis consisted of a series of logistic regressions. It was a series of ANCOVA regressions.
