Think you know your tennis? Quick question. Who has the fastest reaction time on service returns in the men’s game?
Popular opinion says it’s Novak Djokovic, widely regarded as tennis’ best returner. Or Roger Federer, known to read serves quicker than anyone else. But official data, released by Tennis Australia’s Game Insight Group (GIG) has a different story to tell.
The answer is Nick Kyrgios.
Surprised? Based on numbers from Australian Open tournaments from 2014-16, Kyrgios’ average reaction time is a blinding 0.613 seconds, fractionally ahead of Federer (0.618). Djokovic, for the record, is seventh (0.638).
| Player | Seconds to hit return |
| Nick Kyrgios | 0.613 |
| Roger Federer | 0.618 |
| Novak Djokovic | 0.638 |
| Andy Murray | 0.643 |
| Grigor Dimitrov | 0.645 |
| Stan Wawrinka | 0.649 |
| Kei Nishikori | 0.653 |
| Tomas Berdych | 0.659 |
| Rafael Nadal | 0.711 |
For regular viewers accustomed to match data in a certain way — first serve %, second serve %, unforced errors, forced errors, aces — stats such as reaction time can seem like an oddity. Tennis has been slow to grasp complex data and analytics, especially ones that concern strategies and patterns of play. In ESPN Magazine ’s Analytics Issue released in 2012, major sports were ranked according to analytical advancement and tennis was second to last, ahead of just boxing.
But that is changing, going by the deep-dive tactical analyses conducted at the Australian Open this year. For Dr. Stephanie Kovalchik, data scientist at GIG, this has been long overdue. “Ever since the Hawk-Eye challenge system was introduced in 2006, tennis has been sitting on an untapped gold mine. A lot of data that has been floating around is easily observable — a regular viewer can keep track of the number of aces a player has hit. What data analytics should focus on is putting numbers to those concepts that are a bit more complex and not easily observed.”
How do you read this data? For instance, what does Kyrgios’ reaction time signify? GIG describes this metric as ‘the expected seconds a returner takes to make impact with a serve travelling at average speed from the time it passes the net’. But you can’t help but wonder if reaction times are just a function of where players stand. Nadal’s relatively slow reaction time of 0.711 seconds could just be the result of how far he stands behind the baseline to return.
“These descriptive statistics just focus on one aspect of the game. It cannot be construed that the number one player in these metrics is the number one in the world,” Kovalchik says. “All this data should be placed in context and understood in layers.”
GIG released other statistics which could be looked at in tandem with reaction time such as return pressure, speed and accuracy of ground-strokes; a combination could explain why certain players are better than others.
| Player | Seconds for return to pass net |
| Novak Djokovic | 0.425 |
| Kei Nishikori | 0.43 |
| Milos Raonic | 0.432 |
| Tomas Berdych | 0.434 |
| Nick Kyrgios | 0.445 |
| Juan del Potro | 0.446 |
| Roger Federer | 0.457 |
| Andy Murray | 0.458 |
| Grigor Dimitrov | 0.47 |
| Rafael Nadal | 0.488 |
| Stan Wawrinka | 0.489 |
Some data helped endorse established opinion. For instance, who is the ‘hardest working’ player? Using tracking data of player movement, data scientists at GIG combined the speed, direction, and distance covered by athletes into a single number that encapsulates their total work during a rally in units of joules. Predictably, Andy Murray — known for his baseline slogging — was found to have the highest work rate, expending an average of nearly 350 joules per shot.
| Player | Work/shot (in joules) |
| Andy Murray | 348.62 |
| David Ferrer | 341.66 |
| Rafael Nadal | 330.46 |
| Gael Monfils | 318.08 |
| Novak Djokovic | 316.18 |
| Roger Federer | 312.93 |
| Stan Wawrinka | 309.86 |
| Milos Raonic | 309.5 |
| Lleyton Hewitt | 304.46 |
| Nick Kyrgios | 265.83 |
| Ivo Karlovic | 237.37 |
GIG also plans to work with players and coaches to analyse match-play data to help make informed, objective decisions. Kovalchik realises that data analyses might not be the panacea, but they could help with devising game-plays and strategies.
Craig O’Shannessy, a leading tennis analyst, feels data exists only to explain strategy. “To formulate strategy, you need to go to data. The role of data analyses should be to remove opinion and guesswork and replace it with facts.”
Besides being an analyst at major tournaments, including the Australian Open this year, O’Shannessy runs Brain Game Tennis, his own strategy business, and has also been a coach for 20 years, developing players like Kevin Anderson and Rajeev Ram. As a coach, O’Shannessy’s methodology involves collecting videos of his players and running them through a match-tagging software that helps unravel patterns and percentages. “The language of tennis is the language of numbers. When I sit on the sidelines and watch, I get one view. But when I record it and assign a tagging software, I get a different one. I will trust the numbers more than my eye. The naked eye is probably the worst way to evaluate a tennis match.”
O’Shannessy believes more coaches are becoming open to using data. In 2015, Angelique Kerber summoned her coach Torben Beltz after the first set of her opening match at the Bank of the West Classic in California. Since 2008, WTA rules allow coaches an on-court visit of 90 seconds per set. Beltz was down with Kerber — but this time, he was armed with an iPad. The tournament was the first in which players and coaches used SAP-equipped tablets that displayed detailed in-match data. Kerber, then ranked 11, went on to win the tournament.
It is, of course, difficult to judge whether data helped Kerber win, although Beltz said it helped her improve her tactics. But perhaps, the more compelling question is what data analysis could mean for the player-coach dynamic.
O’Shannessy says coaches should focus on giving just the right amount of information. “The art of coaching is essentially being able to look at the numbers and explain it efficiently and succinctly to the players.”
Notable in his experiments was his work with Dustin Brown’s team in 2015. O’Shannessy devised a data-driven strategy that helped Brown defeat two-time Wimbledon champion Nadal, in one of the most memorable matches of the 2015 Championships. “The success of the match was organising the chaos inside Brown’s head, to get him used to the patterns of play employed by Nadal,” says O’Shannessy. “That is the data that matters to me the most — the kind that will decide how matches are won and lost.”
The Keys story
Some of the statistics rolled out by the Game Insight Group are fascinating, especially in popular categories like stroke speeds, and challenge popular assumptions.
For instance, the average forehand of 21-year-old Madison Keys not only ranked fastest among women, at 79.9 mph, but was also ahead of all men. Her average backhand speed (74.1) was also the quickest across both tours.
Indeed, even Keys’ median forehand speed (81 mph), while slower than some male players’, is faster than Nadal’s (79.32 mph), Djokovic’s (79.21 mph), Federer’s (76.18 mph) and Murray’s (74.87 mph).
When the debate surrounding equal pay surfaced last year, bio-mechanists and exercise scientists spoke of how women have less of a physical capacity for tennis than men. But GIG data shows that men’s and women’s ground-strokes are comparable – most male players on average reach speeds of 71 to 83 mph on the forehand, women 70-79 mph.
The men, however, separate themselves on serve. The average first serve for men measures 115 mph, the average first serve for women 99 mph.
