This is the first story in a three-part series examining the analytics boom in tennis.
* * *
On the eve of the U.S. Open, Roger Federer proclaimed, “No return, no wins.” He went on to say that Rafael Nadal was “by far the best” in returning serves, and his insight proved to be prophetic. Nadal broke Daniil Medvedev’s serve three times in the fifth set of Sunday’s championship match in Flushing, Queens.
In 2019, Nadal led the ATP Tour in this facet, having won 38% of all return games, a 5% percent advantage over Novak Djokovic and the rest of the field. Nadal himself said recently, “The two first shots of the point are decisive. You play rallies, yes, but every time the game is moving in a more aggressive way.”
This tracks with one of the prevailing insights being revealed by the sport’s analytics boom. Though long rallies may create more indelible highlights, matches are ultimately won in the balance of “early points.” Craig O’Shannessy, Djokovic’s strategy coach and the proprietor of Brain Game Tennis, frequently notes that 70% of all points are decided by four or fewer shots. The player who wins more of those short points will win the match 90% of the time.
HEALING HANDS: What Sports Tech Can Learn From This P.T. Guru
Data proliferation and analysis might be transforming all sports and businesses, but that movement has only started to accelerate in tennis over the past few years. The industry is now making room for individual strategy consultants such as O’Shannessy, as well as boutique firms such as Golden Set Analytics, Tennis Stat and Tennis Analytics. Even major tech companies are entering the space. SAP provides the WTA with an analytics platform. Infosys does the same for the ATP and is a technology partner for two Grand Slams, the Australian Open and French Open. IBM has partnered with the USTA to develop tools for American players and also powers the fan-facing SlamTracker for the U.S. Open and Wimbledon.
Yet how the data is being incorporated remains closely guarded. “It’s more of a secret-agent kind of role,” O’Shannessy says. “You don’t really want other teams knowing what you’re doing and how much you’re doing.”
Djokovic, for one, says advanced data and video have informed his preparation for every tournament for a decade, putting him ahead of the curve. “I have to do my homework,” he says. “I think it’s something that is so logical and so normal. It’s part of, in a way, my job as well to get myself ready for what’s coming up.” At a recent U.S. Open press conference, Federer deemed analytics “very interesting” and added, “In the future, we’ll only use more of it.” (Federer is reported to employ Golden Set Analytics, although the company would not confirm any current client.)
“I do use statistics, not only about opponents, but also about myself,” says Karolína Plíšková, the No. 2-ranked women’s player. “It’s numbers and they never lie. So I think that you can for sure find something good in that, but of course in the end, you can still win without having good numbers.”
“If you’d asked me this five or six years ago, I would have said most of the top pros are not using it,” says Warren Pretorius, the founder of Tennis Analytics who connected the sport with the Dartfish video analysis system. “Now, I’d be shocked if, in the top 50, there are pros that are not using some form of analytics.”
* * *
Darren McMurtrie joined Tennis Australia as a performance analysis manager in 2008. Around that time, he remembers, ESPN ranked all of the sports by their use of analytics: Tennis was second to last, ahead of only boxing. “Tennis has probably been behind most other sports a couple years,” he says, noting that “it’s definitely picked up” in recent years.
Tennis Australia and its subsidiary, Game Insight Group, have been at the forefront of bringing data-driven methods and analysis to the sport. Helping fuel that research is the revenue generated by the Australian Open. The USTA enjoys a similar benefit from the U.S. Open, given the tournament’s collaboration with IBM, which lends its Watson’s AI capabilities and its Red Hat cloud computing power.
ACCELERATING CHANGE: Revolutionizing Athlete Health, Wealth and Performance
IBM first used Watson to generate automatic highlights. That video indexing capacity was then applied to help USTA coaches tag matches and unearth new metrics. One data point developed last year was Red Steps—shorthand for redirect steps—to quantify the number and rate at which a player changes direction.
This year, IBM and the USTA took Red Steps even further to create Coach Advisor, a new platform that introduces Energy System monitoring. Using Hawk-Eye positional data, Coach Advisor measures both a player’s physiological load (cumulative match exertion) and mechanical intensity (bursts of energy, as measured in watts).
“It’s going to be a really powerful tool,” says Martin Blackman, the USTA’s general manager of player development. “Not just for the coaches to see what’s happening during a tournament, but more importantly, for the entire team to make changes in the player’s strength and conditioning program to optimize their performance and mitigate anything that they’re seeing that’s causing poor performance in tournaments.”
SPORTTECHIE’S UFC SERIES
- All the Tech Being Used at the UFC Performance Institute
- What Happens in Vegas Is Going Global
- Science Is Drawing a Fine Line Between Winners and Losers
While Blackman was reviewing one early round match at this year’s U.S. Open involving an American through the Coach Advisor portal, he saw that momentum shifts in the set coincided with inflection points in each player’s physical output. “Looking at the shifts in the score, especially breaks of serve, and seeing how that aligns with the physical exertion and then being able to click on those games and access the video, is really a rich way to look at the data, the energy output, and then dig into the video to pick up the nuances of what was happening tactically,” Blackman says. “It’s just so powerful.”
Coaches are the conduits of this information, exercising their own discretion about what gets mentioned and what gets filtered out so that players aren’t overwhelmed. “That’s a skill in itself, trying to work out what to tell and not tell the player,” McMurtrie says. “It’s mixing the science with the art.” IBM’s Elizabeth O’Brien, a program director for sports partnerships, says the aim of the new tool is to conduct a “first layer of analysis and point their questions.”
The automation of the process is critical, especially with a short turnaround between matches in a tournament. Adds O’Brien’s IBM colleague, Kristi Kolski, “A lot of it has a lifespan. If you can’t quickly ingest that data and draw the insights out and then act on that data, then it’s useless.”
* * *
Tennis has lagged behind other sports in its embrace of data because it’s an individual sport with a wide gap in tour earnings, complicated travel logistics, multiple governing bodies, and the slow arrival of automation.
“Tennis is very fragmented. Even at the top of our sport, you’ve got the ITF, the WTA, the ATP,” O’Shannessy says of the international governing body and the women’s and men’s tours. “They don’t get along that well. They don’t talk to each other a lot. They don’t share data.”
Hawk-Eye is best known for its review of line calls, but the camera system also tracks ball and player movement around the court. Some Hawk-Eye data is made available via the Infosys and SAP platforms, and full reports can be acquired for a fee. Players, however, can only purchase their own matches, according to two analysts, so aggregating enough meaningful data to extract global trends is difficult for smaller operations supporting one player or only a few clients.
Golden Set Analytics staffers have extensive computer science and statistics backgrounds, with team members holding degrees from Harvard, Yale and Stanford. The company declined to speak in detail; one of its principals likened the group to a consulting firm, positioning itself in the background. (Though the company won’t acknowledge current clients, the principal says it previously assisted Novak Djokovic and the USTA during the 2017 U.S. Open when all four women’s semi-finalists were Americans.)
SPORTTECHIE’S BASEBALL SERIES
- The Unlikely Origin of Baseball’s Launch Angle Revolution
- Major League Baseball Will Be Decided by Swing States
- You See the Ball, You Hit the Ball, You Got It?
- How Outsiders Became the Ultimate MLB Insiders
- Bats to the Future: ‘You Can Fight Technology. You Won’t Win’
“Companies like us are nimble,” the source says, noting that past and present GSA clients have notched 12 men’s and women’s Grand Slam victories in the last five years. “We listen to our clients. We learn right at the forefront. We’re not in it to show off. We’re in it only to provide value. Our only goal is to get our players to win.”
The elite players who can afford to collect their data and hire their own analysts will eventually accrue a good sample size, making patterns in play easier to spot.
“Now, whoever’s pulled out of a hat, there’s a strong focus on where the opportunity is to exploit,” says Lorcan Reen of Great Britain’s Lawn Tennis Association, who has been Andy Murray’s performance analyst for five years. “And you look at the trends. I’ve got a library of players.”
NEW BOX SEATS: Ultra Reality Viewing Will Make Its MLB Debut in the Playoffs
But sometimes even top pros have to reply on rudimentary video. Before Plíšková’s first-round U.S. Open match against Tereza Martincová, whom she had never played, Plíšková resorted to some footage found on YouTube. Yet at the same time, the high-end tech is expanding to the lower ranks of the sport.
PlaySight’s SmartCourt technology automates the collection of several metrics, such as shot speed and location, as well as video replays. At the collegiate level, where scores are inputted, players and coaches can filter reports based on specific scores to check tendencies when up or down in a game or set. (The system was recently approved by the Intercollegiate Tennis Association for also reviewing close calls.) Newer technologies such as Lvision rely on computer vision to compile metrics from any video feed.
Many companies, such as Tennis Analytics, tag matches by hand. Pretorius works with a few elite clients but focuses primarily on junior levels of the sport. His firm charts some 300 hours of competition per week to produce Tennis Intelligence reports that tabulate a portfolio of metrics such as percentage of points won on first or second serve, percentage of points won on an opponent’s second serve, or points won in the first four shots. This data can provide more context in tracking player development and inform training. “That’s the biggest impact that we’re starting to see analytics having, even if it’s simple analytics,” Pretorius says. “Players and coaches have a more focused approach on the practice court.”
NOSTALGIA NEVER DIES: Panini Gives Fans What They Crave With Digital Collectibles
Tennis Australia often sends its players to faraway matches with cameras to record competition when broadcast or streaming options aren’t available. That video is later coded, by hand, with rich detail for future reference. That process can take hours of an analyst’s time, prompting hours of further review by a coach—all for a few minutes that end up being shown to a player. “The problem with tennis is that it’s just all over the world,” McMurtrie says. “You’re not with the coaches as much as you’d like to be, to be able to educate them.”
The same plethora of data that makes it an attractive betting sport also speaks to its potential for advanced statistical analysis. Milan Černý, SAP’s innovation lead for tennis, says some players have “very, very clear patterns, especially at different scores.” He adds, “That’s one of the advantages that you have with our solutions—you can really look into situational analysis and figure out how does a player respond to pressure situations.”
Alison Riske, an American currently ranked No. 34 in the world, is said to have benefited from some insights O’Shannessy provided to her coach and husband before she toppled world No. 1 Ashleigh Barty in Wimbledon’s fourth round. “There are small adjustments, but I think there are big things to know, especially on big points,” Riske says. “To know a player’s tendencies, et cetera, I think is really important.”
Question? Comment? Story idea? Let us know at talkback@sporttechie.com