The revolution will be analysed

A data revolution is coming to tennis to offer players, coaches and fans a deeper understaing of the game within the game.

Jo-Wilfried Tsonga, Practice, 9 January 2017

How are tennis matches won and lost?

Simple questions rarely come with simple answers – particularly in a sport where you can win more points but lose the match, or hit more winners yet leave defeated.

For most of its history, tennis players, coaches and fans have relied on conventional wisdom to answer this question, generally failing to explain the complexity of victory – shortcomings placed centre-stage by a 2012 ESPN program that, in terms of its use of analytics, ranked tennis as the second-least intelligent professional sport on the planet.

But a revolution is afoot – a data revolution, offering players, coaches and fans alike a deeper understanding of the game within the game: the shots, strategies and patterns of play that determine which player wins the point long before the ball hits the back fence.

For Dr Machar Reid, head of Tennis Australia’s Game Insight Group (GIG), the rise of analytics in tennis is long overdue. Since its inception in 2008, the team has built up an impressive body of research and produced scores of scientific papers, making GIG the world’s leading authority on tennis science and helping the sport make up for lost time.

“For the last 40 years, tennis matches have been described in a particular way – first serve percentage, second serve percentage, unforced errors, forced errors, etc.,” says Reid, who worked with Greg Rusedski and Li Na before pursuing a doctorate that focused on the science behind the sport.

“It represented a logical starting point, but we haven't really progressed since.

“If you were to compare that to our contemporaries – big professional sports like baseball and basketball – they have really shifted the dial. Their vocabulary has evolved, and their understanding of the game has probably left tennis a couple of decades behind.”

Reid and his team are at the forefront of the movement to bring tennis metrics to the mainstream. Based at Melbourne Park’s new Tennis HQ, nestled between Rod Laver Arena and the MCG, the GIG team utilise the latest analytics software and hardware and the treasure trove of data produced by the Australian Open to rethink all aspects of the sport, from improvements to the tournament, the prize money structure of professional tennis, and how players select the right racket and strings, to deep-dive tactical analyses of the game’s elite.

In 2016, GIG published their investigation into three years’ worth of Australian Open HawkEye data to identify the critical shots and patterns of play adopted by the likes of Novak Djokovic, Roger Federer, Rafael Nadal, and Andy Murray. Reid and his collaborators at Disney, the AIS, and QUT, were awarded the Grand Prize at the 2016 MIT Sloan Sports Analytics Conference. “It showed that we could compete with the world's most popular sports in the analytics space,” he added.

This team developed a more sophisticated method of interpreting playing style, defining 50 shots and combinations to reach a more nuanced consideration of how the baseline game has evolved – a far cry from prevailing game style descriptions like serve-volleyer, aggressive baseliner, counter-puncher, and the semantically redundant all-court player.

The report’s findings make for interesting reading. Among the ‘Big Four’, Federer and Djokovic’s games were most closely related, while Murray’s game was more similar to that of David Ferrer, Jo-Wilfried Tsonga, and Lleyton Hewitt; perhaps less surprisingly, Rafael Nadal’s game made him an outlier.

The results also illustrate Reid’s belief that the current approach to describing player performance is too rudimentary. In his eyes, the heart of a point is a misunderstood, undervalued commodity. “Serve and return aside, the rally is what tends to put players into winning or losing positions, and becomes more informative in understanding how matches are won and lost.

“The default position for the sport has typically been understanding what the start of a point looks like, and what the end looks like,” Reid explains. “It’s akin to only taking note of the centre bounce and a shot on goal in Australian Rules football – in effect, discounting how the team actually got the ball into position to take that shot.”

GIG’s work will continue to bridge that knowledge gap in 2017, with the recent recruitment of Dr Stephanie Kovalchik, a leading American data scientist and author of the tennis blog on-the-T, substantially adding to the team’s firepower. This year’s Australian Open coverage will include point-by-point Win Predictions, real-time assessments of a player’s shifting probability of emerging victorious. Based on Elo ratings employed by the likes of ESPN’s FiveThirtyEight for American sports, the metric updates the pre-match analysis as the match progresses to give more accurate forecasts of the likely winner. Every tennis player has a chance until the last point is point is won or lost – now we’ll know exactly what their chances are, and which points proved critical in deciding the outcome.

Another innovation set to debut at this year’s Australian Open assesses a player’s performance in pressure situations. GIG’s Clutch Index is a next-level assessment of match statistics designed to demonstrate who played better when it mattered most. “Outside of break point conversion, traditional tennis statistics tend to identify and treat every point the same way,” Kovalchik explains. “This fails to provide a complete view of a player's performance under pressure. Hitting an ace while 30-40 down at 4-4 in the deciding set has a very different meaning to hitting an ace at 40-0 in the first game of the match – in the former, I’m under pressure and showing real clutch performance, while in the latter, I'm not.”

While Reid laments the lag in tennis analytics, he also recognises the opportunity. The GIG team work with players and coaches to analyse individual matchplay data, to make objective decisions about changing their equipment, and to understand more general game trends. “Credit to Tennis Australia – they’ve invested in something that’s not really been explored before in tennis,” he said. “It's not a risk-free proposition but it's progressive and innovative. The players and coaches have hopefully benefitted at various times, and the intent is for others within the sport to do likewise.”

Now to spread the word beyond Melbourne Park – something Reid hopes will lead to the release of more data from tournaments worldwide, giving GIG the chance to broaden their investigations and change the way we think about tennis.

“There’s an opportunity, as other sports have done, to capitalise on the step change we're experiencing in data and technology. This journey is in its absolute infancy for us as a sport, but it will help the game – and more importantly the players – in more ways than many of us can imagine is possible right now.”

Simple questions rarely come with simple answers. But for Reid, and for GIG, the beauty of tennis lies in the nuance.

° C | ° F