1. Business
  2. The future of sport
December 3, 2019updated 04 Dec 2019 2:11pm

Data analytics is rife in tennis, but could AI replace humans entirely?

By Robert Scammell

A robot arm strikes a tennis ball, serving it at 220 miles per hour. From the opposing side of the court, another robot returns the serve and the two machines engage in a 427 stroke rally. An automated umpire decides, instantly, that the ball was out by a hair’s breadth. Silence rings around an empty stadium.

This is the hypothetical, somewhat dystopian image conjured by Chris Brauer, director of innovation at the Institute of Management Studies at Goldsmiths, University of London, at a recent panel discussion about the use of data analytics and technology in tennis.

While steeped in tradition, tennis has embraced technology on multiple fronts: coaching, umpiring and fan experiences.

Since the early 2000s, the Sony-owned Hawk-Eye system has been assisting tennis umpires in making close calls.

At Wimbledon, IBM’s Watson AI analyses fan and player reactions in real-time video footage from matches to create highlight reels just minutes after the end of a match.

Meanwhile, at the ATP Finals in London, similar data analysis is being carried out by digital services and consulting firm Infosys.

The multinational firm has been partnered with ATP for five years, providing features such as its cloud-based platform, which leverages artificial intelligence to analyse millions of data points to gain insights into the game.

This year Infosys launched the ATP Tour app to serve up live scores, stats, news and video in a personalised feed for fans.

Players and coaches can also make use of the Infosys’ Players and Coaches Portal, allowing them to “slice and dice” matches on an iPad with 1,000 data analytics combinations.

Craig O’Shannessy, strategy analyst for the ATP World Tour and a coach for 20 years – including for the likes of Novak Djokovic – says video and data analytics is crucial for giving players an edge.

For example, the data shows that the most important battle for players to win is the first four shots in a rally, which correlates with a 89% win rate for the match.

“It’s about finding out of 100 points, the 10 or 15 that matter the most, and explaining that these are the patterns of play that you want to repeat in these upcoming games to win those matches,” he says.

Could AI replace humans in tennis?

During the discussion, Brauer asked whether the “inevitable conclusion” of technological innovations in tennis was removing humans from the game entirely.

Raghavan Subramanian, associate vice president and head of Infosys Tennis Platform, says it’s a “very philosophical question” and that we can look to the precedent set by other ‘man vs machine’ face-offs.

“In chess, we had [Garry] Kasparov play against the computer. And then of course, we have Jeopardy games,” he says.

“So I think the natural first transition will not be two robots playing against each other, but one robot, possibly playing against the best player today. That’s the first possible bridge before two robots play.”

While AI is already superior to humans at carrying out advanced calculations, robots are still limited when it comes to manoeuvring with the same agility and dexterity as humans.

But as robotics continue to advance, we could see robots capable of running 100 kilometres per hour – more than double the top speed of Usain Bolt.

Such a robot would be capable of retrieving every ball. “That’s a very unfair advantage,” says Subramanian.

ATP chair umpire and manager Ali Nili says that while there could one day be robot players adjudicated by robot umpires, it would be an entirely different sport such as a “robot world cup”.

“At ATP, we’re most proud of our athletes,” he says. “It’s our athletes which makes the tennis exciting. It’s how fast they are, how strong they are being. As human beings, we compare them to us and we’re fascinated by the things that they’re able to do.

“They’re the number one attraction for anyone who comes in, watches tennis, and everything else is secondary, you know, all the data and everything else, because we try to make our athletes more appealing.”

What about coaching a robot tennis player? O’Shannessy is unequivocal:  “No chance, that’s not going to happen.”

Read more: How IBM and Wimbledon are combining technology and tennis