Who would have predicted that Eastern Europe would have taken this year’s Wimbledon by storm? Serbia’s Novak Djokovic and the Czech Republic’s Petra Kvitova emerged victorious from the men’s and women’s finals.
IBM’s SPSS predictive analytics software was not in the game of predicting winners at this year’s tournament. But, in the 21st year of its association with the 125-year-old tennis championship in South West London, and in its own centenary year, IBM did deploy a mixture of predictive analytics software, data visualisation and kinetic tracking.
Mark Crawley, an application architect at IBM engaged on the Wimbledon project, explained that SPSS was used to surface “what people consistently did when they won.”
Five years of historical Grand Slam data, amounting to almost 40 million competitive points, were analysed to identify the key aspects of players’ performance when they played at their best. For each match at Wimbledon, IBM’s PointStream visualisation tool also incorporated specific match history data based on previous tournament meetings. PointStream identified the top three “Keys to the Match” that each player needed to achieve to be at the top of her game, displaying those, along with other data, on a scoreboard online. A “key” could be the percentage of first serves won or the number of points won in fewer than three rallies.
“The benefit [of this technology],” Crawley said, “is you can spot trends across historical data, and from that you can draw conclusions about future trends. Instead of having the data just sit there doing nothing, it can be used. SPSS is very powerful at pulling out what is important and bubbling up the historical data.”
For the first time in a Grand Slam tournament, IBM piloted Player movement tracking on Court 18, using its Second Sight technology. This has been used to track missiles, but at Wimbledon it measured player speed, distance and stamina. It is similar to the Prozone technology that football clubs use to track players.
“This added to the raw tennis data, giving an extra set of statistics around movement,” Crawley said. The firm will work with Wimbledon to see what it wants to do next with player tracking, analytics and data visualisation.
“The PointStream technology has a lot of mileage, in our view. We could develop the keys into a richer application that broadcasters could use.” Coaches could also use the whole analytics, tracking and visualisation package, Crawley said. “We got a lot of interest from coaches and players for the Second Sight technology,” he confirmed. The tracking can disclose information about how far and fast players are running, suggesting changes to diet or training regime to increase stamina. On the other hand, he conceded, in tennis, “running fastest or furthest might not be an advantage.”
For the full story on the IT powering Wimbledon's tennis tournament, see our sister publication, ComputerWeekly.com.