Watson serves up AI at Wimbledon 2017

By | 4 minute read | July 3, 2017

Summer in the UK means two things: mandatory shorts (regardless of temperature) and Wimbledon. For two weeks, the All England Lawn Tennis Club (to give it its full title) buzzes with up to 39,000 on-the-ground punters and countless fans catching a slice of the action online. This year, IBM Watson serves up four digital and AI innovations to help tennis fans get right to the heart of the action.

A brief history of IBM and Wimbledon

IBM and Wimbledon are old bed fellows – for almost 30 years we’ve been Wimbledon’s Official Technology Partner, overseeing and driving changing technology to create the best experience for athletes and fans alike. Stopwatches and electronic nets in the ‘90s gave way to the Hawkeye system, and The Player Report (which started off in 1994 as a basic statistical analysis of service direction and point-by-point match history) got a serious upgrade thanks to big data and real-time analytics to digitize player tactics. Now, thanks to the abundance of channels, devices and content available to viewers, the technological spotlight is on engaging fans with a super-personalized experience.

This year features four new (or updated) cognitive innovations: IBM SlamTracker, with real-time scores and insights; ‘Ask Fred’, a Watson-powered cognitive assistant; automated video highlights powered by AI, and a ‘What Makes Great’ solution, analyzing sports coverage across 44 years to give a new perspective on athletic prowess.

Watson the tennis expert: What Makes Great

One of the most exciting things about Wimbledon is the chance to see greatness in action. Great stars in the making. Great comebacks from near defeat. Great passion, commitment and coolness in the face of overwhelming pressure.

But what makes a ‘great’ tennis player? Is there any objective way to tell, and can big data analysis help us to understand the difference between success and failure? In an effort to answer the greatness question, Watson, IBM’s super computer, has been gorging on enormous unstructured data sets, including 53,713,514 tennis data points, Wimbledon annuals, social media commentary and interviews (amounting to 11,208,192 words), and 6,349 Telegraph articles.

The resulting insight is what informs ‘What Makes Great’ – a new feature offering unsurpassed perspective on what makes a great tennis player. ‘What Makes Great’ uses 44 years of sports coverage to examine everything from key physical performance measures to mental and emotional attributes such as passion and response to pressure.

What Watson is offering is a unique synergy between computer and human expert. While computers historically have been good at number crunching and little else, Watson has made the leap from analysing purely structured data sets (numbers) to unstructured data sets (words). The result is a third perspective that intersects the subjective human and the passionless digital responses to yield fresh insight.

SlamTracker and big data

IBM SlamTracker with Cognitive Keys to the Match is the viewer’s ticket to understanding styles of play and match metrics, and pinpointing the crucial tipping point between winning and losing. This cross-platform app mines real-time data from courtside statisticians, chair umpires, radar guns, ball position and player location. It’s had a ton of updates since last year to include more detailed metrics, like pace of play and baseline proximity, which are set against a backdrop of information taken from 8 years of Grand Slam tennis data – around 41 million data points all told. Together, this information will help fans identify winning patterns and styles, spot trends in a player’s approach to the game, and understand their performance in the light of past experience.

At the match outset, fans can launch the SlamTracker to check out which tactics to look out for from two players going head to head, and follow their progress against these keys in real time. Plus, there are added insights for ‘pressure situations’, which will display historical performance data for a player who has found themselves in a similar situation in previous matches.

‘Ask Fred’ – a new hat for Watson’s AI concierge

Visitors to Wimbledon will know that the grounds themselves are a bit like a small city, and that finding your way as a newbie can take some time. For those on the ground, there’s ‘Ask Fred’ – a Watson-enabled bot (named after Fred Perry) whose job is to help visitors familiarize themselves with the layout, locate dining options and facilities, and generally get the most out of their experience. ‘Ask Fred’ uses a Natural Language Interface (NLI) to understand questions such as ‘where can I buy strawberries’, or ‘where are the closest toilets’, making it simple to use.

Watson turns director

In case Watson didn’t have enough to do, the AI system will be donning its directorial hat and picking out key moments of a match based on analysis of player movement, match data and crowd noise. This automated approach will help editorial teams put together 90 second highlights packages from the 6 main court shows at speed, so that video content can be available minutes after the end of a match.

Find out more and join the #WhatMakesGreat debate

We’re on a quest to discover what makes a great tennis player. To find out more or to have your say in the debate, use the tag #WhatMakesGreat. Interested to know more about Watson and IBM’s cognitive solutions? Take a look at our website.