Just last weekend, fans around the world were on the edge of their seats watching the nail-biting finale to the 2019 Wimbledon Tennis fortnight.
Millions of viewers tuned in to watch as reigning champion Novak Djokovic won against Roger Federer in an epic near-five hour match, but few may have stopped to consider the technology behind the action on Centre Court.
A thirty year partnership
The force behind everything from IT operations during the tournament to fan-facing digital offerings such as apps and websites, IBM has been a feature of Wimbledon for the past thirty years as the official technology partner.
In that time, tennis as a sport has gone from being steeped in tradition to embracing the benefits cutting-edge technology can offer, including artificial intelligence, as the sport continues to harness the benefits of big data for players, coaches and fans.
Sam Seddon IBM Wimbledon Client Executive, explains that for Wimbledon, its reputation for high quality extends to the digital world:
“Wimbledon describe themselves quite nicely as like putting on a wedding. So you want to have the most fantastic wedding in terms of the details and everyone having a wonderful day. But they invite half a million people and then they stream it to 20 million over 13 days and they need 24/7 security. So what we’re doing here is providing the technical foundation for that to happen.”
Data is key
The foundation of this is data. According to IBM, 4.5 million tennis data points are generated each year at Wimbledon, and since 1990 the company has collected, catalogued and analysed over 62.8 million of them. Seddon explains that IBM processes many different forms of data during the tournament with the aim of generating greater insights:
“We capture various types of data sources. So scores and statistics, we also integrate all of the player position data into our feed…and then we’ve also got data in the form of photography, data in the form of video, and all of that we have to make available around the world.
This quantity of data drives many aspects of IBM’s role in the tournament. IBM has been responsible for the live scoring at Wimbledon since the early 2000s, broadcasting real-time data not just in the grounds of the tournament but around the world. However, this has evolved in recent years to bring fans increasingly individualised services through Wimbledon.com.
Seddon explains that personalisation is at the heart of this:
“They’ve got a lot of different users of their website, from fans to bencholders, to members, to players, and they want to give them all the same high-quality experience. They’ve got sign on for the first time on Wimbledon.com for the this year. So from a fan’s perspective you can have a personalised order of play, and you can go up onto Henman Hill and vote for what you want to see on the livescreen TV. So there’s an element of give-back in return for the data….all of that is looking at how we can deliver a more personal experience.”
This year, with the aim of better connecting with fans around the world, IBM and Wimbledon have also developed a web app to provide a service for Wimbledon audiences in territories with lower bandwidth and less developed mobile hardware, such as the over 900 million fans in India.
According to Seddon, the scale of the operations increased by 55,000% during the tournament, meaning an “elastic and flexible” set of cloud capabilities is vital:
“We have something called hybrid cloud, so there are components that require a low network latency, such as scores which are displayed around the grounds and also web publishing…and then from a public-facing perspective we’ve got our public cloud, so what we’re doing there is all the web serving. We’re pushing everything out to the fans, and that’s also where we access our own Watson AI capabilities using our public cloud. And it’s all managed through our private cloud, so that’s where we’ve got our security elements.”
As with any high-profile event, security is key, with this risk factor extending to cybersecurity. Seddon believes that a high standard of cybersecurity has now become a key part of the tournament’s brand:
“We can see all of the security events coming into all of our data centres…and we can monitor all of the security elements of the infrastructure. We had around 200 million security events during the championships so we can track the source location…We have to make it all very secure. Wimbledon as business is driven through brand quality, that’s how their business model works. They’re a premium event…so digital cybersecurity is key for their brand.”
AI offers increased personalisation
Known primarily for its AI expertise, IBM has been harnessing the technology to not only offer detailed insights through match analysis, but also in how the tournament attracts and communicates the half a million fans that pass through the gates, and the many more that interact online. In 2018, IBM and Wimbledon launched a Facebook messenger chatbot for the tournament to relay information to fans.
One of the more unusual aspects of IBM’s deployment of AI at the tournament is its use of visual recognition API. IBM Watson is able to analyse real-time video footage of a match and assess excitement levels from factors such as how loudly a crowd cheers or the expressions made by a player. With multiple matches taking place everyday, Seddon explains that this streamlines the process of delivering highlights to fans:
“What this is doing is ranking every single point on ten courts…and every point is ranked for excitement level, so how much the crowd are cheering, how excited and animated they are, how much the players are gesticulating. And then match data, so is it a critical moment in the match? And then we can rank each point…the digital editors can then look at this and say ‘I’m interested in seeing the most interesting break points’ and they can use this to find different things to create content around… The digital editors that would have done this previously can now focus on other content that they want to give to the fans.”
This means that highlights reels can be produced just two minutes after the end of a match.
For this year’s tournament, IBM also addressed the issue of AI bias. Using Watson Open Scale, the AI system can now recognise levels of noise and excitement levels of players, allowing it to remove bias when searching for highlights from players with a particularly vocal fanbase or those who are particularly animated on court.
However, personalisation is not just reserved for spectators. IBM’s AI tools enable players and coaches to unlock the insights from data. The numerous data points generated by a tennis match, such as percentage of aces, number of break points, and even the effeciveness of particular positions on the court or playing styles, can be used to spot trends that even the most experienced coach might miss, and how this can then be passed onto players at the top of their game.
Seddon explains that players are able to access this information soon after a match, and can be an important coaching tool:
“Players get a full match report broken down by statistics and strategy types. They’re able to get a match video with match statistics within it, so if they want to they can just jump to the break points etc.
“On Wimbledon.com there’s something called Keys to the Match and what that is is AI analysis of grand slam data. We use machine learning against all of those data sets so that we can identify trends for players, in particular head-to-heads…You can look at patterns, styles of play and identify key focus areas. If you use these particular tactic areas, you’re more likely to win. You can see that tracking in real-time as the match is being completed.”
But what is next for the historic partnership combining technology and tradition? Seddon says that a focus on personalisation is set to continue for the years to come:
“We’ve got a few things underway. The digital convergence element will continue…personalisation and speed of video content creation. We’re looking at whether we can apply our AI highlights to the huge archive of video footage…that element of personalisation again.”