This year’s US Open tennis tournament is going to be providing fans with statistics so detailed they’ll predict likelihood to win, potential upsets, and which matches will be the most compelling to watch. All of that new data is thanks to IBM Watson machine learning tools trained to analyze every detail, no matter how small, to get a machine’s-eye view of each of the 254 scheduled singles matches.
IBM, which has been the US Open’s digital partner for 30 years, promised new experiences for fans and additional ways to interact with the tournament, all thanks to two new technologies it’s launching for the event: IBM Power Rankings with Watson and Match Insights with Watson, both of which run on IBM Cloud with Red Hat OpenShift.
Power Rankings and Match Insights both use Watson’s AI and natural language processing capabilities to provide deep insights into the tournament, even up to deciding which statistics are relevant or exciting enough to bother showing to audiences watching on ESPN or via the United States Tennis Association’s daily “The Changeover” show.
In particular, Power Rankings is described by IBM as “AI-powered daily rankings of player momentum that will be used at the US Open for the first time this year.” Typically, IBM said, tennis tour rankings use 52 weeks of historical data to rank players. IBM Power Rankings “takes this further by focusing on a player’s most recent history. It combines advanced statistical analysis, Watson Discovery’s NLP capabilities and IBM Cloud to analyze player performance data, mine media commentary and measure player momentum,” IBM said.
The Power Rankings leaderboard will be posted the Friday before the tournament begins (Aug. 27) and will be updated as the tournament goes on.
Match Insights, first launched at last year’s tournament, is the second tool being used by IBM at this year’s open. Match Insights “uses Watson Discovery to create AI-generated fact sheets ahead of every single match during the tournament,” IBM said. New features introduced this year include the ability to analyze media mentions to extract key insights and what IBM calls “By the Numbers,” by which Watson uses natural language generation to translate statistics into sentences.
Along with the new capabilities added to Power Rankings and Match Insights, three new features were added that may be able to help tennis fans decide which matches to watch. Likelihood to Win is a new confidence value expressed as a percentage; Ones to Watch is a pre-tournament list of players whose IBM Power Ranking is five or more positions higher than their actual tour rank; and Upset Alerts will be applied to matches that IBM identifies as favoring the underdog.
How IBM’s US Open innovations can be more broadly applied
IBM describes the work it’s done for the USTA and US Open as built on the same “hybrid cloud infrastructure that includes IBM Cloud, Watson for Cyber Security, and IBM Watson’s enterprise-grade AI capabilities” IBM uses to digitally transform clients in many different industries.
As one example, IBM cited HSBC bank as a customer using Watson Discovery, the technology behind Match Insights and Power Rankings. HSBC uses Discovery and neural networks to build its EquBot AI investment platform that “ingests and learns from vast amounts of company financial data to uncover new investment opportunities for their first of a kind AiPEX and AiMAX indexes.”
SEE: Digital transformation: A CXO’s guide (free PDF) (TechRepublic)
“The same natural language processing innovations in IBM Watson Discovery that we are applying to the US Open to get the most relevant and recent insights to fans are also being applied by businesses today,” said John Kent, program manager, IBM. “This AI can be used to help employees automate the process of searching through large volumes of complex documents and text, synthesize high volumes of information to make better decisions and take actions more quickly.”
In short, like many other complicated analytics platforms, the software at the US Open that will be on display from Monday, Aug. 30 until Sunday, Sept. 12 could deliver the same level of depth to any organization’s business analytics practice. Tune in to get a firsthand look.