Artificial Intelligence and predictive analytics in sports: a blessing for some, a nightmare for others
Why didn’t he play more defense during the final minutes? Why did he let him play instead of keeping him on the bench? How can teams still win the championship without having a sugar daddy from Qatar as investor?
If there’s one thing sport fans love to debate about, it’s the coaching strategy of their team. In sports there is a large opportunity to change how athletes and teams push their limits. Imagine the possible gains by making game-changing decisions based on data instead of gut feelings. And what if athletes could extend and improve their career by predicting injuries and preventing them by a small tweak in their training programs? The answer to all of this is sports analytics. Artificial Intelligence (AI) and predictive analytics tools in other sectors such as finance and retail are already well-known. Think about how chatbots are becoming our own personal finance assistants. Or what about how AI is rewriting the retail playbook by learning about the customers and knowing what they want when they want it, even before they know it themselves.
But how exactly are AI and predictive analytics influencing sports?
One of the most recent applications is the Sports Performance Platform of Microsoft. It’s a tool that includes machine learning and AI in its algorithms to come up with data-driven decisions for athletes and teams. In short, it provides solutions for the locker room, performance lab and the side-line. The platform contains algorithms to prevent injuries, make game-changing decisions based on each athlete’s game-day availability and alter the training scheme to keep athletes performing. For instance, the platform predicts the recovery time while taking into account factors like distance sprinted, temperature,…
A small Danish football club, FC Midtjylland, went from being close to bankruptcy to winning their first championship title. Were they acquired by a sugar daddy from Qatar who invested millions to buy new players? No, quite the opposite. Instead, they switched steering angle and took the analytical route. Each player is now analyzed twice a month and receives an individual training plan. Analytical experts provide the coach with insights during half-time to alter the game plan based on in-game statistics. The management now also uses insights from analytical models that suggest new players. They thus completely changed their approach from using their heart to using their brain.
You may get the idea that only football is influenced, however, many other sports are joining this analytical revolution. In tennis, IBM introduced their tool ‘SlamTracker’ to predict the winner of a match. Using historical data, Slamtracker determines a player’s pattern of play, propensity of using their forehand, willingness to volley,… This is then compared with live game footage to predict the winner. Next to this, IBM used this year’s Wimbledon tournament to introduce their all-knowing AI chatbot ‘Ask Fred’ to improve the customer experience by answering questions on: accessibility, tickets, facilities, what to do, food and drinks, travel and transport,… Fans not at the venue benefit from an AI system that selects spectacular moments to show based on players’ facial expressions and crowd noise. Moreover, with the help of recognition software and forecasts on how close a match will be, fans are offered suggestions for future matches.
In the NBA, the most popular basketball competition, they also embraced AI and predictive analytics in their coaching strategies. For example, models are able to predict whether a player that is in a certain position will either try to score or pass (and to whom). Another leading example is the Formula 1 where they use AI and analytics on real-time data to gain every drop of performance improvement and make race-changing decisions.
Sports and analytics: a match made in heaven for all?
Of course not, for the sports betting sector it is seen as a double-edged sword. Although AI models are used to set the rates at which people can bet on sport events, they promise to be a bookmaker’s worst nightmare as they can predict game results and enhance the odds for people placing the bet. For example, OhMyBet is an AI tool using machine learning algorithms that selects upcoming events to predict the most probable winner. Only a few are picked as the ones with higher risks are sorted out. OhMyBet currently has an accuracy of 85%! A real nightmare for betting agencies, right?
Despite these examples, many sport teams have not (yet) made the switch in their mindset and are not using data the way they should. All these applications have proven the influence of AI and analytics in every aspect of sports, going from performance to injury prevention, fan experiences and betting results. AI and analytics in sports can be a match made in heaven for players, teams and supporters but might be a nightmare for bookies. One thing is for sure: sport has to stay partially emotional due to the element of surprise and an inevitable dose of unpredictability. Now tell me, will you still criticize coaching decisions the same way you did before?