Right now, in the tech industry, there are two words on everyone’s lips – artificial intelligence. The success of projects like DALLE-2, MidJourney, and ChatGPT has sent reverberations around the globe, with some believing we see a wave of technology equal to – or perhaps even greater than – the roll-out of the internet in the 1990s. There is much talk of job roles disappearing, including the roles of marketers, writers, artists, and even lawyers. While some have pointed to a dystopian future, others say we should harness technology as a tool to help us, not as a replacement for our own jobs and creative roles.
There are, of course, many uses for AI, and one of the most interesting areas is in the modeling of predictions, including sports predictions. We have already seen this with perhaps the world’s most famous supercomputer, IBM’s WATSON, which has been making NFL fantasy football predictions in partnership with ESPN. Fantasy sports isn’t quite the same as sports betting, but there are enough similarities in both the tech and the application. To be fair, WATSON has been doing okay, although it certainly isn’t delivering results like some sports predicting oracle. Otherwise, everyone participating in the multi-billion-dollar fantasy sports industry would be using it.
Premier League predictions were off the mark
So far, however, AIs used for sports betting have had mixed results. To give you an example, in the summer of 2022, an AI supercomputer predicted the outcome of the 2022/23 Premier League this season. It predicted that, like last season, Manchester City and Liverpool would be the top two (with City triumphing), followed by Chelsea and Spurs. Arsenal, who currently top the real table by five points, and who are clear favorites in the Premier League betting for the title, were 5th. Manchester United, who have been largely excellent this season and are comfortably in the Top 4, was placed in 6th by the AI.
As such, the AI used in this case was wide off the mark. Yes, there is still plenty of the season to go, but teams like Liverpool and Chelsea have been terrible, and they will struggle to make it to the Top 4. Moreover, everyone who follows the Premier League knew that Newcastle – now one of the world’s richest teams – would do well in 2022/23, given the Magpies’ transfer business, so why did the AI model put them in 14th behind teams like Leicester and West Ham, both of whom have relegation troubles to worry about?
Unstructured data remains a problem for AI
You might say that the problem lies in the way that AI analyzes data. Computers are good at crunching numbers in all types of sports when it comes to something called structured data. That relates to tangible stuff, like previous results, head-to-heads, and so on. The programs can eat up millions of points of data almost instantly, so it has a supreme advantage over a ‘human’ punter. However, AI is mostly inferior to analyzing unstructured data, i.e., the stuff we can’t quantify with numbers. Newcastle’s form is a good example of that. The club got new owners and a new manager and had some smart recruitment (without going over the top in transfer spending), but there was also an air of positivity around the club. Fans can see it – can a computer?
But here’s the rub: Proponents of AI believe it will become infinitely smarter in dealing with unstructured data. And that could be a literal game-changer. Anyone can use AI like ChatGPT right now, so could bettors conceivably run their betting picks past an AI before placing a wager? Could they use it to beat the bookmakers? Would sportsbook companies have to react if we all suddenly become very adept at betting?
These questions remain to be answered. But the explosion of AI technology at the beginning of 2023 has made many proponents believe that we will see disruptions to many industries. Nobody, not even future supercomputers, will be able to predict everything 100% correctly – that’s one of the reasons, so many of us love sports. But computers could soon be much better than we humans at predicting, and it’s going to be interesting to see its impact on the industry.