
Artificial intelligence has certainly influenced pro sports. Most elite sports franchises, such as NFL and NBA teams, will be using the technology as part of their data analysis departments, helping in areas like scouting, performance analysis and injury prevention.
As the technology can help model outcomes for sports, such as predicting a player’s injury record, it also follows that it would help with sports betting. But to what extent is AI being adopted in that area? It is somewhat complicated.
Perhaps the best place to start is by saying that AI, certainly consumer-facing AI, is not ready yet to crack the code of sports predictions. Some believe that it never will, though, for now that’s a moot point.
AI has limits in its usefulness
Nevertheless, it remains the case that feeding an AI model Thursday’s NFL game odds and asking it to come up with some kind of winning formula does not work.
Perhaps specialist AI models fed on massive amounts of data could provide a winning formula, but there are limits to how effective that can be.
One of the issues is a fundamental one: Statistics can only tell you so much about the outcome of a sports game.
The world’s greatest basketball player can stink out the court on any given night; the most dominant soccer team can lose to the worst team in its league; legends like Tom Brady, LeBron James, and Lionel Messi have all had bad games in their careers. It just happens in sports.
So where does AI help? By comparing statistics to odds, we can argue that AI can be beneficial, even if, as mentioned above, stats don’t tell you everything.
By that, we mean AI can be used to analyze comparative value. In a nutshell, it’s a case of AI saying, “the odds say this, but the stats say that.” In another sense, you are looking for AI to pinpoint where odds-setters have erred.
As you can gather, sportsbooks don’t often make mistakes. Moreover, they might be using AI themselves to ensure that their analysis of the market is up to scratch. However, there can be slight discrepancies and anomalies that the AI models may be able to flag.
Good data can help
All of this is data-dependent. Consumer-facing AI models are going to be watching games on your behalf, and then suddenly be able to say, “Hey, the Atlanta Falcons’ Bijan Robinson is much better than the statistics say.”
There are some AI models that can be used to analyze sports games and plays, but these remain on the experimental side. For now, you would be limited to data.
The type and source of data are important. The NFL’s website, for instance, has plenty of data, but the win-loss records of teams, points scored, conceded, touchdowns per player, and so on will probably only get you so far.
There are specialist sports data agencies, but it is their livelihood, so they don’t give it away for free on the live web.
Indeed, one of their most prominent customers is sportsbooks, so the AI could be working off the same sets of data as the people setting the odds.
What should be clear in all of this is that there is no magic wand. In sports betting, AI is useful for data analysis, pinpointing value in the odds. In time, it will get better at that, too.
But note that the betting industry will have to respond to AI, just like every other one. It may impact how odds are set, so that those companies can protect profits.
And, well, if there is a future scenario where everyone can win at sports betting just by asking AI to make picks, there will no longer be a sports betting industry to facilitate it.