Zion Market Research reported in 2019 that the global sports betting market was valued “around $104.31 billion” in 2017 and is expected to reach $155.49 billion “by 2024, growing at a healthy CAGR (compound annual growth rate) of 8.83% between 2018 to 2024.”
With that much money at stake, gamblers will obviously turn over every stone for an edge. While this could materialize as studying expert opinions or team tendencies and matchups, another option is to utilize machine learning to predict outcomes for you.
The problem? Bookmakers have access to the same technology.
Richard Bartels, a data scientist with Vantage AI, explained in a recent article with KDNuggets.com:
“Both the bettor and the bookmaker can be equally skilled in predicting the outcome of a match, however the bookmaker sets the rules for the bet and thereby guarantee themselves a profit in the long run,” Bartels writes. “The way they do this is by controlling what is called the payout.”
But bookmakers can make mistakes and/or fail to accurately assess the winning probabilities of a given game or match.
That’s where machine learning — using a custom loss function — can provide an advantage. This model can evaluate the odds using both current and historical data, and identify areas where the bookmaker failed to set the proper odds. So rather than trying to predict the exact winner, machine learning will instead identify “high risk” areas where a bookmaker’s application was less than perfect. It’s no guarantee — but it has been shown to potentially help a bettor gain an advantage over the current system.
The Flip Side To Machine Learning And Sports Betting
Simplebet, a betting startup that recently raised $11 million from investors such as Philadelphia 76ers co-owner David Blitzer and tennis legend Andre Agassi, is ready to launch, and it’s relying on machine learning to set it apart.
Instead of focusing on winners of games, fights, or matches, Simplebet is bringing live betting to a new, more granular level. Users can bet on games, but they can also bet on events within the games. Will a pitcher throw a strike on his next pitch? Will Lebron James make his free throws?
Such a quick-hitting concept can’t be executed using human labor alone, so Simplebet turned to machine learning to quickly assess and analyze, then present new scenarios with fresh odds to keep viewers engaged.
“The average NFL fan watches 52 minutes of a game. If you can increase engagement from 52 minutes to 57 minutes, that has enormous value when the NFL goes out to do rights deals,” Chris Bevilacqua, a media consultant and one of Simplebet’s three co-founders, recently told Sportico in an interview. “More people are watching for longer. That’s their currency.”
The concept isn’t fully refined and complete, but there’s no doubt the early stages were made possible by machine learning.
If you’re in the sports betting sector and you need to find engineers with a background in machine learning, artificial intelligence, or data science, contact us at firstname.lastname@example.org.