Sports bettors and bookmakers are always looking for an edge. Sports betting is a massive industry that has grown exponentially over the last couple of years with legal expansion in US markets. In the United States, growth has already exceeded expectations. Here are sports betting and AI — the ultimate gamble.
According to ESPN Chalk, since the Supreme Court passed the decision to let states individually handle sports betting, over 20 billion dollars has moved through US sportsbooks.
With that much money on the line, the sportsbooks want the most accurate information they can get so they can set the closest lines possible. Sports betters wish to be able to have enough information to be able to spot advantageous lines. An advantageous line is a spot where the sportsbooks may have missed something, or the public or sharp money pushed the line into a now favorable position.
With machine learning and artificial intelligence, the next step in information aggregation is obvious: AI. In short, it’s all about stats.
Betting on sports isn’t all just you against the bookies. Football Index, which uses StatsPerform A.I., describes itself as a football stock market. They compile market-leading data to help people around the world buy and sell shares of footballers. Their idea of using AI-powered, stats-driven data gives people a much better look at players and whether they want to buy into them or not.
Before, all the average person had to go on was media mentions and what they saw in the games they could watch. But most of us don’t have the time to track every stat and do a comparative situational analysis of each, along with historical data and how it compares not only to different situations but different players.
Football Index also uses StatsPerform’s AI to power their matchday rankings and in-play dividends. So, Football Index has created an interesting new market-take on predictive sports analytics.
STATS VQ is a newer machine learning system that sportsbooks use to create player props lines based on historical data of individual performance.
“STATS collects the richest sports data globally and transforms it through revolutionary AI to unlock the past, present, and future of everything sport. The pioneer of live sports data, STATS, continues to speed innovation in the industry with AutoSTATS, the first-ever AI-powered technology to collect comprehensive sports data from any television broadcast.” AP News.
It isn’t just Vegas books and offshore outlets that utilize artificial intelligence for sports betting purposes.
Savvy handicappers have been aggregating data to create models for years and years. Now, with the ease of access to powerful computer programs, similar data can be run in simulations thousands of times in a short period to figure out the most likely outcomes.
Sites like SportsLine and Oddshark tout their computer prediction models. The picks to win on the money lines are usually pretty accurate, but it gets a little tougher when it comes to point spreads.
For example, in Oddsshark’s last 100 NFL computer predictions, they have won 62 and lost 38, but with “favorite” prices, this doesn’t mean there has been a profit. ATS, Oddshark’s AI is only 49-48-3. So, if you consider a -110 average price, their computer models are not profitable.
In my opinion, computer picks are quite useful as a barometer.
So, computer picks are quite useful as a barometer — but without the human touch, they are almost worthless. Things like momentum and personal issues are hard to quantify.
Statistical analysis is nothing new. It’s what we handicappers and sportsbook operators have always used to build prediction models. But AI takes data aggregation to a new level. NBA Teams started using AI a while back to learn more about their players’ habits.
Speaking of players, here’s where it gets interesting and perhaps a little scary. Top sportsbooks use AI not only for running simulations based on historical data to set betting lines but to profile their players as well.
With proprietary algorithms, sportsbooks can track and monitor their players’ actions and go back through the historical data to create models on how their players (us) will react to the lines they set. They’ll pay particular attention to the time a wager is made, how line movements affect wagers, who is making the wagers (what type of player: bankroll, betting history, etc.) to build profiles.
With this information, they can adjust from traditional betting lines that are designed to land right in the middle to more profitable betting lines that are set based on a desired and expected reaction from bettors. So, it’s important to keep this in mind as you are looking over the betting lines. Has a line been inflated or deflated to get you to react a certain way?
Having your own power ratings even more important.
You should already know what the point spread should be on a given game before even checking the opening lines. This way, the sportsbooks can’t manipulate your actions as easily. Because you can choose to opt-out of any betting lines that don’t hold a clear advantage.
A.I. is changing the sports betting industry in many ways. But it’s all just data when it comes down to it, and sports analytics are nothing new. There are so many intangibles that a computer cannot pick up on that a human touch will always be needed, at least for the foreseeable future of sports predictions. So, fear not.
AI in the sports world.
It should also be noted that what is commonly referred to as AI in the sports world isn’t really artificial intelligence, at least not in most cases. It’s more of a marketing ploy. For it to be actual AI, the program would have to be able to learn and adapt without interference from the programmers. Essentially it would have to be able to write its own programming as it evolves. Simply put, aggregation of data and simulations does not equal the definition of artificial intelligence.
If you really want to beat the bookmakers, people should focus on algorithmic models that focus on finding the consistent patterns of when and where the bookmakers tend to set the betting lines inconsistently. Instead of focusing on a specific sport or data related to certain teams, focus on specific linemakers and exactly when they make errors.
Remember it is not the public, sport, game, or other teams you’re betting against; it’s the bookmaker and the line they set. So, if we can effectively profile them, the way they try to profile us, we can once again gain an advantage as bettors.