In a fascinating application of artificial intelligence, Matthew Bakowicz, a professor at American University, and his 32 students have been analyzing data for the 152nd Kentucky Derby. Their goal was not merely to decode betting strategies but to demonstrate how AI can lead to more informed decision-making in various professional fields.
Having previously worked as an operations manager for DraftKings Sportsbook and Racebook, Bakowicz found the intersection of AI and sports betting particularly intriguing. He noted that the upcoming Kentucky Derby provided a prime opportunity to apply this technology, which is often seen as both revolutionary and contentious.
Bakowicz emphasized that the primary aim of the project was to create a classroom experience that mirrored financial modeling, integrating elements of sports analytics. He expressed satisfaction with the results, even as he acknowledged the unpredictability inherent in horse racing.
“Is this going to be perfect? No, it will not,” Bakowicz admitted, citing significant variables such as the race's distance and the unusually large field of competitors. The experiment aimed to foster an understanding of AI's potential rather than to produce a flawless algorithm.
The students tasked the AI with maximizing a hypothetical bankroll of $1,000 across six races at Churchill Downs, focusing particularly on the Derby, which is the 12th race on the card. They have previously conducted similar projects in other sports contexts, such as basketball and the NFL Draft.
Bakowicz explained that the approach taken was conservative, adhering to responsible gambling principles while still aiming for entertainment and potential profitability. The AI was trained to identify high-value betting opportunities.
In its latest iteration, the program analyzed races 8 through 14, allocating half of the $1,000 bankroll to the Derby. According to the AI's predictions, Commandment (6-1 odds) is favored to win, with Silent Tactic (20-1) expected to place and Chief Wallabee (8-1) likely to show.
The betting strategy proposed by the AI includes a $100 wager on Commandment to win and $50 on Commandment to place, alongside other suggested bets on Silent Tactic and Danon Bourbon. However, the unpredictable nature of horse racing already affected some outcomes, as one of the suggested horses was scratched from the race.
The use of AI in horse racing betting has sparked debate, particularly with the rise of Computer-Assisted Wagering systems that not only predict outcomes but also exploit inefficiencies in betting pools.