Racing with Intelligence
by Trey Nosrac
He began to lecture me the second I stepped into his kitchen.
“Artificial intelligence will destroy pari-mutuel wagering. The average handicapper, even an excellent handicapper, cannot compete with complex algorithms.”
“Destruction? Artificial intelligence? Algorithms? Dude, I haven’t even had my morning coffee yet.”
I headed towards the Keurig and slid my trusty”The Man, The Myth, The Legend “coffee mug under the spout. When my pod of Italian Roast clicked into place, I asked, “Now what’s this Silicon Valley BS?”
He continued to speak as if he were conducting a morning seminar,”Not many people understand that the ramifications of artificial intelligence are both amazing and dangerous. One of the dangers is that people too many people believe they understand it.”
“Alright, alright, there’s going to be a war between artificial intelligence and natural stupidity. I already have my NS Tee shirt. Which side are you joining? And why did you drag in horse racing?”
“Last night, I did some research. AI is already involved in horse racing more than most people realize.”
“Everyone knows the days of a guy with a pencil in his fedora hat are gone.”
“Trey, they have been gone for 30 years. Your game needs some serious renovation. I have a concept that could give your game more time.”
“I’m listening.”
“Horse racing should turn the tables. Racetrack management should use artificial intelligence to their advantage and squeeze the juice out of wagering data preemptively. AI should design the races to make every race uber-competitive. I can think of several ways to create equitable races, assign posts, drivers, stagger starts, etc.”
The Keurig gave a final hiss. I grabbed my mug and took a sip. “So, you are suggesting every race be a blanket finish, no long odds, no short odds. Just photo finishes every race?”
“Yes, competitive races all the time would be an improved entertainment product.”
“But what about gambling? How does your reclassification idea stop the artificial intelligence floodgates?”
“Nothing will stop artificial intelligence, but horse races created by it will minimize the wagering on AI.”
I gave him a blank stare, so he continued, “Scientifically presenting races that move from obvious data-based selections to races determined by less predictable random events is the way to go for your sport. Gambling against data machines makes any handicapper toast.”
I tilted my head and tried to wrap my mind around his madness, “But there will always be a winner, always be a second place, always a third-place finisher.”
“Yes. Look, no system can call every race correctly every time. There will always be variables like human error, horses throwing a shoe, a tiny misstep, or a horse just having a bad day. Gamblers move from math against a machine, where they will lose, to races determined by data that is much less measurable, especially drivers.”
He continued, “Using artificial intelligence to design races will make every race more competitive. A race can have 500,000 data points available to predict the results, but no machine has access to a driver’s mind, will know if a horse has a tiny blister under his shoe, or when a goose will run onto the track from the infield and spook a horse.”
I took another slug of Italian Roast, “What’s the point if a gambler can’t feel superior when his horse comes in?”
He fired back with questions, “What’s the point of playing a game you cannot win? How much fun would it be to sit down and play a game of chess against your computer set to a level where human Grandmasters cannot compete?”
I took another slug and got serious, “This is one of those ideas that sounds good in theory, but I see a dark side. Suppose I own a very nice 3-year-old trotter, one of the best in the state. I have invested a lot of money and time. My horse deserves to win. My horse deserves to crush the field. I don’t want a stakes race manipulated to give others an artificial chance to beat me.”
He nodded, “Great point, I was thinking about that myself. My first thought is to segregate stakes racing. Maybe all two-and three-year-old racing remains exempt from artificially created races. After all, there is already inherent randomness in a two-year-old trotter or any green horse. They still have small databases.”
I picked up the ball, “After stakes racing, horses all have larger databases, and then they shift into your level the field artificial intelligence concept.”
“Yes, or maybe a structure where a horse needs to race a dozen times before becoming eligible for AI-generated racing.”
I took another sip and said, “The idea of AI-generated races would need to make sense for owners. They carry a lot of the load.”
“Of course. The AI constructed races for older horses would need to make financial sense, enough sense that forfeiting their power to choose classifications and drivers would be palatable. Owning a quality horse must remain a real perk. Excellent racehorses will race for vastly improved purses than lower-quality horses and have a chance to win every time they race.” Still excited, he almost yelled, “These details can be solved using AI and algorithms.”
“Hey! I thought I was at war with artificial intelligence.”
“You are, but don’t be stupid about it.”