My ChatGPT Horse Betting Experiment at Exeter

Attending the races is a classic Exeter Student experience. It’s a day for thrifted flat caps, dodging interviewers, and hoping to turn a student loan into a small fortune.

Tired of my usual strategy of leaving the track with an empty wallet, I decided to try something new. This time, I brought in a specialist: ChatGPT was going to be my personal bookmaker.

I traded my own questionable gut feelings for algorithmic advice. I fed ChatGPT all the odds, horse names, and gave it a £50 budget to work with. I was hoping it could finally change my luck. Here’s what happened.

Race One The Safe Bet

For the first race, ChatGPT suggested a conservative approach. It advised a £5 win bet on the “strong market favourite,” Hot to Go (5/4), and a £5 each-way bet on an outsider, Grey Lagoon (9/1).

Standing in a crowd of students, I felt a bit more professional than usual. The AI’s logic seemed sound: back the favourite and take a small risk for a potentially bigger payout. I was feeling confident.

That confidence didn't last long.

Hot to Go started strong but faded quickly, finishing in a disappointing eighth place. Grey Lagoon had a similar fate, challenging for a top spot before tiring and finishing fifth, just outside the placement spots.

After just one race, I was £10 down. ChatGPT had stumbled at the first hurdle.

Race Two The Comeback

After a quick trip to the bar to soothe my financial woes, it was time for the next race. Despite my shaking faith, I committed another £10 to the AI's plan.

ChatGPT’s picks were another favourite, Of Corse I Can (3/1), and a solid outsider, Blackacre (7/2).

The race was a nail-biter. Of Corse I Can was in a two-horse battle for the lead from the very beginning. It was neck and neck until the final moments, when my horse pulled ahead to become the clear winner.

A £20 return landed in my hands, covering my earlier losses. Blackacre finished in a forgettable fourth place, but it didn't matter. I had broken even. ChatGPT had redeemed itself, and the comeback was officially on.

Race Three Money Down the Drain

The excitement from the second race was short-lived. ChatGPT quickly returned to its losing form. My hopes were now on a £10 each-way bet on Bluegrass (9/2) and a £5 win on Courageous Strike (3/1).

It was a total disaster. Bluegrass came in at a dismal seventh, while Courageous Strike managed a slightly better but equally useless fifth. Neither horse was ever a real contender.

ChatGPT had put me back in the red, and my pint suddenly tasted a lot more bitter.

Race Four Out of the Running

Race four presented a new problem. Diplomatie, ChatGPT's top pick, was a non-runner. With no time and poor phone signal to consult my digital bookie, I had to rely on my own intuition.

Trying to win back some of my losses, I placed a £5 win bet on the favourite, Busy Being Busy (7/4). The result was heartbreaking. In a photo finish, my horse lost by just a few inches.

I was now £20 down and left wondering what could have been if ChatGPT’s original pick had run.

Race Five The Outside Chance

Facing the real possibility of going home with a loss, I asked the AI for one last roll of the dice. In a surprising change of strategy, ChatGPT advised me to “go against the grain” for the final race.

It suggested a bold £10 win bet on Riskintheground (7/2). The heavy favourite, Blueking D’oroux (2/9), offered almost no value. A win here would cover all my losses and put me up £15, but a loss would leave me £30 down.

Both horses started well, and it quickly became a two-horse race between ChatGPT’s pick and the favourite.

As they entered the final stretch, Blueking D’oroux made a strong challenge, but Riskintheground held on, pushing ahead to win by three-quarters of a second.

Against all odds, ChatGPT had pulled it off. I was £15 up. It wasn't a fortune, but it was a profit, and more than enough to celebrate with a night out.

For more of the latest news, guides, and gossip, check out The Exeter Tab on Instagram, TikTok, and Facebook.