Originally published on Medium in 2024, republished here as I move my writing to my own site.

I’m visiting family in Illinois this weekend, which means we can sports bet. I made a GPT with access to weekly NFL Player Stats: Fantasy Football Stats Cruncher

Today I used it to help me make (semi) informed sports bets on the day’s games.

Travis Kelce

Travis Kelce TD bet odds

Travis Kelce to score a TD — estimated at 66% likelihood of No, and 44% likelihood of yes. What has he done previously?

ChatGPT showing Kelce's historical TD data

Hmm, seems like 67% chance of not scoring a TD to me.

Another Kelce data point

This data point tells me nothing.

Kelce defensive context

Ahh the context I needed.

So the odds were set slightly below his historical probability for this season, and on top of that, he is playing a good defense. Seems like the Swifties may be betting with their hearts.

Lamar Jackson

Lamar Jackson passing yards odds

Lamar Jackson Passing Yards are set to a 53% odds scenario no matter if you choose over or under.

Lamar's historical passing yard data

He’s gotten more than 211.5 yards 53% of the time.

KC defense passing yards allowed

KC has let up 211.5 yards 53% of the time.

So it’s two 53% scenarios? How does that work? Is the bet right on? Do I multiple them? I’m not sure, but I’m just going to take the over because even a square bet is a pretty good bet in sports gambling.

Wish me luck!