Netflix is reportedly paying at least $60 million in purses to make history in its first-ever, live, non-pay-per-view sports broadcast. The streaming giant's venture into live programming pits 27-year-old YouTuber-turned-boxer Jake Paul against Mike Tyson, 58, in a November 15 clash at AT&T Stadium that's already drawing intense scrutiny from prediction markets and sports analysts.

“I’m here to make $40m and knock out a legend,” Jake Paul told interviewers.

"This fight is not going to change my lifestyle financially," Tyson said. "I feel I can beat this guy."

Some experts say Paul will get the lion’s share of the purse, though others say it’ll be a 50/50 split.

Either way, there’s a lot of betting going on—and as yet another public service to the degen community, we wanted to see what AI could tell us about this particular instance of the sweet science. So we built a GPT—basically a custom AI agent powered by OpenAI's GPT-4o—to cut through the noise and deliver data-driven fight predictions, continuing our exploration of artificial intelligence in sports forecasting.

After a satisfactory Kentucky Derby analyzer with HorseGPT and a baseball expert correctly calling the Dodgers' World Series victory, we turned our attention to boxing's most talked-about matchup.

We called it BoxingGPT, which you can spar with yourself, or read on.

Our combat sports AI digested decades of boxing data, including historical match outcomes, performance metrics, and detailed technical analysis of fighting styles—well, at least decades of Tyson’s fights, because El Gallo de Dorado is relatively new at this and half as old as Iron Mike.

We loaded the model with research papers on age impact in heavyweight boxing, punch statistics from Paul's previous fights, and Tyson's complete career data. The process mirrors our successful approach with HorseGPT, except it analyzed humans instead of horses.

After processing all this data, our AI made its mind: Unbelievably, Tyson holds a slight edge over Paul—but the odds will ultimately depend on the fight’s duration. Our bot sees Tyson as a sprinter, whereas Paul seems more like a marathon runner.

A short fight will favor Tyson, in other words—but once he gets tired, he’s done.

Quoth our bot: “In a hypothetical matchup, Tyson’s raw power and aggressive style make him the favorite. His path to victory hinges on landing early and putting pressure on Paul, who would need to play a disciplined, defensive game to survive. The younger Paul has a chance if he can weather the early rounds and exploit Tyson’s age-related decline in the later rounds, but Tyson’s extensive skill set and experience make him a formidable challenge.”

This runs counter to all the conventional wisdom (including CBS Sports and FanDuel Sportsbook) which favors Paul to win with odds ranging from -215 to -310—meaning that bettors would need to wager between $215 and $310 to win $100. Polymarket bettors also see Paul as the likely winner.

That’s the tl;dr. If you’re interested in the fine print, read on.

When AI and humans think differently

BoxingGPT breaks down the fight into distinct phases:

Early Rounds (1-3): Tyson would likely come out aggressively, aiming to get inside Paul’s reach and unload powerful combinations to the head and body. Paul might try to keep Tyson at range with jabs and straight rights, but Tyson’s head movement and relentless forward pressure could make this difficult. Paul’s best strategy would be to remain calm, stay at range, and avoid early exchanges where Tyson could dominate.

Middle Rounds (4-6): If the fight goes into the middle rounds, then Tyson’s stamina could start to be tested. This could open opportunities for Paul if he’s managed to stay relatively unscathed. Paul’s conditioning advantage would allow him to control the pace more effectively if Tyson begins to tire. Tyson would likely adjust his output to conserve energy, picking moments to close the gap.

Later Rounds (7-8): If the fight extends into the later rounds, then Paul would have an advantage in terms of stamina and size. By this point, he could begin landing more shots if Tyson’s defense and head movement start to slow. Paul might win rounds simply by staying active and controlling distance if he can avoid Tyson’s powerful counters.

Our bot’s predictions don’t appear to be aligned with the analyses shared by other experts with more skin in the game, who put Jake Paul as the favorite, mainly due to the 30-year difference between the two pugilists.

Polymarket shows Paul as a 62% favorite, with Mike Tyson having less than a 30% shot of prevailing. Our model suggests these odds may be overvaluing youth and underpricing Tyson's experience—particularly his 88% career knockout rate, which remains an outlier even in adjusted aging curves.

Martialbot, another specialized boxing platform, takes an even stronger stance, giving Paul a 99% chance of victory. “This should be a confident win for Jake Paul at 99-1 winning odds. Mike Tyson really doesn't have much to present against Paul,” the site argues.

That said, Martialbot’s analysis said the same thing about Jake Paul’s fight against Tommy Fury in 2023, and it ended up completely wrong. The site also gave Clifford Etienne 84% chance of victory against Tyson, and the fight ended in round 1 with Tyson knocking Etienne out in the first round.

In other words, the Tyson-Paul fight appears to be pretty similar to his match against Etienne in terms of expectations vs strategy—so you better be quick, Mike.

The betting odds from DraftKings SportsBook favor Paul as a -210 favorite, with Tyson as a +170 underdog—basically, to win $100, you must bet $210 on Paul or $58.82 on Tyson. Tyson is +250 to win by knockout, meaning a $100 bet would return $250 if he wins by knockout, while Paul is +140 to get the finish, returning $140 on a $100 bet if he stops Tyson.

But there’s something our AI and the betting markets agree: The fight is expected to be fast-paced and unlikely to reach the final bell, with odds at -130 for an outcome under 6.5 rounds. Paul is more likely to win by KO or TKO at +125, while a decision victory for him is less favored at +320. For Tyson, the most probable path to victory is also by KO at +260, with long odds of +1100 for a decision win, reflecting his age and conditioning.

Round-by-round odds favor Paul’s chances in the middle rounds, while Tyson is given his best shot early on, with his odds for a KO win reaching as high as +6500 by the eighth round—meaning a $100 bet on Tyson winning in the eighth would return $6,500. A draw is an unlikely outcome, sitting at +800, so a $100 bet would return $800 if the fight ends evenly.

Whether our BoxingGPT joins the success of HorseGPT or follows Martialbot's tracks of missed calls remains to be seen. Until then, our AI suggests keeping a close eye on those first three rounds—they might be worth more than all the prediction markets combined.

Edited by Andrew Hayward