Advanced Poker Training’s Million-Sim Model Backs Greg Mueller
Lucas Jumalon was the clear numbers favorite, but Steve Blay still made the three-bracelet veteran his official pick.
Advanced Poker Training ran 1,000,000 simulations of the 2026 WSOP Main Event final table and found Greg Mueller to be Steve Blay’s official winner pick, even though the model gave Lucas Jumalon the best mathematical chance to win. The company’s numbers had Jumalon winning more than 40% of the time while starting with roughly 35% of the chips, but Blay still made Mueller his official winner pick.
Blay said Jumalon’s edge made sense because of ICM pressure. In his view, the other players could not attack as freely because every confrontation carried big payout consequences, and he said the chip leader’s simulated win rate translated into only about 2% more prize money than his ICM expectation.
Jumalon was not completely safe in the model, either. PokerNews reported that he could still bust in 9th place in 1% of the simulations.
Rami Hammoud was the biggest disappointment for the model, as his average finish came in nearly 5% below his ICM value. Blay pointed to Hammoud being “technically an amateur,” scouting reports that say he can be a little overaggressive, and a bad seat draw with Jumalon two seats to his left, where he can wait in the blinds with a massive stack and maximum leverage when Hammoud reaches the cutoff or button.
Mueller and Jamie Shaevel were the biggest overperformers in the simulations, with Mueller singled out as the one who kept jumping off the page. Blay said that came down to decent table position and a lot of experience, and he described Mueller as a three-time WSOP bracelet winner with dozens of cashes dating back before Jumalon was born, plus a former professional hockey player.
PokerTraining.com’s final-table profile put Mueller fourth of the nine players with 48.5 million chips. It listed his live tournament winnings at $3,540,270 before the final table and his biggest prior cash at $460,841, while noting that his bracelet run included a 2009 Limit Hold’em title, a $1,500 Shootout win, and a $10,000 H.O.R.S.E. bracelet in 2019.
Blay said each bot can be tuned across more than 40 characteristics, including reactions to ICM pressure and how willing a player is to risk tournament life, and said the framework can simulate three WSOP Main Event final tables per second. He also compared the approach to election forecasting and pointed to 2016, when Advanced Poker Training correctly identified Qui Nguyen as the most likely winner from just 100 simulations.
The final table was set with Jumalon on 194 million chips, Rami Hammoud on 79 million, and Jamie Shaevel on 56 million. The final nine each locked up at least $1 million, with $10 million for first place, and play is set to resume on Aug. 3.
