A supercomputer has crunched through 1 million simulated finishes of the 2026 WSOP Main Event final table, and the math points heavily toward chip leader Lucas Jumalon. The tool’s own creator, however, is picking against his own numbers, backing a different player entirely as the final nine get ready to play down to a champion.
The Mind Behind the WSOP Main Event Simulation
The model comes from Steve Blay, a Florida based academic and software developer who runs Advanced Poker Training. Blay built an AI powered simulator that assigns each remaining player more than 40 configurable characteristics, covering everything from how they respond to ICM pressure to their blind stealing tendencies and continuation betting patterns. The current version of the tool can simulate three complete final tables every second, which is how it reached the 1 million mark well before the nine players even sat down to play.
Blay’s track record gives the project some real credibility. Back in 2016, running just 100 simulations on far more limited hardware, his model correctly picked Qui Nguyen to win the Main Event even though Nguyen did not hold the chip lead heading into the final table. A decade later, with vastly more computing power behind it, Blay’s tool is being taken seriously by a poker media ecosystem that usually treats bracket style predictions with a healthy dose of skepticism.
Jumalon’s Chip Lead Shows Up Clearly in the Numbers
Lucas Jumalon, the 22 year old from Spokane, Washington, bagged 194 million chips to end Day 7, giving him close to 35 percent of the chips in play and nearly two and a half times the stack of his nearest competitor. That advantage carries straight through the simulation results, with the model giving Jumalon a 40.4 percent chance to win the bracelet and the roughly 10 million dollar first prize that comes with it.
No other player comes close to that number. Rami Hammoud is next at 13.2 percent, followed by Jamie Shaevel at 11.3 percent and Greg Mueller at 10.9 percent. Michael Gagliano sits at 7.0 percent, Mario Boos at 6.5 percent, Lauri Saaskilahti at 5.2 percent, Han Feng at 3.0 percent, and Evagoras Evagorou rounds out the field at 2.5 percent. The spread reflects how much of a mountain a short stack has to climb at a nine handed WSOP Main Event final table, even against players with a real edge in skill.
Why Chip Stack Alone Rarely Tells the Whole Story
Big stacks win WSOP Main Events more often than not, but poker’s history is full of final tables where the leader going in did not leave with the bracelet. Understanding why requires looking past raw chip counts and into how ICM pressure changes optimal strategy as the table shortens, a topic covered in more depth in Poker Pro Academy’s WSOP guide. A stack that looks dominant nine handed can shrink quickly if its owner plays too passively while shorter stacks apply pressure through repeated shoves and re-shoves.
That gap between raw chip equity and eventual win probability is exactly what a simulation like Blay’s is built to capture, weighing not just stack size but each player’s tendencies under the specific pressure of a televised final table with life changing money on the line.
The Model’s Creator Picks Against His Own Numbers
Despite Jumalon’s commanding statistical edge, Blay is not backing the favorite. “The player who kept jumping off the page in my simulations was Greg Mueller,” Blay said, explaining his personal pick. His official call: “Greg Mueller for the win, baby!” It is an unusual move for the man who built the model to publicly pick a player sitting fourth in his own win probability rankings, but Blay has pointed to patterns in Mueller’s simulated play, rather than the raw percentage, as the reason for his confidence.
That kind of gap between a model’s output and its creator’s gut read is part of what makes these predictions compelling to poker fans rather than just a dry statistics exercise. It also mirrors the split reaction across the poker media landscape, where some outlets have leaned on prediction market data pointing to Jumalon as the heavy favorite, while others have highlighted underdog storylines among the shorter stacks.
What Comes Next for the Final Nine
With the final table set to play out imminently, Jumalon’s simulated 40.4 percent edge will be tested in real time against eight players hoping the numbers are wrong. Whether the outcome favors the model’s math or Blay’s personal hunch for Mueller, the final table adds another data point to a technology that has already proven, at least once before, that it can spot a champion the raw chip counts alone would never suggest.
