Supercomputer Predicts 2026 WSOP Main Winner After 1 Million Simulations

Last Updated on July 30, 2026 by Bala Kumar

The Setup: 1 Million Simulated Final Tables

With the 2026 WSOP Main Event final table just days from resuming, the poker training platform Advanced Poker Training has thrown serious computing power at a simple question: who wins? The company, run by Florida-based academic and software developer Steve Blay, simulated the entire nine-handed final table one million times.

Blay isn’t new to this game. Back in 2016,  the last time the WSOP Main Event had a similar break between the final table and the champion being crowned,  his model correctly picked Qui Nguyen to win even though Nguyen wasn’t the chip leader. That call was made using just 100 simulations. This year’s model ran 10,000 times more iterations, powered by a system capable of simulating three full final tables per second.

The Numbers: Full Win Probabilities

Here’s how the one million simulations broke down by finishing position for each of the nine finalists:

Player1st2nd3rdBust 9th
Lucas Jumalon40.4%20.4%13.1%1.0%
Rami Hammoud13.2%15.3%14.2%5.4%
Jamie Shaevel11.3%13.4%13.7%6.5%
Greg Mueller10.9%13.5%13.5%6.7%
Michael Gagliano7.0%10.0%11.3%10.3%
Mario Boos6.5%9.4%10.9%11.4%
Lauri Saaskilahti5.2%7.9%9.8%13.9%
Han Feng3.0%5.3%7.2%20.8%
Evagoras Evagorou2.5%4.7%6.3%24.0%

Jumalon, who bagged a commanding 194,000,000 chips (roughly 35% of the total chips in play), came out on top nearly 41% of the time — but that dominance wasn’t total. The simulations had him busting in 9th place, dead last, in 1% of runs.

Prize Money vs. Chip Equity

Perhaps the most interesting output wasn’t who wins, but who over- or under-performs relative to their pure chip-count (ICM) value:

PlayerStarting ChipsICM $ ValueSimulated Avg. WinningsEdge
Lucas Jumalon194,000,000$6,188,930$6,306,953+1.91%
Rami Hammoud79,000,000$4,054,198$3,862,553-4.73%
Jamie Shaevel56,000,000$3,414,290$3,605,675+5.61%
Greg Mueller48,500,000$3,177,214$3,568,945+12.33%
Michael Gagliano46,500,000$3,111,062$2,996,098-3.70%
Mario Boos44,000,000$3,026,466$2,899,613-4.19%
Lauri Saaskilahti37,500,000$2,795,602$2,673,088-4.38%
Han Feng25,000,000$2,296,565$2,231,290-2.84%
Evagoras Evagorou22,500,000$2,185,672$2,105,788-3.65%

Greg Mueller stands out here, outperforming his chip-based expected value by more than 12% — the single biggest edge of anyone at the table.

Interpreting the Data

The raw numbers only tell part of the story. Blay flagged a handful of results he found especially noteworthy.

Jumalon Is Powerful, But Not Invincible

Blay says Jumalon’s dominance comes down to simple math: holding roughly 35% of the chips, he won outright in over 40% of simulation runs, largely because no one else at the table can pressure him without taking on outsized ICM risk. Still, Blay admits he expected the chip lead to translate into a bigger edge than the roughly 2% bump over pure ICM expectation. And despite his massive stack, Jumalon still busted in last place in 1% of the runs — a reminder that nothing is locked in.

Simulator Says Hammoud to Underperform

Rami Hammoud was the model’s biggest disappointment, finishing nearly 5% below his ICM value on average. Blay chalks this up to a mix of factors: Hammoud’s relative inexperience as an amateur, scouting reports suggesting he can play too aggressively, and — above all — a brutal seat draw. Hammoud sits two seats to Jumalon’s right, meaning whenever he’s in prime position to steal blinds, Jumalon is often sitting behind him with a monster stack and maximum leverage.

Shaevel and Mueller: The Final Table Sleepers

If you’re looking for an underdog, Blay points to Jamie Shaevel and, especially, Greg Mueller as the two biggest overperformers — both bring strong table position and deep experience. Shaevel has eight career Main Event cashes and is a respected Los Angeles cash-game regular, while Mueller, a three-time WSOP bracelet winner and former professional hockey player, has been cashing WSOP events since before Jumalon was born and is comfortable performing under pressure.

It’s a Rich Man’s World

All five of the shortest stacks at the table underperformed their ICM expectations in the simulation — a result Blay says wasn’t surprising. Every pay jump at a final table represents life-changing money, which makes it far harder for short stacks to find profitable spots without risking elimination.

The Science Behind It All

Blay says the model runs on more than 40 configurable behavioral traits per player-bot, covering things like reaction to ICM pressure, willingness to gamble for tournament life, and blind-stealing tendencies,  a framework built up over decades. He compares the approach to modern election forecasting: rather than polling every voter, the model blends historical data, tendencies, and statistical patterns into a probabilistic picture that, while imperfect, tends to be directionally accurate.

The Expert’s Official Winner Prediction

Despite the numbers clearly favoring Jumalon, Blay isn’t picking the favorite. <cite index=”23-1″>”The player who kept jumping off the page in my simulations was Greg Mueller,”</cite> he says, putting his name behind Mueller to win it all and citing the value of experience under pressure over pure chip equity.

The 2026 WSOP Main Event final table resumes play on August 3 at Horseshoe and Paris Las Vegas, with $6.2 million in ICM equity on the line for chip leader Lucas Jumalon and a $10 million top prize up for grabs.

FAQs

Leave a Reply

Your email address will not be published. Required fields are marked *

Poker Platform

QQPoker (QQPK) is a real-money poker platform offering cash games, tournaments, crypto payments, and Triton Poker satellites.

 
 
 
 

ACR Poker is a trusted real-money poker platform offering large tournaments, crypto payments, and games for both beginners and experienced players.

 
 
 
 
ClubWPT Gold is a WPT-backed sweepstakes poker platform with tournaments, cash rewards, and built-in hand analysis tools.

Follow Stake’s official social media channels for the latest poker news, promotions, and updates.