POTD #321 FT Friday: Our Final Main Event Hand
Greg Mueller folding JJ
Poker is a game of valuation. In cash games you need to determine how much your hand is worth. In tournaments you need to determine both how much your hand is worth and how much your stack is worth. Then you need to synthesize the two valuations to create a harmonious strategy. The most common way one values their stack is ICM, specifically Malmuth-Harville ICM. ICM might seem very complex, but it’s a pretty simple model— ICM is a model that says if you have 39% of the chips you will win the tournament 39% of the time. If I had to create a naive model to estimate the value of someone’s stack in a poker tournament, I might land on something similar to ICM. The fact that ICM is simple and its outputs line up with a naive model is a feature, not a bug.
Making a model more complex does not always make it more accurate. If it gives you an output that is nowhere close to a naive model, the odds are more likely the model is over/under-fit than it’s revolutionary. ICM gets you in the ballpark and that is often enough. If a perfect MTT model provides an output that disagrees with ICM— let’s say, one that says if you have 39% of the chips you should win the tournament 42% of the time— how much would that actually change your strategy? I suspect the changes will be subtle and on the margins.
If you face a shove at a FT and you know that one model wants 48% to call off, while another wants 50% to call off, and your hand has 49%, your decision is a gamble on which model is more accurate. It’s easy to spend a lot of time debating over which model is right, but when facing most opponents you should focusing on how much equity you have vs. your opponent’s actual range. Different sims might have different thresholds to call all-in, but they’ll be in the same ball park, while your equity vs different humans’ ranges can easily vary between 5-10%. You’ll make a lot more money calling one pip too wide vs. a player too loose, or hero-folding vs. a nit, than you will poking small holes in ICM. Fortunately, you’re allowed to do both.
In today’s hand, Greg “FBT” Mueller faced a minraise and large shove in what could have been a three-way all-in with five left in the World Series of Poker Main Event. Everyone (I hope) understood that he had an easy call for chips, but it was a tough technical spot at a high stakes FT. However, he was against two opponents who had some unbalanced tendencies, which made his decision easy from my view on the sideline. I think he may have been too concerned with the math puzzle of the hand and not concerned with the human puzzle.
World Series of Poker Main Event The Final Table
(1.5M/3M/3M) (SB/BB/BBA)
5th: $2.25M, 4th: $2.75M, 3rd: $3.75M, 2nd: $6M, 1st: $10M
Video of the hand Pokernews Update
Han Feng (25M/8.3BBs) makes it 6M UTG with A♦️A♣️, Lucas Jumalon (376.5M/125BBs) folds, Lauri Saaskilahti (59.5M/20BBs) makes it 50M with 9♠️9♥️, Michael Gagliano (26M/8.6bbs) folds the SB, Greg Mueller (63/21BBs) folds J♦️J♣️, Han Feng calls.
Lauri wins on the thrilling runout of T♦️8♦️3♦️J♥️Q♠️.
A reminder if you are not currently a paid subscriber you can get
20% off 1 month
25% off 3 months
30% off 1 year
At the above links.
