Button Vs Blinds, SPR Thresholds And The ICM Squeeze Trap

I busted ninth at a final table two summers ago, and I have replayed the hand more times than I would like to admit twenty-three big blinds. The chip leader opened from the button, a short stack flat-called from the small blind, and I looked down at a hand that was, in pure chip terms, an easy shove. So I shoved.

The chip leader called with something worse than my hand and hit. That part was variance. The part that was not variance is that I had made a play that was correct in a currency I was not being paid in.

Three ideas explain what actually went wrong, and they stack on top of each other: who has the advantage in button versus blind pots, what stack-to-pot ratio does to your postflop options, and why the squeeze becomes the single most expensive button on the table once real money is on the line.

Button Versus Blinds Is Not Meant To Be Fair

Start with the geometry, because everything downstream depends on it.

The button opens the widest range at the table because only two players are left to act and both of them will be out of position for the rest of the hand. That is a structural gift, and good players take it aggressively.

The big blind’s counterplay is pot odds. Against a standard 2.5bb open with the small blind folding, there are four big blinds in the middle, and it costs you one and a half more. You need roughly twenty-seven percent equity to continue, which is why the big blind defends enormously wide, often more than half of all hands. That is not looseness. It is arithmetic.

The small blind is the worst seat in poker, and it is not close. You are out of position, you have a live player behind you, and flatting invites the big blind to squeeze. Most solid tournament players collapse the small blind into something close to raise-or-fold for exactly this reason.

Here is the consequence people miss. Because the big blind defends so wide, the big blind’s range on most flops is wide and capped, while the button’s range is wide and uncapped. On an ace-high board, the button has more aces. On a low connected board, the big blind has more of everything else. Those two facts drive almost every postflop decision in the most common pot in modern tournaments.

SPR Is Decided Before The Flop

Stack-to-pot ratio is your effective remaining stack divided by the pot when the flop lands. The concept was introduced by Matt Flynn, Sunny Mehta and Ed Miller in Professional No-Limit Hold’em, and it is the most useful number nobody calculates at the table.

The thresholds that matter in practice:

• Under 3. You are committed with a lot less than you think. Top pair with a decent kicker is frequently a stack-off hand. Overpairs are gold. Draws lose value because there is not enough money behind to justify chasing.
• Three to six. The awkward zone. Top pair is now a strong hand that cannot comfortably absorb three streets of pressure. Two pair or better is your stack-off threshold most of the time. This is where most tournament money changes hands and where most players are worst.
• Seven to thirteen. Sets, straights and big draws. Top pair is a one-street or two-street hand. If you find yourself putting a third barrel in with a single pair here, something has gone wrong earlier in the hand.
• Above thirteen. Deep. Implied-odds hands like suited connectors and small pairs climb in value, and the discipline required to fold overpairs goes up with them.

The important part is not memorising the bands. It is realising that you choose your SPR before the flop and then live with it.

Where The Two Ideas Collide

Run the numbers on the same matchup at different stack depths, and the point makes itself.

Hundred big blinds deep, button opens to 2.5 and the big blind calls. The pot is about five and a half, stacks behind are about ninety-seven. Your SPR is close to eighteen. That is a hand where top pair is a modest holding, and you are playing three streets of chess.

Same seats, same action, but the big blind three-bets to ten and the button calls. Pot around twenty, stacks around ninety, SPR roughly four and a half. Top pair just became a stack-off candidate. Nothing about the cards changed. You changed the game by choosing a sizing.

Now bring it to a final table, where twenty-five big blinds is a normal stack. Button opens to 2.2, big blind calls, pot is about five, stacks behind are around twenty-three. SPR under five. You are one bet and a raise from playing for everything, and you got there without anyone doing anything unusual.

That is the thing about deep tournament play. The SPR collapses on its own as blinds climb, and every decision you took for granted at a hundred big blinds needs rebuilding.

ICM Makes Every Chip Asymmetric

Cash chips are linear. A hundred chips is worth exactly twice fifty. Tournament chips are not, and the Independent Chip Model is the tool that quantifies the difference, converting stacks into a share of the prize pool based on the payouts still outstanding. The underlying maths comes from a 1973 model David Harville built for horse racing, which tells you something about how long people have been trying to price uncertainty.

The consequence is brutal and simple: the chips you can lose are worth more than the chips you can win. Doubling your stack does not double your equity, but busting takes all of it.

And it does not hit everyone equally. Medium stacks carry the most ICM pressure at a final table because they have the most to lose and the least protection. The short stack is already near the floor. The chip leader can absorb a loss. The player with twenty-three big blinds and five people still to bust is the one paying the highest premium on risk.

If this sounds like formalised loss aversion, that is because it more or less is. Kahneman and Tversky showed decades ago that people weight losses more heavily than equivalent gains, and ICM is the rare situation where that instinct is not a bias to correct but a real feature of the payout structure.

Why The Squeeze Is The Most Expensive Button At A Final Table

Now back to my hand.

The squeeze looks irresistible under ICM. An opener and a caller both have ranges that cannot comfortably continue, and both of them are supposed to be feeling the same pay-jump pressure you are. On paper, you are printing.

The problem is who is left behind the play. When you squeeze the chip leader, you are attacking the one player whose risk premium is lowest, using the stack whose risk premium is highest. They can call you with hands they would fold to an equal stack, because busting you costs them a fraction of what busting you costs you. Your fold equity is not what your instinct says it is.

There is a second trap underneath. The squeeze is a chip-EV play by construction. You compute it in big blinds. You pay for it in finishing positions. A shove that shows a small chip-EV profit can be a sizeable dollar-EV loss, and nothing at the table tells you which one you are looking at.

What I do now: I squeeze when I cover the opener, not when they cover me. As the medium stack with pay jumps live, I fold hands I would have shoved two years ago, and I feel slightly sick doing it every single time. That feeling is the price of the correction.

What Actually Improved My Game

Reviewing hands away from the table did more than any in-session realisation. So did drilling the same spots repeatedly at low stakes, which is the honest argument for a private table: playing poker with friends lets you rehearse squeeze and defence spots without a pay jump punishing every mistake you make while learning.

It is also worth knowing how thoroughly this game has been mapped. When Brown and Sandholm’s Pluribus beat elite professionals at six-handed no-limit hold’em, it did so with strategies that leaned on bet sizing and position rather than on reads, building on earlier work where DeepStack reached expert level in heads-up play. The blind-versus-button structure the solvers converged on is not a fashion. It is the shape of the game.

None of which makes results predictable. Skill is real and measurable in the aggregate, with one study of the World Series finding that players identified in advance as highly skilled returned over thirty percent while everyone else lost money. But that is an average across thousands of entries, not a promise about your next final table.

Part Nobody Puts In A Strategy Guide

Variance in tournaments is enormous, downswings last longer than feels reasonable, and the correct play loses constantly in the short run. Play with money you can afford to lose entirely, set limits before you sit rather than during a session, and treat a losing stretch as information about the sample size rather than about your worth. If any poker game has stopped feeling like a game, the National Council on Problem Gambling runs free, confidential support around the clock.

I still think about ninth place. But I fold that hand now.

Disclaimer

References

• Card Player. Pot-Limit Omaha: The Stack-to-Pot Ratio (SPR), Part I. Card Player Poker Magazine. https://www.cardplayer.com/cardplayer-poker-magazines/66224-poker-year-in-review-26-1/articles/21011-pot-limit-omaha-the-stack-to-pot-ratio-spr-part-i
• Independent Chip Model. Wikipedia. https://en.wikipedia.org/wiki/Independent_Chip_Model
• Kahneman, D., and Tversky, A. (1979). Prospect theory: an analysis of decision under risk. Econometrica, 47(2), pp. 263-291. https://doi.org/10.2307/1914185
• Brown, N., and Sandholm, T. (2019). Superhuman AI for multiplayer poker. Science, 365(6456), pp. 885-890. https://doi.org/10.1126/science.aay2400
• Moravčík, M., Schmid, M., Burch, N., Lisý, V., Morrill, D., Bard, N., Davis, T., Waugh, K., Johanson, M., and Bowling, M. (2017). DeepStack: expert-level artificial intelligence in heads-up no-limit poker. Science, 356(6337), pp. 508-513. https://doi.org/10.1126/science.aam6960
• Levitt, S. D., and Miles, T. J. (2014). The role of skill versus luck in poker: evidence from the World Series of Poker. Journal of Sports Economics, 15(1), pp. 31-44. https://doi.org/10.1177/1527002512449471
• National Council on Problem Gambling. Help and Treatment. https://www.ncpgambling.org/help-treatment/