Short Stack Poker Strategy

What counts as short stacked

"Short stacked" gets used a few different ways, and they're worth telling apart:

• Relative to the field. You're short if you have fewer chips (in big blinds) than the average stack at your tournament. This can be a variable target that shrinks as the blinds go up, even if your own stack hasn't lost a hand.
• Under ~20bb effective. A rough, widely-used benchmark. Below it, stack-to-pot ratios get small enough that opening hands commits a real fraction of your stack, and preflop decisions start to dominate the hand. The button's solved 20bb chart is the clearest place to see the benchmark bite: it is still recognisably an opening range rather than a pure shove.
• The smallest stack. "The short stack" (with a definite article) usually just means whoever has the fewest chips in the hand, at the table right or in the tournament right now, regardless of the actual number.

They're the same underlying idea (not much room behind your stack) measured in different ways. What actually changes your strategy is the number in big blinds, so that's what this guide - and every chart on this site is built around.

Why it's worth focusing on

Being short isn't just something to endure until you can play "real poker" again. It comes with real benefits.

• Easier decisions. Under about 10bb, almost every hand worth playing is a shove-or-fold decision with no postflop street to navigate at all. Fewer branches means fewer ways to go wrong.
• Easier to study. One street (preflop) to focus on, which can give you a massive edge over the field. That's a small enough number of permutations to actually memorize solved charts for, which isn't realistic for the full tree of a 100bb postflop hand.
• Better equity realization. A shove doesn't just risk your stack, it denies your opponent the rest of the hand, and makes sure that you never end up folding a hand that could have been a winner. Many hands have real equity against a calling range but would struggle to realize much of it across five streets of postflop play out of position. However when jammed preflop, it gets to cash in all of that equity in one go, plus whatever fold equity the shove itself earns on top. That's why solved shoving ranges are consistently wider than they "should" be by raw equity alone. This is also important to keep in mind for exploitative decisions as you continue to hone your skills.
• ICM. Once real money is on the line at pay jumps, chips stop being worth a fixed amount - busting is disproportionately costly compared to what you gain from winning a marginal pot. Short-stack decisions are where that pressure is sharpest, which makes this the highest-leverage spot in the tournament to actually understand ICM.

Push/fold: the strategy

Push/fold is exactly what it sounds like: instead of open-raising to a size and having to decide whether to continue against a 3-bet (or playing post flop), you commit your whole stack or you fold. It's a simplification, but it's not an approximation once you're short enough - below roughly 10bb effective there's rarely enough stack left behind to profitably raise-fold, so if a hand is good enough to open at all, it's usually good enough to get all-in. That makes push/fold the actual chip-EV-optimal strategy in this range, not just a simplified stand-in for one.

The button's solved opening range at two depths shows why. At 10bb it shoves 40% of hands, 69 of them at full frequency. Drop to 5bb and the range widens to 47% - shorter means a shove risks less in absolute terms, so more hands clear the bar.

Not sure how to read the colors or the split-fill cells above? Start with How to Read a Push/Fold Chart. For the full picture of how a range changes from 5bb up through 15bb effective - and where it stops being a pure shove-or-fold decision - see Push/Fold Strategy by Stack Depth.

Position changes the range as much as depth does

The same stack size plays completely differently depending on who's left to act behind you. Opening under the gun means the whole table can still wake up with a hand, so that range stays tight even when you're short. Opening from the button, with only the blinds left, is a different game entirely - the range widens sharply because there's so much less behind you to punish it. The blinds have their own asymmetry on top of that: defending against an under-the-gun open is a much tighter decision than defending against a button open, because the range you're facing is tighter too. See Big Blind Defense for the solved charts behind that comparison.

Table size pulls the same lever, in a way that is easy to guess wrong. The button always has only the blinds left to act behind it, so a 9-max button looks like it should play the same as a 6-max one. The solved charts disagree: at 10bb the 9-max button shoves 48% against 40% at 6-max. Eight players have posted an ante rather than five, so there is more dead money sitting in the middle to win uncontested, and that extra reward pulls another few percent of hands over the line.

Depth and position aren't the only things that move a range - the objective you're solving for does too. Below is the button's opening range at the same 10bb effective stack, solved two ways: chip-EV, and ICM on a real bubble payout structure with both blinds also sitting short (this is the one ICM example free accounts can see - full ICM coverage is every position, depth, and payout structure, and it's a Pro feature). Chip-EV shoves 40% of hands here; under ICM pressure that drops to 32% - every hand that changes moves the same direction, tighter, purely because busting on the bubble costs more than the chips themselves are worth.

Layering in ICM

Near the bubble and every pay jump after it, chip-EV stops being the right model on its own. ICM converts everyone's stack into real tournament equity given the payout structure, then re-solves the same push/fold question against that instead of raw chips. The practical effect can vary based on the stack depths involved. Generally there is a risk premium where risking your stack is a lot worse compared to Chip-EV. However you can apply this same pressure to other opponents if you are the covering stack. Start by learning a good Chip-EV baseline strategy, then focus how ICM can impact different aspects of your strategy. Push/Fold is still a dominant strategy in heavy ICM scenarios, however some players also mix in min-raises at the top and bottom of their range once ICM stacks get a little deeper than pure push/fold territory. That's a real, important idea, but it's a sizing decision this site's charts don't currently model, so treat it as something to be aware of rather than something to look up here.

Stage of the tournament matters too, independent of your own stack: the same 15bb stack plays differently three tables from the money than it does one spot from a pay jump, even though the chip-EV math hasn't changed at all - only the payout structure sitting on top of it has.

Adjusting to how your opponents actually play

Everything above assumes opponents who play back at you close to optimally. Real opponents don't, and once you have a real read - not a guess off one hand, but a sustained pattern, the highest-EV range against that specific player isn't the balanced one anymore. GTO vs. Exploit covers when that adjustment is worth making, and Exploiting Tight and Loose Opponents shows the actual hands that move: wider against a big blind who folds too much, tighter against one who calls too much - and it's worth noting those two adjustments aren't mirror images of each other, since widening against folding creates new bluff candidates from nothing while tightening against calling just erodes the fold equity behind hands you were already shoving. It gets more interesting still with two different reads to weigh at once - a tight player in one blind and a loose one in the other, or a loose big blind behind a short stack that's pricing in its own ICM survival. That combination is exactly what the app's ICM exploit trainer is built to drill.

If the tables you play are low stakes, that adjustment is worth more than it is anywhere else - your opponents deviate further, so the gap between the balanced range and the highest-EV one is wider. Push/Fold Strategy for Low Stakes sets out the order to do it in, and why going straight to the exploit without the baseline first is just guessing.

This is the same border-coded comparison Practice shows you after every hand: your answer's range against the GTO baseline for that spot, with any cell that moved meaningfully ringed to show which way. Below, the button's 10bb shove range widens from 40% at GTO to 50% once the big blind is tagged as folding too often - every ringed cell is a hand that only makes sense once you know this specific opponent's tendency.