Push/fold strategy: when a shove is the only sensible bet left.
Below roughly fifteen big blinds, raising small stops working — any raise commits your stack anyway, so the fold you were saving does not exist. What is left is all-in or nothing, and unlike most of poker it can be worked out exactly rather than felt out.
Under about fifteen big blinds a raise commits you anyway, so all-in or fold are the only two coherent actions left. A shove is worth making whenever your opponents fold often enough to cover what a call costs you — and that bar drops fast as you get shorter. At twenty big blinds you need them to fold roughly 67% of the time. Below about 2.9 big blinds you need no folds at all: the shove is profitable even when it is always called.
Why a short stack has only two moves.
Push/fold is not a simplification for beginners. It is what is left once your stack is too small for a raise to mean anything, and there are four reasons the format takes over.
A raise already commits you
Open to 2.2 big blinds from a twelve-blind stack and you have put a fifth of it in. If somebody shoves back you are getting a price you cannot refuse, so the fold you were hoping to keep does not exist. The raise was an all-in with extra steps.
All-in maximises the only weapon you have
Short stacks cannot outplay anyone after the flop — there is no flop. What you can still do is make folding correct for everyone else, and the largest bet available is the one that does that most often.
The pot is now worth taking
Blinds and antes that were rounding errors at 60 blinds are a serious prize at 10. Winning them uncontested adds a quarter of your stack, which is why open-shoving is a profitable habit rather than a desperate one.
It makes the game solvable
With only two actions and no post-flop play, the whole decision collapses to one equation. That is why push/fold is the one part of tournament poker where you can genuinely know the right answer rather than estimate it.
How often do they have to fold?
A shove is worth making when the blinds you pick up unopposed cover what a call costs you on average. That is the entire decision, and it can be written down. Here is the answer at every depth a short stack actually faces.
The third column never moves. Whatever your stack, a shove nobody calls wins the same 2.5 bb — the blinds and the ante, and no more. What changes is the second column: the more chips you have, the more a call costs you. A fixed prize against a growing risk is why the required fold frequency climbs from 1% to 67% across this table, and why the same hand is a clear shove at one depth and a punt at another.
Below about 2.9 big blinds the last column disappears. At that depth the dead money is large enough relative to your stack that the shove shows a profit even if you are called every single time. Fold equity stops being part of the calculation, which means the only mistake still available to you is folding.
Worked at 35% equity when called — roughly what a shoving range holds against the range that calls it — into 2.5 bb of dead money. Move either assumption and every row moves with it; the shape of the curve does not.
What it costs to wait for a hand.
Folding feels like the safe move because nothing happens when you do it. Something does happen: the blinds arrive anyway, and the price they charge has an exact equivalent in shoves you did not make.
The two middle numbers are the same number. An orbit of folding costs 2.5 bb, and a shove that everybody folds to wins 2.5 bb. So sitting out one orbit does not cost you “a few chips” — it costs you one successful steal, which you now have to make again just to get back to where you started. Two orbits of patience is two steals. Nobody thinks of it that way, which is why it is the most expensive habit at a short stack.
And it usually buys nothing. Over the 9 hands of an orbit you will pick up a genuinely strong holding about 37% of the time — so roughly 63% of the time you pay the orbit and are handed nothing for it, at a stack that is now shorter than the one you were unhappy with.
The honest counterpoint: blinding down does make the next shove easier to justify. A 15-blind shove needs folds 59% of the time; by 5 blinds that has fallen to 20%. But the bar comes down because the stack behind it has shrunk, not because your position improved. A shove that gets through at 15 blinds leaves you with 17.5; the same shove at 5 leaves you with 7.5. Cheaper to make, worth less to win.
Position, and the hands that travel well.
The equation tells you what a shove needs. These four things decide whether you are getting it. All of them are visible at the table before you look at your cards.
Count the players behind you, not your seat
Position matters here only because it changes how many hands can wake up behind you. Shoving under the gun nine-handed asks eight players to fold; the same shove on the button asks two. That is the entire positional adjustment, and it is why your range can roughly quadruple in width across one orbit.
You are not shoving from a seat. You are shoving through a number of opponents.
Showdown strength beats playability
A shove wins by fold equity first and equity-when-called second, which favours hands that hold up against a calling range rather than hands that flop well. Any ace, any pair and suited broadways are shoves long before suited connectors are, because you will rarely see five cards.
K-9 offsuit shoves from more seats than 7-6 suited does, and the gap widens as you get shorter.
Shove into the stacks that can fold
A big blind with twelve blinds behind can fold and still have a tournament. One with three cannot, and will call with almost anything. Look left before you shove: the correct range depends more on who is behind you than on what you hold.
The same hand is a clear shove into one player and a punt into another.
Never limp, never min-raise, never call
Limping surrenders fold equity, min-raising commits you without buying it, and calling a raise leaves you playing a flop with a stack too small to bet twice. All three are ways of paying the price of a shove without collecting its benefit.
At this depth, every action that is not all-in or fold is a slow fold.
The same equation, three different games.
Nothing about the maths changes as you get shorter — only the answer it gives. Pick the band you are actually in and play the one after it as well, because you will be there within two orbits.
15–20 bb · The upper edge
Still just deep enough to raise-fold, and already shallow enough that shoving is often better.
- Open-shove from late position, but keep a raising range from early seats.
- Do not shove into a table that has shown it will call wide — the bar is high here.
- Prefer shoving over a limper; their limp is dead money and their range is capped.
- Look for a spot before this becomes 12 blinds and the decision is made for you.
Use the last depth where you still have two options, because at 67% folds required, a shove needs a genuinely tight table.
8–15 bb · Push/fold proper
The depth this entire page is about — all-in or fold, nothing in between.
- Shove or fold. Every other action costs you the fold equity you are relying on.
- Widen dramatically as the players behind you decrease.
- Attack big blinds who have enough chips to fold and enough sense to do it.
- Take the first reasonable spot rather than the perfect one two orbits later.
At 46% folds needed around ten blinds, a shove clears the bar against almost any normal table.
Under 8 bb · Nearly forced
Fold equity is fading and the maths stops needing it.
- Shove any two from late position when it folds to you.
- Stop waiting for a hand — the blinds are arriving faster than the cards.
- Expect calls. Below this depth the shove is profitable anyway.
- Do not fold into the money at this stack; blinding out finishes in the same place.
At 20% required folds, you are barely relying on folds — and below about 2.9 blinds, not at all.
Why prize money makes your shove better.
Everything above prices your shove. This prices their call — because a shove only works if folding is right for the people behind you, and near the money it becomes a lot more right than a chip count suggests.
Four players left and three get paid $500 / $300 / $200. You are on 6,000 chips and you shove. Here is the equity each player behind you needs before they can call — a number that would be exactly 50% for every one of them if chips were money.
Not one of them can call at 50%. On average they need about 56%, and the hands that sit between those two figures — the middling aces, the small pairs, the suited broadways — stop being calls and become folds. That is a straight transfer to you: the fold frequency your shove needs from the table in section two does not change, but the frequency you actually get goes up.
Which is the real reason short stacks get paid near the money. Your shove is not better because you are desperate. It is better because everybody else has something to lose and you have already lost most of it — and the closer the field gets to a pay jump, the further their bar climbs above yours. The ICM guide works through where these numbers come from; the bubble guide covers what to do with the rest of your table while it is happening.
Four ways short stacks throw it away.
Every one of these is a way of paying the price of a shove without collecting what a shove buys. They are also, in order, the four most common things you will see at a short-stacked table.
Waiting for a hand
The single most expensive habit at a short stack. Every orbit costs you blinds and antes whether or not you play, and the hand you are waiting for arrives after the stack that made it worth playing has gone.
Min-raising instead
It looks safer and it is strictly worse. You commit a fifth of your stack, you buy almost no fold equity, and you cannot call the shove that comes back. A raise you cannot defend is an all-in you did not get paid for.
Shoving pretty hands over strong ones
Suited connectors need a flop you will usually not see. Against a calling range you want raw showdown strength — aces, pairs, big cards — because the hand almost always runs to five cards uninterrupted.
Ignoring who is behind you
The chart in your head assumes an average table. A big blind with three blinds behind calls with anything and destroys your fold equity; one with fifteen folds correctly and hands you the pot. Look left first.
The push/fold cheat sheet.
Everything above, compressed to what to do and what to stop doing once your stack drops under fifteen big blinds.
- Shove or fold below about fifteen blinds — treat every other action as a leak.
- Widen sharply as the number of players behind you falls.
- Prefer raw showdown strength over hands that need a flop.
- Check the stacks behind you before choosing a range.
- Don't min-raise a stack you cannot defend with.
- Don't wait for a premium — the blinds arrive faster than the cards.
- Don't shove into a player too short to fold.
- Don't limp. There is no depth at which it beats shoving or folding.
You can work this one out at the table.
One equation, not a chart.
A shove needs folds often enough to cover what a call costs. That is it. At twenty blinds the bar is about 67%; by ten it has fallen to roughly 46%, which almost any table clears.
Under three blinds, folds stop mattering.
Below about 2.9 big blinds the dead money alone makes the shove profitable even against a certain call. At that depth the only mistake available is folding.
Count the players, not the seat.
Position matters only because it changes how many hands can beat you to the pot. Eight players behind is a different game from two, and your range should move by more than most players are willing to move it.