Push or Fold Poker: the Short Stack Strategy

At some point in a tournament, poker stops being a four-street game. Your stack becomes too small to support a raise plus a c-bet, and every hand reduces to a single question: all in or muck. That simplification is not an admission of weakness, it is the mathematically correct strategy.

This guide gives you the thresholds, the ranges by position, the calling ranges, and above all the adjustments that separate mechanical application from genuinely profitable play: ICM, antes and opponent profile.

Why push or fold exists

Three mechanisms stack up as your stack shrinks.

1. Raising already commits your tournament. At 10 big blinds, opening to 2.2 and c-betting represents half your stack. You are playing a pot you cannot leave, without the benefit of maximum pressure.

2. Fold equity is your main asset. Shoving maximises the chance everyone folds. A partial raise invites 3-bet shoves and strips that weapon from you.

3. Postflop play becomes impossible. With an SPR near 1, there is no postflop decision left to make. You may as well delete the street.

Thresholds by stack depth

The stack that matters is always the effective stack, the smaller of the two in play. If you have 30 big blinds and the big blind has 9, you are playing a 9 big blind spot.

Shoving ranges by position

These ranges are starting points for a tournament with antes at a nine-handed table. Percentages show the share of hands played.

At 10 effective big blinds:

At 15 effective big blinds:

Two immediate readings: position matters more than the hand, and moving from 10 to 15 big blinds roughly halves the ranges in early position.

Calling ranges facing a shove

This is the most neglected half of the topic, and by far the most expensive.

Facing a 10 big blind shove, in the big blind:

The rule that structures everything: your calling range is always tighter than their shoving range, because you have no fold equity. Only one way to win remains: show down the better hand.

Nash ranges and their limits

Nash charts describe the equilibrium of a two-player game where both sides can only shove or fold: neither improves their expectation by deviating alone.

Three limits to know before applying them blind.

1. Nash assumes your opponent plays Nash. If your big blind only calls 12% when they should call 25%, your correct shoving range is far wider than the chart. That is the classic exploitative gap described in the GTO versus exploitative guide.

2. Nash reasons in chips, not money. ICM changes the value of a chip depending on the payout structure. A chip-neutral shove can be a clear money loss near the bubble.

3. Nash stops at two players. With several players behind you, the cumulative chance that one of them calls climbs quickly, and the range tightens.

Use them as a reference point, never as an instruction.

The effect of antes

An ante adds dead money before the cards are even dealt, which raises the reward for stealing.

Concretely, a shove risking 10 big blinds to win 2.5 needs far less equity than one risking 10 to win 1.5. Ignoring the ante leads to playing far too tight in the late stages.

ICM adjustments

Near the bubble or a pay jump, a chip gained is worth less than a chip lost. Two opposite consequences, often confused.

Your calling range tightens sharply. Calling a shove and losing ends your tournament. A call that is chip-correct becomes a clear money loss. On a tight bubble it is not unusual to fold offsuit AJ against a shove that would be an obvious call early on.

Your shoving range can widen. If your opponents feel the same pressure, they defend less often. Your fold equity rises, so the steal becomes more profitable.

The asymmetry is the key point: ICM pressure makes calling expensive, not shoving. It favours whoever takes the initiative, which is equally true in PKO formats, where the bounty adds its own value to eliminating a player.

The 15 to 25 big blind zone

This is the most technical depth, because three options coexist.

Min-raise-fold. You open to 2 big blinds keeping the option to fold to a 3-bet shove. Correct with hands that keep postflop value and lose against a shoving range: medium pairs, suited connectors.

Direct shove. Reserved for hands that dislike being called and benefit from maximum fold equity: weak aces, offsuit K9, marginal hands in late position.

Small blind limp-shove. Against a big blind who attacks limps too often, limping with the intention of calling a raise becomes profitable with strong hands.

The simplifying rule: shove the hands that want everyone to fold, raise the ones that want to see a flop.

Exploitative adjustments by profile

Nash charts assume a theoretical opponent. The real one has a profile.

The chip leader near the bubble is the special case worth remembering: they are the only player free of ICM pressure, so the only one who can call wide. Avoiding shoves into them without a real hand is an adjustment in its own right.

Auditing push or fold spots in GrindLab

Short stack spots are exactly the kind of decision you can validate afterwards, with numbers.

1. Replay a short stack hand in the Lab in tournament mode and set the calling range you were assigning your opponent.
2. Adjust the risk premium to account for the ICM pressure of that moment: the verdict often flips sign between an early spot and a bubble spot.
3. Compare the EV of shoving versus folding, then of calling a shove you faced.
4. Find your late-stage spots through tournament results tracking to see whether your exits come from shoves that were too tight or calls that were too wide.

Classic mistakes

1. Waiting for a hand at 6 big blinds. The cost of waiting exceeds the gain from selection. At that depth your range must be very wide, especially in late position.

2. Calling too wide on the bubble. The symmetrical error, and the most expensive. Elimination costs more than doubling up gains when a pay jump is close.

3. Applying Nash without looking at the opponent. Nash is equilibrium, not exploitation. Against a big blind who folds too much, the correct range is far wider.

4. Forgetting the players behind. A shoving range that is correct against one player becomes far too wide against three still to act.

5. Ignoring antes. Without counting them you play no-ante ranges in an ante tournament, and you get blinded out.

Key takeaways

• Under 10 effective big blinds, pure push or fold is the correct strategy.
• Position matters more than the hand: the button range is three times the UTG range.
• Your calling range is always tighter than their shoving range.
• Antes widen ranges by 5 to 8 points.
• ICM tightens calls far more than it tightens shoves.
• Nash is a reference point, adjusting to the opponent profile is what prints.

GrindLab lets you replay a short stack spot in tournament mode, set a risk premium for the ICM pressure of the moment and see the EV of shoving, calling and folding in one verdict. Try it free during the beta →