Poker Bubble Strategy: Play the Money, Not the Cards
On the bubble, dollar equity and chip equity split hardest. Computed with this site's ICM engine: a mid-stack call that breaks even at ~50% in chips needs 61.5% equity on a standard bubble — and 73% on a satellite bubble, where folding QQ can be correct. The playbook: big stacks attack (nobody can afford to call), medium stacks tighten calls but keep shoving, and short stacks shove before they blind out.
Most tournament mistakes cost a fraction of a buy-in. Bubble mistakes cost whole cashes, in both directions — calling off your tournament with the second-best hand, or folding away a mountain of free chips because everyone else was scared too. This spoke of the tournament strategy hub puts computed numbers on both errors.
The bubble math, computed
Four players, three paid ($500/$300/$200), stacks 40/25/20/15bb. The site's Malmuth-Harville ICM engine values the stacks like this:
| Stack | ICM equity |
|---|---|
| 40bb | $331 |
| 25bb | $261 |
| 20bb | $226 |
| 15bb | $182 |
Now the 20bb stack shoves and the 25bb stack considers calling. Recomputing both outcomes: win and the caller's equity rises from $261 to $373 (+$112); lose and it crashes to $82 (−$179). Break-even equity = 179 / (179 + 112) = 61.5%. In chips the same call needs about 50%. That 11-point ICM tax is the bubble in one number — and it is exactly why the shoverprofits: every fold they force collects the whole pot with no showdown risk. The ratio of loss to gain ($179/$112 ≈ 1.6) is the "bubble factor" — the bigger it is against a particular opponent, the tighter you call.
Big stack: this is your harvest
- ✓ Raise relentlessly into medium stacks. They face the worst risk/reward at the table — the math above says they must fold hands that beat your range, and they know it.
- ✓ Put people all-in; don't get put all-in. Your immunity only works while you are the one applying it. Avoid big confrontations with the other large stack — that is the one player who can revoke your license.
- ✓ Target the players who want to ladder. Watch who folds their blinds without a fight. Some short stacks are gambling to survive; others are white-knuckling into the money. Only the second group is your ATM.
Medium and short stacks: survive without surrendering
- ✓ Medium stack: tighten calls, keep first-in aggression. The 61.5% number applies when you call. When you shove first, fold equity works for you instead — so steal from the players who cannot call, and stop calling the players who can hurt you.
- ✓ Short stack: blinding out is also a decision. Folding into the money works only if an even shorter stack will bust first. If you are the shortest, a shove with fold equity — see the push/fold chart baseline — beats paying blinds until your shove threatens nobody.
- ✓ Count stacks between you and the money. Two shorter stacks at other tables change everything about your risk tolerance. In online tournaments, keep the lobby open on the bubble.
- ✓ Hand-for-hand changes nothing about the math — it only removes the option to stall. If your plan was folding every hand, it still is; if it was shoving on the weak blinds, do it before the big stack re-seats the table order in their head.
The satellite bubble: fold everything, including queens
Satellites pay identical prizes, which sends the bubble math vertical. Computed example: 4 players left, 3 seats worth $1,000 each, stacks 8000/5000/4000/3000. Seat equities: $905, $792, $714, $589. The chip leader considers calling the 5000-stack's shove: winning gains $95, losing costs $260 — a 73% break-even. For scale, QQ against a completely randomhand wins about 80% and AA about 85.1% (site-simulated preflop equities, 300k trials per matchup) — against any realistic shoving range, even queens can be a fold, and the chip leader already owns 90.5% of a seat by folding everything. This is the one spot in poker where "never call" is close to literal. Full mechanics in what is a satellite, and the satellite bubble calculator runs the exact numbers for your stacks.
Bubble fear is symmetric leverage: every dollar of equity the tight players protect is a dollar the aggressive ones farm. If you finish bubbles with the same stack you started them with, you are not "playing it safe" — you are paying the players who aren't. The goal is not to survive the bubble; it is to be the one taxing it.
Where these numbers come from
- • ICM equities and break-evens: computed with the site's Malmuth-Harville implementation (app/lib/icm.js — the engine behind the ICM calculator and satellite calculator) for the exact stacks and payouts stated in each example. Break-even % = loss / (loss + gain).
- • QQ ~80% / AA 85.1% vs a random hand: the site's simulated preflop equity dataset (300,000 trials per matchup with the site's own evaluator).
- • Limits: ICM ignores blinds, position, and skill. Real bubble ranges also depend on who is at your table — these numbers are the floor of the analysis, not the ceiling.
Common questions
What is the bubble in a poker tournament?
The bubble is the stage just before the money — when one more elimination means everyone left gets paid. The player who busts in that last unpaid spot is said to 'bubble.' It is the point of maximum ICM pressure in the whole tournament: the difference between min-cashing and nothing is on the line for every all-in, so correct play deviates furthest from simple chip-count poker.
How should I play on the bubble?
It depends almost entirely on your stack. Big stack: attack — everyone else is mathematically forced to fold, so raise wide and put medium stacks all-in for their tournament lives. Medium stack: tighten your calling ranges dramatically (in our computed example a chip-flip call needs 61.5% equity, not 50%) but keep shoving playable spots where you act first. Short stack: do not blind out — a shove with fold equity beats folding into a guaranteed slow death, unless you can visibly outlast an even shorter stack.
Why is calling worse than shoving on the bubble?
Because the shover wins the pot immediately when everyone folds, while the caller must actually have the best hand at showdown. ICM widens that gap: computed with this site's ICM engine (4 players, 3 paid, $500/$300/$200, stacks 40/25/20/15bb), the 25bb stack calling the 20bb shove gains $112 in equity by winning but loses $179 by losing — a 61.5% break-even, versus roughly 50% in pure chips. The same asymmetry is why big stacks get to shove 'any two' profitably.
How is a satellite bubble different?
It is the same effect at maximum volume, because every prize is identical — first place wins nothing extra. Computed example: 4 players left, 3 identical $1,000 seats, stacks 8000/5000/4000/3000. The chip leader's stack is already worth $905 of a $1,000 seat. Calling the second stack's all-in risks $260 of equity to gain just $95 — a 73% break-even. Even QQ against a completely random hand only wins about 80% (site-simulated), so on many satellite bubbles folding literally everything, including monsters, is correct. Run your own spot in the satellite bubble calculator.
What is a bubble factor?
A shorthand for how much ICM inflates the cost of losing: the ratio of what you lose by losing an all-in to what you gain by winning it. In our computed bubble example the 25bb stack's factor against the 20bb stack is $179/$112 = 1.6 — losses hurt 1.6x more than wins help. In pure chip poker the factor is 1.0; on satellite bubbles it can climb to 2.7+ (ours is $260/$95 = 2.7). The higher your bubble factor against a given opponent, the tighter you call and the more you should be the one shoving.
Tools for the bubble
Keep reading
New to poker? Take the free 15-minute crash course →