ICM for micro-stakes players
Updated Nov 13, 2026
ICM — the Independent Chip Model — converts everyone's stack into their realistic share of the prize pool. Near the money, chips and money stop being the same thing, and ICM says one thing over and over: losing chips hurts more than winning the same chips helps. A double-up never doubles your money, and a bust-out takes all of it.
By Roger Irwin (Rutland34), FindPokerTournaments — put together with help from other players and the training resources linked on this page.
📚 Keep learning
How it works
Early in a tournament, 1,000 chips is just 1,000 chips. On the bubble, 1,000 chips might be the difference between cashing and going home with nothing — and the next 1,000 chips you win are worth much less than the 1,000 you would lose. So a call that is easy math in a cash game ("I only need 47% equity, I probably have 55%") can be lighting money on fire late in a tournament, because the pot odds are only half the calculation. The other half is what your stack is already worth, sitting still, doing nothing. You do not need a solver subscription to use this. You need one worked example and the habit of asking "what does folding keep?" before you ask "what might calling win?"
Four left in an $11 event, three paid $55, $33 and $22. You hold A♠ K♠ in the big blind against the chip leader's shove. Pot odds say you need 47%, and AKs has about 55% (estimate). But folding keeps a stack worth $29.27 in prize money, while calling and winning only raises it to $42.30 — so the call needs 69% to break even, and burns roughly $6 every time you make it. The full math is worked step by step below.
The worked example: an $11 bubble, with real numbers
Four players left in an $11 event. Three get paid: $55, $33, $22 — fourth gets nothing. Small field, simple numbers, same shape as the nightly micros you actually play.
Stacks: the chip leader has 6,000. You have 4,000 — second stack, and you're in the big blind, with 200 of it already in. The small blind has 2,900 after posting 100, and the short stack has 2,000.
The chip leader shoves. You look down at A♠ K♠.
By cash-game math this is not a decision. You have to call 3,800 more into a pot that will be 8,100 — you need 47% equity to break even, and AKs against a big stack's bubble shoving range is around 55% (estimate — nobody knows a stranger's exact range, but that's the neighborhood). Snap call, right?
Now run the money.
If you fold, your 3,800 remaining chips are worth $29.27 in expected prize money. Not $40, not "second place money" — $29.27, because ICM shares the pool by everyone's stack and finish chances. That's what you're protecting.
If you call and win, you have 8,100 chips — your stack more than doubles. Your money does not: you're now worth $42.30. Your chips went up 102%. Your money went up 45%. Read that again, because it's the whole guide in one sentence.
If you call and lose, you're out in fourth. $0.
So the call is a coin you're buying with $29.27 to win an extra $13.03. You need to win it 69% of the time just to break even. AKs wins about 55%. Calling here doesn't cost you a little — the call is worth about $23 in expectation against a fold worth $29.27. Every time you make this call, you burn roughly six dollars — more than half the buy-in — while feeling like you played it brave.
That's the risk premium: the gap between the 47% the pot odds demand and the 69% the money demands. Twenty-two points of equity, appearing out of nowhere, because fourth place pays zero and first place doesn't pay infinity. This is an extreme spot — a stone bubble, where the premium is at its biggest — but the premium is present, smaller, in every late-tournament call you make.
Why the medium stack gets it worst
Look at the same hand from the short stack's chair. The 2,000-chip player faces the same shove. If he folds, his stack is worth $18.49 — he's in real danger of blinding out in fourth anyway. If he calls and wins (pot 4,300), he's worth $30.09. His breakeven is 61% — a big premium over the 47% pot-odds number, but nowhere near your 69%.
That's the medium-stack squeeze, and it's the cruelest seat in tournament poker. The big stack has chips to spare. The short stack has little to protect — folding doesn't save him, so he can call closer to the pot-odds math. You, the second stack with two shorter stacks at the table, have the most money to lose and the least reason to gamble: if you stay out of the way, somebody else probably busts first. The big stack knows this. That's why he's shoving. His range against you can be wide — when he loses an all-in against you he still has chips and a live stack; when you lose one against him, you have fourth place and a story. He risks a slice of his equity. You risk all of yours. When a big stack leans on you near the bubble and you're not the short stack, the correct response is usually a tighter calling range than your instincts scream for — and turning your own aggression up against the other medium stacks, who are just as squeezed as you.
One more consequence: ICM pressure doesn't start at the final table. It starts building when the field gets to roughly 20% above the money — 120 players left with 100 paid, that kind of neighborhood — and it peaks on the exact bubble. If your calling standards don't change at all between 200 left and 4 left, they're wrong somewhere.
Final tables at micro stakes: the jumps are $40, not $40,000
Everything above was a bubble. The final table is the same math in a softer voice, because now every bust-out pays something and the question becomes the size of the jumps.
Take a typical $22 nightly final table — say a $990 prize pool paying five spots $420 / $255 / $165 / $95 / $55 (check your own lobby; the shape is what matters). Climbing from fifth to fourth is worth $40. Fourth to third is $70. These are real dollars, and laddering one spot is never nothing — but compare the TV version, where a pay jump is a house. At micro stakes, players over-ladder: they fold for an hour to climb a $40 rung while the blinds saw their stack in half, and arrive at fourth place with six big blinds and no fold equity. The jump from fifth to fourth is $40; the difference between arriving at three-handed with 25 big blinds versus 8 is worth far more than $40 in every direction that matters. Ladder when a shorter stack is about to bust — that part is free. Don't pay for the ladder with your whole stack's future.
How to actually use this at the table
You're not going to run an ICM calculation mid-hand, and nobody is asking you to. What changes is the order of the questions. Before a late-tournament call, run these three:
- What does folding keep? If shorter stacks are about to eat the blinds, your stack is worth real money standing still. The closer your stack is to "probably cashes by folding," the more a call has to offer.
- Who busts if I wait? If the answer is "very likely the short stack, within an orbit," your call needs to be a monster, not a math edge. If the answer is "nobody — I'm the one blinding down," the premium shrinks and you play closer to pot odds.
- What does winning actually buy me? Not "double my chips" — doubling your money almost never happens. Winning buys you a bigger share and a better seat to attack from. Price it as that, not as a lottery ticket.
That's the whole skill. The players who look fearless at final tables aren't ignoring this math — they've just done the three questions so many times they answer them in two seconds.
How PKOs change it
Everything on this page assumes the only money is in the prize pool. In a progressive knockout, part of the money is sitting on heads, and bounty money is different: it's paid immediately, in cash, and it doesn't care about pay jumps. A bounty worth two buy-ins attached to a short stack can flip a clear ICM fold into a clear call, because the bounty pays you in dollars, not in tournament equity. The adjustments are small early (a starting bounty is worth about a quarter of the buy-in) and huge late, when a big bounty can outweigh a pay jump entirely. The full method — convert the bounty to chips, add it to the pot, re-run the equity you need — is in our PKO guide. The short version: ICM tightens you up; bounties loosen you back, by an amount you can calculate.
What ICM can't tell you
ICM is a model, and it has blind spots you should know about before you worship it:
- It assumes everyone is equally skilled. If you genuinely outplay your table, you're worth more than the model says, and gambling a little more than it advises is defensible. At micro stakes, fields are soft — but be honest about how much of that edge is real and how much is a story you tell yourself on the bubble.
- It doesn't see the future. Position, who's about to take the blinds, the maniac two seats left — ICM knows none of it. Use it to set your ranges, then let the table adjust them.
- It doesn't cover deals. At a real final table with a deal on the table, the math changes completely — that's a negotiation, not a model output.
- It fades heads-up. Two-handed, there are no more jumps to protect, and it's back to chips = money. Play cards again.
- It is not an excuse. "ICM fold" said about every playable hand for three hours is just folding. The model punishes marginal gambles. It does not tell you to fold the nuts, and premiums shrink dramatically once you're in the money or a short stack is about to bust.
Strategy tips
- Ask "what does folding keep?" before "what might calling win?" — your stack has a money value standing still.
- When a big stack leans on you near the bubble and you are not the short stack, tighten your calling range and do your own bullying downward, at the other medium stacks.
- ICM pressure starts around 20% above the money, not at the final table. Adjust your calling standards before the bubble, not on it.
- At micro final tables, ladder when a shorter stack is about to bust — but never pay a $40 jump with your whole stack's future.
- Remember that a double-up never doubles your money. Price a call by what winning actually buys: a bigger share and a better seat to attack from.
- In PKOs, price the bounty in dollars before you fold: a big bounty can flip an ICM fold into a clear call.
Common mistakes
- Calling by pot odds alone near the bubble. If your only math is "I need X% and I have Y%," you are playing cash poker in the part of the tournament where cash poker bankrupts people
- Thinking a double-up doubles your money. It never does — that is the whole point
- Being the medium stack who "won't be bullied." Your pride is exactly what the big stack is shoving into
- Over-laddering at micro final tables. A $40 jump is not worth blinding down to a push-fold stack
- Applying bubble ranges in the money, or money ranges on the bubble. The premium is biggest exactly on the bubble and shrinks fast on both sides of it
- Ignoring the bounty half of a PKO. Pure ICM in a PKO folds too much, the same way pure pot-odds thinking in a freezeout calls too much
🎯 Sharpen your game
Hand-picked places to get better at this format. Free stuff first. Community marks working players, coaches and smaller creators worth supporting. Links checked Nov 5, 2026.
📖 Free articles & videos
- ICM Explained: Tournament Poker Endgame Strategy ↗ Article
ThinkGTO · The same model in more depth, with a bubble section and a common-mistakes list that matches what actually happens at the micros. - What is ICM? ICM Poker Strategy with Fedor Holz ↗ Article
PokerCode · Good on ICM's limits: "a concept, not a solution" is the right attitude, from a player who has made these calls for real money. - The Mathematics of Bounties ↗ Article
PokerStars Learn · The PKO side of this page, done rigorously: how a bounty changes the equity you need, with worked examples. - GTO Wizard: MTT (playlist) ↗ Video
GTO Wizard · Final tables, ICM and late registration, worked through hand by hand; the fastest way to see what the premium does to real ranges.
📚 Books & audiobooks
- The Mathematics of Poker ↗
Bill Chen & Jerrod Ankenman · 2006 · The math under ICM and everything else on this page. Not light reading; it's the source the summaries are summarizing. - Secrets of Professional Tournament Poker ↗ 🎧
Jonathan Little · 2021 · Endgame and bubble play in plain language, from a tournament regular — the readable version of most of this page.
🔎 Tools
- SharkScope ↗affiliate link (affiliate link)
Look up the players left at your table: ability ratings and finish distributions tell you who is likely to protect a stack and fold to pressure, and who is calling off. How to use it
𝕏 Players to follow on X
- Dara O'Kearney @daraokearney ↗
Player on X · Wrote the book on satellite strategy and posts constantly about ICM spots at exactly these stakes; the short-stack end of the squeeze, explained by the guy who lives there.
Find these tournaments
All tournaments → PKO and bounty tournaments → Sunday majors →
FAQ
What is ICM in poker?
ICM — the Independent Chip Model — converts everyone's stack into their realistic share of the prize pool. Near the money, chips and money stop being the same thing, and ICM says one thing over and over: losing chips hurts more than winning the same chips helps. A double-up never doubles your money, and a bust-out takes all of it.
What is ICM in simple terms?
A way of valuing tournament chips in real money. Because busting pays nothing and doubling your stack never doubles your money, losing chips costs more than winning the same chips gains — so late-tournament calls need more equity than pot odds alone suggest.
Do I need a solver to use ICM?
No. The three questions in this guide — what folding keeps, who busts if you wait, and what winning actually buys — cover most real spots. Solvers and trainers are for checking your guesses afterwards; see the Study page.
Does ICM apply in PKOs?
Only partly. Bounties pay cash immediately and ignore pay jumps, so a large bounty loosens correct calling ranges back up. The PKO guide has the conversion method.
Related guides
Freezeout · Deepstack · PKO (progressive knockout) · Satellite, step and feeder · Bankroll management · Using a randomizer · Bubble play: the folds, the punts, and when to go for it
Play responsibly. Tournaments are entertainment, not income. Set a budget and time limit, and never chase losses. 18+/21+ depending on site and state. Help: 1-800-GAMBLER.
18+/21+ only. Minimum age depends on the site and your state. Play responsibly. Gambling problem? Call 1-800-GAMBLER.
Affiliate disclosure: Some links may be affiliate links, and we may earn a commission if you sign up. That doesn't change the listings or the order they appear in.
Not legal advice. Online poker and sweepstakes laws vary by state, so check your local laws before you play. Of the sites we track, only PokerStars (US) on FanDuel, BetMGM Poker (US) and WSOP Online are state-licensed in the US, and only in the states they list (NJ/PA/MI, NJ/PA/MI and NV/NJ/MI/PA). Other state-licensed US rooms exist that we don't track yet (for example BetRivers). Offshore and crypto sites (Ignition, ACR, BetOnline, CoinPoker, Phenom) are not licensed in the US. GGPoker doesn't accept US players. CoinPoker's own terms list the US as a Restricted Territory, even though US players currently play there. Sweepstakes sites exclude some states.
Schedules come from each site's public pages, and each listing shows its source and the date we last checked it. Items marked unverified conflict across sources or are out of date. Always confirm in the poker lobby. Stake tiers use face value: a 110 SC or 110 CHIP buy-in tiers alongside a $110 buy-in. SC and CHIP are not cash — won SC is redeemed under each room's sweepstakes rules.