Type the surviving stacks and the payout structure into the tool above. It returns each player's slice of the prize pool under the Independent Chip Model — the number that ought to be driving your bubble and final-table decisions.
In a cash game a chip is worth its face value and nothing more. In a tournament that stops being true: the first chip you win is worth more than the last one, because the money is paid by finishing position rather than by chip count. The Independent Chip Model (ICM) is the standard way of pricing that difference. This page explains how the calculator above does its work, walks an example through by hand and shows you where the model gives way. It is one of the free tools supporting our poker strategy guide.
What ICM is measuring
ICM takes two things — the stack of every player still alive and the payouts still to be won — and gives back each player's equity: the average prize money they would walk away with if the tournament were settled many times over from this exact spot, with every chip equally likely to end up anywhere. It knows nothing about skill, position, blind levels or who is about to act. It is purely a function of stacks and payouts.
The Malmuth–Harville model
The model, developed by Mason Malmuth and formalised out of David Harville's horse-racing work, assumes the probability of finishing first equals your share of the chips in play. The probability of finishing second is built up by removing each possible winner in turn: given that player A has won, B's chance of coming second is B's share of the chips excluding A's. Third place repeats that step with two players taken out, and so on down. Each player's equity is then the sum across finishing positions of (probability of that finish × the prize attached to it).
A worked example: three players left, three payouts
Three players are left with 10,000 chips in play and a ₦1,000,000 prize pool paid ₦500,000 / ₦300,000 / ₦200,000. Stacks: A 5,000, B 3,000, C 2,000.
Probability of finishing first
Straight from the chip shares: A = 5,000 / 10,000 = 50%; B = 30%; C = 20%.
Probability of finishing second
Take A: if B wins (30%), A holds 5/7 of the remaining 7,000, giving 0.30 × 0.714 = 21.4%; if C wins (20%), A holds 5/8 of 8,000, giving 0.20 × 0.625 = 12.5%. That totals 33.9%. By the same route: B = 0.50 × 3/5 + 0.20 × 3/8 = 30.0% + 7.5% = 37.5%; C = 0.50 × 2/5 + 0.30 × 2/7 = 20.0% + 8.6% = 28.6%.
Probability of finishing third
Simply whatever remains: A = 100 − 50 − 33.9 = 16.1%; B = 100 − 30 − 37.5 = 32.5%; C = 100 − 20 − 28.6 = 51.4%.
Multiply through by the payouts
A: 0.500 × ₦500,000 + 0.339 × ₦300,000 + 0.161 × ₦200,000 = ₦250,000 + ₦101,790 + ₦32,140 = ₦383,930. B: ₦150,000 + ₦112,500 + ₦65,000 = ₦327,500. C: ₦100,000 + ₦85,710 + ₦102,860 = ₦288,570. The three equities add back up to ₦1,000,000.
| Player | Stack | Chip share | Chip EV (share × ₦1,000,000) | ICM equity | Difference |
|---|---|---|---|---|---|
| A | 5,000 | 50% | ₦500,000 | ₦383,930 | −₦116,070 |
| B | 3,000 | 30% | ₦300,000 | ₦327,500 | +₦27,500 |
| C | 2,000 | 20% | ₦200,000 | ₦288,570 | +₦88,570 |
The chip leader's 50% of the chips is only worth 38% of the money, while the short stack's 20% of the chips is worth 29% of it. That compression is the entire point of ICM: the chips already sitting in front of you are worth less per unit than the chips you are putting at risk, so a confrontation that is neutral in chips is negative in money for whichever stack is bigger.
Where ICM bites hardest
The money bubble
With one elimination left before the money starts, the leap from ₦0 to a min-cash is the biggest single step in the whole structure relative to the stacks. Short stacks should routinely fold hands they would happily shove on chip-EV grounds; big stacks should squeeze, because the medium stacks cannot call without risking a zero. An all-in that is a 55% favourite on the bubble is very often a money-losing call.
Final tables and pay jumps
Every pay jump manufactures another small bubble. With three left and a short stack about to blind out, the two larger stacks should be steering clear of each other almost completely — in the example above, if A and B get it in, the winner gains far less than the loser drops, while C picks up equity doing absolutely nothing. That is ICM pressure, and it is why the standard advice runs: tighten up against stacks that cover you, widen against the stacks you cover.
Satellites
In a satellite where every prize is identical — say ten seats worth ₦1,000,000 apiece — ICM reaches its extreme. Once you hold enough chips to lock a seat, every extra chip is worth precisely nothing and any all-in is an error. Folding pocket aces pre-flop can genuinely be the right play.
ICM set against chip EV
| Chip EV | ICM (money EV) | |
|---|---|---|
| Unit | Chips | Prize money |
| Assumes | Every chip is worth the same | Chip value falls as your stack grows |
| Correct for | Cash games; early tournament stages with flat structures | Bubbles, final tables, satellites, any pay jump |
| Typical error if misused | Calling too wide on the bubble | Playing too tight early, when pay jumps are negligible |
Early in a big-field tournament, with hundreds still playing and the payouts a long way off, ICM and chip EV give you virtually the same answer and you can simply play for chips. The gap opens up as the field thins and the next pay jump grows into a meaningful chunk of the average stack. A practical trigger: once the next pay jump is worth more than roughly 10% of your stack's cash value, run the numbers.
Where the model falls short
- It ignores skill. A stronger player's chips are worth more than the model allows, a weaker player's less. ICM assumes every chip in the pool is equally likely to be won.
- It ignores position and blinds. A short stack on the button with the blinds about to slip past holds more real equity than the same stack posting the big blind next hand. Future-game models try to patch this.
- It slightly overvalues short stacks. Because it assumes chips get redistributed at random, it understates how often the short stack is forced in before the bigger stacks clash.
- It assumes independence. Harville's second-place formula is an approximation known to carry a small bias in horse racing, and that bias travels straight into poker.
- It is only a snapshot. ICM prices the position as it stands right now and says nothing about the next orbit, which is exactly what a shove-or-fold decision is trading on.
Getting the most out of the calculator
Enter every stack still in play
Include the players at other tables — ICM is computed across the whole field, not just the table in front of you.
Enter the payouts still to come
Copy the actual structure from the tournament lobby. Where deals are common in your room, enter the deal payouts instead.
Read off each player's equity
Hold your equity up against your chip share. A wide gap means ICM pressure is high, so tighten against the stacks that cover you.
Test the alternatives
Move the stacks to reflect a call or a fold and watch how your equity shifts. Repeat with different assumptions about how often you win the hand.
Poker strategies
Push/fold ranges, bubble play and final-table adjustments that put ICM to work.
Read the strategy guideHow to play poker
Hand rankings, betting rounds and tournament structure, right from the start.
Start hereCompare poker rooms
Field sizes, rake, software and payout speed lined up side by side.
Compare roomsBankroll management
How many buy-ins tournaments demand, given the variance they carry.
Plan your bankrollFrequently asked questions
What does ICM mean in poker?
Independent Chip Model. It converts tournament chip stacks into each player's expected share of the prize pool, assuming your chance of finishing first equals your share of the chips and then computing the lower finishes recursively. Most bubble and final-table strategy is built on it.
Why is a chip leader's equity below their chip share?
Because the prizes are capped: the winner takes first-place money whether they finish with 51% of the chips or all of them. Chips beyond what is needed to win add nothing, so each additional chip is worth less than the one before. In our example, 50% of the chips is worth only 38% of the pool.
When should I switch from chip EV to ICM?
The moment the next pay jump becomes significant next to your stack: the money bubble, final-table pay jumps and satellites. Early in a large tournament the two models agree closely enough that you can just play for chips.
How accurate is ICM?
It is a useful approximation rather than the literal truth. It disregards skill, position and blind levels, tilts slightly in favour of short stacks, and leans on Harville's second-place formula, which carries a known small bias. Newer future-game simulations refine it, but ICM is still the standard starting point.
Can I use this calculator to work out a final-table deal?
Yes. Enter the stacks and the remaining payouts, and the equities it returns are the standard ICM chop figures. Many rooms offer an ICM deal option; set it against a chip-chop, which simply divides the pool by chip share and therefore favours the big stacks.

