ICM Poker: The Independent Chip Model Explained
Last updated: May 11, 2026
ICM (Independent Chip Model) converts your tournament chip count into real dollar equity — and the result is always less than a proportional share of the prize pool. A player holding 50% of all chips in a 3-way tournament with a $1,000 prize pool does not own $500 in equity; ICM typically yields closer to $420–$460 depending on the pay structure. Understanding this non-linearity is the foundation of correct tournament strategy from the bubble through the final table.
What Is ICM in Poker?
ICM — the Independent Chip Model — is the mathematical system that translates tournament chip counts into real dollar equity. The core insight is simple but counterintuitive: in a tournament, chips are not money. A chip is worth different amounts depending on how many you have relative to the field and where those positions sit in the prize structure.
In a cash game, every chip has a fixed dollar value. Buy in for $100, and each chip represents a fixed fraction of that $100. In a tournament, the prize pool is split across finishing positions — typically 10–15% of the field cashes, with amounts concentrated at the top. A player who doubles their chip stack does not double their prize equity, because they were already in contention for the top prizes and adding chips beyond a certain point cannot increase the 1st-place prize they can win.
Mason Malmuth rediscovered and applied ICM to poker in 1987, and it became the standard framework for evaluating tournament decisions, especially those involving all-in confrontations, final table deals, and bubble play. Today, every serious tournament solver — from ICMIZER to HRC — uses the ICM framework as its foundation.
The practical consequence is that winning chips is less valuable than it appears, and losing chips is more costly than it appears. A player going from 10,000 to 20,000 chips gains less equity than a player going from 10,000 to 5,000 loses. This asymmetry — known as the diminishing marginal value of chips — is why tournament players fold spots that would be clear calls in a cash game.
The ICM Formula — How Chip Stacks Become Dollar Equity
The ICM formula computes each player's prize equity by summing the probability of finishing in each paid position multiplied by that position's prize. The probability of finishing 1st equals your chip count divided by total chips in play. Probabilities for 2nd place and below are calculated iteratively — by assuming each possible 1st-place finisher, removing that player's chips, and recalculating the process recursively.
ICM Equity Formula
Equity(player i) = Σₖ P(player i finishes in position k) × prize[k]
P(i finishes 1st) = chips[i] ÷ total_chips
P(i finishes 2nd) = Σⱼ≠ᵢ [P(j finishes 1st) × chips[i] ÷ (total − chips[j])]
Worked Example — 3-Player Tournament
Stacks: A = 3,000 chips | B = 2,000 chips | C = 1,000 chips (total: 6,000)
Prizes: 1st = $500 | 2nd = $300 | 3rd = $100 (total: $900)
- P(A 1st) = 3000 ÷ 6000 = 50%
- P(A 2nd) = P(B 1st) × 3000/4000 + P(C 1st) × 3000/5000
- = 0.333 × 0.75 + 0.167 × 0.60 = 25.0% + 10.0% = 35.0%
- P(A 3rd) = 1 − 50% − 35% = 15%
- ICM equity A = 0.50 × $500 + 0.35 × $300 + 0.15 × $100
- = $250 + $105 + $15
- ICM equity A = $370
- ICM equity B ≈ $311 | ICM equity C ≈ $219
Player A holds 50% of chips but only 41% of prize equity ($370 of $900). This is the core distortion ICM measures — chips and dollars are not proportional.
The larger the prize differential between positions, the greater the ICM distortion. A winner-take-all structure produces less ICM distortion than a structure where 1st pays 10× the minimum cash. Understanding the prize structure of every tournament you enter is essential for calibrating how much ICM pressure to apply to your decisions.
ICM Pressure: When It Peaks and Why
ICM pressure peaks at the bubble — the moment when exactly one more elimination will pay everyone still at the table. In a 100-player tournament with the top 10 paid, the bubble occurs at 11 players remaining: 10th place earns the minimum cash while 11th place earns $0. The gap between making the money and missing it is the single most powerful ICM leverage point in any tournament.
During the bubble, players on a short stack can fold their way to the money even with as few as 5–6 big blinds. A player with 7 BBs folding pocket tens on the bubble is not a mistake — it is mathematically correct when the chip leader is applying all-in pressure and busting means $0. ICM pressure also creates significant leverage at every pay jump, not just the initial cash bubble.
Pay jumps at 50th place, 27th, 18th, 9th, and the final 3 all create secondary ICM pressure spikes. Skilled tournament players track the remaining players relative to pay tiers throughout the event — not just at the initial money bubble — and adjust their aggression level accordingly at each threshold.
How to Adjust Your Strategy for ICM
The primary ICM strategic adjustment is tightening calling ranges dramatically while widening shoving ranges if you are the chip leader. These two adjustments move in opposite directions and reflect the asymmetric nature of ICM pressure across different stack sizes at the same table.
Consider the bubble call: you hold 50 big blinds and a 100 BB chip leader shoves all-in. In a cash game at 50% equity you would break even. Under ICM, you might need 60–65% equity to call profitably because busting costs you not just your chips but your entire remaining prize equity — while folding costs only the current hand and preserves your tournament life and minimum cash. The same logic applies to every pay jump — any elimination that moves you up a prize tier is worth ICM equity even without winning a chip.
Tighten calling ranges near pay jumps
The closer you are to a pay jump, the stronger your hand needs to be to call an all-in risk. Near the bubble, hands like AJo or 99 that are standard calls in cash games become folds against many all-in shoves. ICM calculators quantify exactly how tight: in a 9-player SNG, calling off your 30BB stack on the bubble typically requires hands in the top 10–15% of all holdings.
Chip leaders should shove wider, not tighter
As the chip leader near the bubble, your ICM pressure is near zero — you cannot bust in one hand. This means you can shove a very wide range of hands and force shorter stacks into difficult ICM decisions. A chip leader can profitably shove any two cards against a 10–15 BB stack that is actively trying to fold into the money. Position matters: shoving from the button and small blind applies maximum pressure.
Factor in the pay jump size, not just survival
A $500 pay jump from 5th to 4th place is worth more ICM equity than a $100 jump from 15th to 14th. Adjust your calling threshold proportionally to the size of the pay jump at risk. In satellite play, every pay jump may be equal (all seats are the same value), which creates the most extreme ICM adjustments of any format.
Position amplifies ICM adjustments
Being in late position near the bubble means you have information about who has folded before you act — and the option to isolate short stacks or steal blinds with less risk of running into a premium hand. Early position requires wider folding ranges than late position because you face more potential callers, each of whom could have a strong hand.
A useful mental shortcut: every time you are considering calling an all-in in a tournament, ask “what is my equity if I fold?” That folding equity — the ICM value of your stack with no risk — sets a floor. Only call when the expected value of calling, probability-weighted, exceeds that floor by a meaningful margin.
Chip EV vs. ICM EV: The Critical Difference
Chip EV treats every chip as equal and ignores the prize structure entirely. ICM EV converts chip counts into dollar equity at every decision point. The difference between these two frameworks explains why tournament strategy diverges fundamentally from cash game strategy — even when the cards are identical.
The clearest illustration is the 50/50 coin flip for all chips. In chip EV terms, the flip is exactly break-even. In ICM EV terms, it is always negative for the player who is not desperate:
3-Player SNG Example — $500 / $300 / $100 Prizes
Player A (3,000 chips) flips vs Player B (3,000 chips)
Chip EV: break-even (3,000 chips either way on average)
ICM EV if A wins flip: A has 6,000 chips → ICM equity ≈ $460
ICM EV if A loses flip: A busts → ICM equity = $100 (3rd place)
Expected ICM EV = 0.50 × $460 + 0.50 × $100 = $280
Pre-flip ICM equity with 3,000 chips ≈ $311 → flip costs ~$31 in EV
The flip costs approximately $31 in expected prize equity — despite being chip-EV neutral. This loss comes from the asymmetry: winning 3,000 chips gains $149 in ICM equity ($460 − $311), but losing 3,000 chips loses $211 in ICM equity ($311 − $100). The 50/50 odds cannot compensate for this lopsided dollar outcome.
The practical implication is that maximizing chips is not the goal in tournament poker. Maximizing prize equity is. These objectives align early in a tournament when ICM distortion is small, but diverge sharply near the bubble and final table where finishing position differences translate to large dollar amounts.
Final Table ICM Dynamics
At the final table, ICM distortion reaches its maximum. The chip leader holding 50% of all remaining chips does not hold 50% of the remaining prize pool in equity — in a typical 9-handed final table structure, they might hold 35–40% of prize equity depending on the pay jumps between positions 1 through 9.
This is why final table deals — also called ICM chops — are common in live poker. Players calculate their current ICM equity and negotiate a cash settlement that reflects stack sizes more fairly than simply playing out the remaining hands. A 3-handed deal with stacks of 60% / 25% / 15% of chips might yield payouts of $12,000 / $9,000 / $7,000 from a $28,000 remaining prize pool, rather than $16,800 / $7,000 / $4,200 implied by a proportional split.
Heads-up play also carries ICM nuance. With 2 players remaining and only 1st and 2nd place prizes left, ICM reverts to near-linear behavior — chip counts now map almost proportionally to equity because the only uncertainty is which of the 2 players finishes first. A 70/30 chip split translates to roughly 70/30 equity split for 2 players, making heads-up poker closer to chip-EV play than any other tournament phase.
3-Handed ICM Deal Rule of Thumb
When negotiating a 3-way deal, the short stack almost always benefits from dealing out (they get their ICM equity guaranteed), while the chip leader often benefits from playing on (they can apply maximum pressure). Knowing ICM equity numbers before deal discussions prevents leaving thousands of dollars on the table in negotiations.
Satellites and Winner-Take-All: Special ICM Cases
Satellites represent the most extreme ICM scenario in poker. In a satellite, every player who finishes in a paying position receives an identical prize — usually a seat in a larger tournament worth a fixed dollar amount. The bottom of the money and the top of the money receive the same prize. This eliminates any incentive to accumulate chips once you have a safe stack.
Consider a 10-seat satellite where 10 players receive a $1,000 tournament entry. If you have 15 big blinds when 11 players remain (1 away from the money), your correct strategy is to fold every single hand — including aces — unless you are forced to post and someone with fewer chips than you has already committed to a pot. With 11 players and 10 seats, you are already mathematically a 90% favorite to cash without risking a single chip.
The ICM math is stark: your 15BB stack is already worth approximately $900 in ICM equity (90% chance of cashing a $1,000 seat). Calling an all-in — even with 60% equity — risks losing $900 for a gain of zero, because doubling your stack adds no prize value whatsoever. The only prize that exists is the seat, and you almost have it locked up.
Winner-take-all structures (rare in practice but theoretically important) produce the opposite: minimal ICM distortion. When only 1st place is paid, the prize equity equals chip EV exactly, because each chip represents an equal share of 1st-place probability regardless of stack size. Standard tournaments fall between these extremes — which is why studying the specific prize structure before every tournament you enter is fundamental to correct ICM-adjusted play.
Definitions
Frequently Asked Questions
What is ICM in poker?
ICM stands for Independent Chip Model — a mathematical framework that converts tournament chip stacks into real dollar equity based on the prize pool distribution and stack sizes at the table. Unlike cash games where every chip is worth exactly its face value, tournament chips have non-linear value because prize pools are top-heavy. A player with 50% of all chips does not win 50% of the prize pool; they win a smaller share because they can only claim 1st place, not every prize simultaneously. Mason Malmuth rediscovered and popularized ICM for poker in 1987, adapting it from earlier combinatorial work. As a simple example, consider a 3-player tournament with equal stacks and a $300 prize pool ($150/$100/$50). Before a single hand is played, each player holds exactly $100 in ICM equity — one-third of the prize pool — regardless of chip count. ICM is used exclusively in tournament formats: multi-table tournaments (MTTs), sit-and-gos (SNGs), and satellites. It never applies to cash games.
How does the ICM formula work?
The ICM formula calculates each player's prize equity by summing over all finishing positions the product of their probability of finishing in that position and the corresponding prize. Probability of finishing 1st is simple: your chip count divided by total chips in play. Finishing probabilities for 2nd, 3rd, and beyond are iterative — you remove each possible 1st-place finisher, recalculate the remaining stacks, and recurse. Here is a worked 3-player example: stacks of 3,000 / 2,000 / 1,000 (total 6,000 chips), prizes of $500 / $300 / $100. Player A (3,000 chips): P(1st) = 3000/6000 = 50%. P(2nd) = P(B wins 1st) × 3000/3000 + P(C wins 1st) × 3000/5000 = 0.333 × 1.0 + 0.167 × 0.6 = 0.333 + 0.100 = 43.3%. P(3rd) = 1 − 0.50 − 0.433 = 6.7%. ICM equity for A = 0.50 × $500 + 0.433 × $300 + 0.067 × $100 = $250 + $130 + $6.70 = $386.70. The calculation confirms that A's 50% chip share converts to only about 43% of the $900 total prize money — not 50%.
When does ICM pressure peak?
ICM pressure peaks at the bubble — the spot in a tournament where one more elimination will pay everyone remaining. The bubble creates the starkest possible asymmetry: busting out on the bubble means $0, while surviving even one more hand guarantees at least the minimum cash prize. In a 100-player tournament with the top 10 paid, the bubble occurs at 11 players remaining. The player in 11th place gets absolutely nothing if they bust, while 10th place might earn $200 or more. This asymmetry means a short stack with 5–6 big blinds can profitably fold their way to the money even holding hands as strong as pocket queens, because the ICM value of cashing exceeds the chip EV of calling. ICM pressure also spikes at every significant pay jump — when 5 players are left and 4th place pays $500 but 5th pays $200, for example, that $300 pay jump creates an ICM incentive to tighten dramatically. Big stacks at the bubble face almost no ICM pressure and can apply maximum leverage on middle stacks, who feel it most acutely.
How should I adjust my strategy for ICM?
The core ICM strategy adjustment is tightening your calling ranges dramatically near pay jumps and the bubble, while simultaneously recognizing that shoving ranges can actually widen for chip leaders who face no ICM pressure. As a concrete example: imagine you hold 50 big blinds at the bubble and a 100 BB chip leader shoves all-in. In a standard cash game, you would need roughly 50% equity to call — a break-even decision. Under ICM pressure, you might need 60–65% equity or more to call profitably, because busting means $0 while folding preserves your tournament life and the minimum cash. Your calling range might tighten from any pair and most suited connectors to only premium hands like AK, QQ+. Conversely, if you are the chip leader, you can shove extremely wide because your opponents cannot call you without risking their tournament life. Pay-jump awareness matters throughout the tournament: every time a player busts and you move up a prize tier, a small equity boost occurs even without winning a single chip. Think of this as the “ICM tax” on calling — any call that risks elimination costs you more in ICM terms than a pure chip EV analysis suggests.
Does ICM apply to cash games?
No — ICM has absolutely no application to cash games. In a cash game, every chip represents a fixed dollar amount: a $1 chip is always worth exactly $1, a $5 chip is always worth $5. You can buy in, cash out, add chips, or leave at any time, and the value of your stack is perfectly linear. There is no prize pool, no pay structure, and no finishing positions to consider. ICM only applies when a tournament structure creates non-linear chip values through a prize pool that rewards finishing positions differently. Multi-table tournaments pay a top percentage of the field with amounts that decrease as finishing position drops. Sit-and-gos pay out 1st, 2nd, and sometimes 3rd. Satellites award seats to a larger tournament to everyone who finishes above a certain threshold. All of these formats create ICM dynamics because surviving longer is independently valuable beyond the chips you accumulate. Cash game players who study tournament poker must consciously learn to think in ICM terms — the chip-EV instincts developed in cash games actively work against correct tournament decision-making near pay jumps.
What is the difference between chip EV and ICM EV?
Chip EV (expected value calculated in chips) treats every chip as equal and ignores the prize structure entirely. ICM EV converts chip counts to dollar equity at every decision point, accounting for the diminishing marginal value of additional chips in a tournament. The clearest illustration is a 50/50 coin flip for all chips in a tournament. In chip EV terms, the flip is exactly break-even: you expect to end up with the same number of chips you started with on average. In ICM EV terms, the flip is always negative EV for the player who is not desperate. Here is why: suppose you have 10,000 chips and your opponent has 10,000 chips in a 2-player SNG paying $100 to 1st and $0 to 2nd. Your ICM equity is exactly $50 (50% of 1st place). If you win the flip, your equity jumps to $100 — a gain of $50. If you lose, your equity drops to $0 — a loss of $50. So the flip is break-even in this heads-up case. But add a 3rd player: now winning the flip does not give you $100 in equity because other finishing positions still exist. You gain less from winning than you lose from busting. This asymmetry — losing chips always hurts more than gaining an equal number helps — is the defining characteristic of ICM and explains why tournament poker demands a fundamentally different strategic framework than cash games.
How does stack size affect ICM pressure?
Stack size determines both the magnitude of ICM pressure you face and the strategic options available to you. Big stacks (above average) experience minimal ICM pressure because they can absorb a loss without busting — they have the freedom to gamble, apply pressure on shorter stacks, and accumulate chips without existential risk. Short stacks (under 10 big blinds) experience a different kind of relief from ICM pressure: they are so desperate that chip EV and ICM EV nearly converge — when you are about to blind out anyway, calling a shove at 40% equity may actually be correct. Middle stacks experience the most severe ICM pressure. They have enough chips to fold into the money but not enough to absorb a big loss, and they cannot afford to call all-ins from big stacks without risking elimination. This creates what players call “payoff asymmetry” — the middle stack wins relatively little in ICM terms from accumulating chips (because they were already somewhat safe) but loses enormously from busting (because they were close to cashing or moving up a pay tier). The concept of the “ICM tax” applies most heavily to middle stacks: every call they make costs them extra in ICM terms beyond the pure chip risk, which is why tight, selective play from middle-stack positions near the bubble is almost always the correct ICM-adjusted strategy.
Related Guides
Calculate your real tournament equity
Use RiverOdds to get exact hand equity for any all-in decision — then apply ICM pressure thresholds to make the right call every time.
Try the RiverOdds Calculator →