ICM Calculator

Convert tournament chip stacks to dollar equity using the Independent Chip Model. Use it to evaluate final-table deals or analyze risk-of-elimination spots.

Chip Stacks

Total: 12,000
Player 141.7%
Player 233.3%
Player 325.0%

Payout Structure

1st50%
2nd30%
3rd20%
Total: $1,000

Tournament Equity

Player 1$359.72
36.0% of pool-5.7% vs chip share
Player 2$334.92
33.5% of pool+0.2% vs chip share
Player 3$305.36
30.5% of pool+5.5% vs chip share
ICM compresses big stacks down and big-blind-survival up. Small stacks gain equity vs their chip share — that's why short stacks should call wider near the bubble.

When to use ICM

  • Negotiating a final-table chop — "chip-chop" is unfair to short stacks
  • Estimating risk premium for shoves near the bubble
  • Comparing two stack distributions to see who's actually winning
  • Studying optimal calling ranges in late-stage MTTs

Computed via the Malmuth-Harville model. Doesn't account for skill differences, position, or future blinds — it answers "what's the dollar value of these chipsright nowassuming everyone plays equally?"

Guide

ICM in poker — what it is and when it matters

ICM is the math that separates serious tournament players from amateurs. Get this right and you'll fold profitable cash-game spots, push profitable tournament-only spots, and negotiate chops at the right number. Below: the math, the intuition, the situations where ICM bites hardest, and the common mistakes that cost players final tables.

What ICM actually means

ICM stands for Independent Chip Model. It's the math that converts your tournament chip stack into its expected dollar (or buy-in) value at the current point in the tournament.

The key insight: tournament chips don't have linear value. In a cash game, a $100 stack and a $200 stack are exactly 2x apart in dollar terms. In a tournament, a 100-chip stack and a 200-chip stack are NOT 2x apart in dollar value because of the prize structure. Doubling your chips never doubles your equity in the prize pool.

Why? Because once you bust, you're out — you can't lose more than your entry. The prize pool is bounded, but losing chips is unbounded. That asymmetry means each chip you have is worth slightly less than the previous one. ICM math captures exactly how much less.

The Malmuth-Harville model

The most-used ICM formula is the Malmuth-Harville model. It calculates each player's expected dollar value as the sum across all possible finishing positions, weighted by the probability of finishing in each spot.

Probability of finishing 1st = (your chips) / (total chips). The probabilities of finishing 2nd, 3rd, etc. cascade conditionally — you place 2nd if someone else wins first, you place 3rd if two others busted before you, etc.

The model assumes no skill difference. Everyone plays equally well. Real-world ICM is a bit different — better players have higher EV than the model predicts — but the calculator gives you the no-skill baseline, which is the right number for chop negotiations.

Bubble pressure (where ICM bites hardest)

On the bubble of a tournament (one spot before the money), ICM creates massive pressure on short and medium stacks. Here's why:

  • Short stacks have nothing to lose AND everything to lose. Going broke = $0. Surviving the bubble = at least min-cash. Even bad short-stack hands gain value because elimination is so costly relatively.
  • Medium stacks must fold premium hands. A standard cash-game shove with KK becomes a fold against a chip leader because risking elimination on the bubble has higher dollar cost than the chips won.
  • Big stacks bully without resistance. They can shove on bubble medium stacks knowing the medium stack must fold even premium hands. Real EV for the chip leader.

Practical example:4 players left, 3 paid. Chip leader 50%, two medium stacks at 20% each, short stack 10%. The medium stacks should fold AKs to a chip-leader shove — risking elimination has more dollar cost than the chips gained. That's the bubble effect.

Final-table chops (negotiating the right number)

When 3-9 players are left at a final table and the prize pool is significant, players often negotiate a chop (deal). ICM is the math behind "fair" chop offers.

Three types of deals you'll see:

  1. 1. Pure ICM chop: remaining prize pool divided by ICM equity. Each player gets exactly their ICM dollar value. No room for negotiation; the math decides.
  2. 2. ICM with a save: ICM-based chop, but a small chunk (often 10-20% of prize pool) is left to play for, so the eventual winner gets a bit more.
  3. 3. Chip-chop: proportional to chip stack, ignoring payout structure. This always favors big stacks unfairlyand is a bad deal for short stacks. Don't accept chip-chops unless you're the chip leader.

When a deal is offered, run it through this calculator. If the offer matches your ICM equity, consider accepting (variance reduction is real). If it's below, counter or refuse.

Common ICM mistakes

  1. 1. Not adjusting calling ranges on the bubble. Many players call shoves with hands like AJ, 99 in cash-game-correct spots. On the bubble these are clear folds. ICM tightens calling ranges by 10-30% depending on stack distribution.
  2. 2. Accepting chip-chop deals as a short stack.Chip chops always favor big stacks. As a short stack you're giving up dollar equity. Negotiate ICM or refuse.
  3. 3. Bullying TOO wide as a chip leader. Yes, ICM lets you bully — but not infinitely. Shoving too wide means losing a few showdowns, dropping out of chip lead, and giving up the bubble pressure entirely.
  4. 4. Ignoring future game dynamics. Pure ICM assumes no future blinds. Real tournaments have escalating blinds — a short stack with 10bb is different from a short stack with 5bb in a few orbits. Account for this when ICM math is close.

Why ICM doesn't apply in cash games

In cash games, every chip is worth its face value. Doubling your stack doubles your dollar value (you can rebuy for the same amount you'd cash out). There's no prize pool with a bounded payout structure — just direct chip-to-cash conversion.

So in cash games, decisions follow pure pot odds and equity. There's no bubble pressure, no payout-structure adjustment, no chop math. Cash game players who watch tournament videos and try to apply ICM are almost always making the wrong adjustment.

The exception: very deep heads-up cash games where blinds become significant relative to stack. Even then, the effect is small.

Frequently asked questions

What does ICM stand for in poker?

ICM stands for Independent Chip Model. It's the math that converts tournament chip stacks into expected dollar value, accounting for the non-linear relationship between chips and prizes due to the bounded prize structure.

When does ICM matter most?

On bubbles (one or two spots before the money) and at final tables, especially when there are pay-jumps. ICM pressure is highest when short stacks risk being eliminated. ICM matters less mid-tournament when stacks are deep relative to blinds.

Is ICM the same as a chop calculator?

Chop calculators use ICM math under the hood. When 3-9 players reach a final table and discuss a deal, the 'ICM chop' is the fair distribution based on each player's chip stack. Chip-chops (proportional to chips, ignoring payouts) are different and always favor big stacks.

Why does ICM say I should fold AK on the bubble?

Because risking elimination on the bubble has high dollar cost. Even with 65% equity (AK vs random), the dollar value of busting (you go from medium-stack alive to $0) is worse than the dollar value of winning the all-in. ICM tightens medium-stack calling ranges enormously on the bubble.

Can I use ICM in heads-up tournaments?

ICM is mostly trivial heads-up — there are only 2 finishing positions (1st and 2nd), so the math reduces to chip percentage × first-place prize + (1 - chip%) × second-place prize. It still works for chop math but doesn't create the bubble effect that 3+ player ICM does.

What's the difference between ICM and chip EV?

Chip EV (cEV) is your expected change in chips. ICM EV ($EV or M$EV) is your expected change in dollar/buy-in value. They differ in tournaments because chips don't translate linearly to dollars. A play with positive cEV can have negative $EV — that's the entire reason ICM exists as a framework.

Are there better models than Malmuth-Harville?

More accurate models exist (FGS — Future Game Simulation — accounts for future blinds; ICM-Sit-and-Go variants account for short-stacked dynamics). But Malmuth-Harville is the industry standard for chop math because it's simple, fast, and within ~1-3% of more complex models for typical situations.

How accurate is this calculator?

Within ~1-2% of any other ICM tool for typical tournament situations. We use the standard Malmuth-Harville implementation, same as poker training sites and tournament software. Not a substitute for true tournament solvers (ICMIzer, HoldemResources) for high-stakes pro play, but more than accurate enough for studying chops and bubble decisions.

Related