ICM stands for Independent Chip Model. It is the standard method for calculating chip equity in poker tournaments.
Here is the problem ICM solves: three players remain. One has 5,000 chips. One has 3,000. One has 2,000. The prize pool is $1,000 for first, $600 for second, $400 for third.
If they all fold forever, eventually one person will win. But how much is each player's stack actually worth? That is what ICM calculates.
The chip leader ($5,000 stack) has about $520 expected value. The second player ($3,000 stack) has about $460. The short stack ($2,000 stack) has about $240.
These add up to $1,220, not $2,000, because ICM accounts for the fact that not all the chips will be played out (some go to players busting and exiting).
How ICM Changes Play
A short stack might think: I need to gamble to catch up. My chip count is small. I am doomed.
ICM tells a different story. Your stack is worth $240. If you fold every hand, you will lose maybe $10-20 in small blinds before you exit. You go from $240 to $220. Not good, but not disastrous.
If you get all-in with 50% equity, you win $240 and exit with $480+, or you lose and go home with $0. The expected value is $240, same as before. You have not changed your situation.
But if you get all-in with 25% equity, you are now gambling below your chip value. You are making a bad deal. The short stack, surprisingly, should tighten their range and fold more often (at least until blinds take their course).
The Math
ICM uses iterative calculation. You remove one player at a time. Each removal assumes chip loss proportional to their chips.
Start with three players (5,000 / 3,000 / 2,000 chips). Remove the shorty. They bust with probability 2,000 / 10,000 = 20%.
When they bust, the remaining chips redistribute to the two survivors proportionally. The chip leader's (5,000) share of the remaining (10,000) is 50%. They win $600 extra (second place becomes first).
Calculate expected value for each player by weighting all possible exit orders.
The calculation is tedious but conceptually straightforward.
When It Matters
ICM matters most in heads-up (two players) and near-bubble scenarios (three-six players).
Early in a tournament, ICM has little relevance. Chips are chip. Positional advantage matters more than chip counting.
Near the bubble (when three or four players remain and payouts change significantly based on who is eliminated), ICM becomes critical. A player who understands ICM plays the bubble correctly. Tight on the short stack. Loose on big stacks.
Practical Application
A player uses ICM to answer: what is my all-in equity need?
If ICM says my stack is worth $240, and I am about to bet $100 to win $150, what is my equity need?
I need to win: 100 / (100 + 150) = 40%.
If my cards have greater than 40% equity against my opponent's range, I should get it in. If less, I should fold.
Many players play the bubble without ICM. They play by feel. ICM turns this into math.
Limitations
ICM assumes that all chip stacks will eventually be played out (which happens in real tournaments). It assumes random chip removal (which does not match actual play). It assumes equity is split by chips (which is roughly true).
Because of these assumptions, ICM is a model, not reality. But it is the best available model for equity calculation.
Sharp players use ICM as a starting point and adjust for opponent tendencies, position, and stack situations.
The Edge
A player who understands ICM gains a consistent edge near the end of tournaments. They make more correct decisions on the bubble. They preserve equity through tight play when short. They extract equity through aggressive play when big.
Over many tournaments, this edge compounds. ICM awareness is not just academic. It is profitable.
