Poker Tournaments: Structure, ICM and Strategy

How it works

Equal buy-in, equal starting chips (no cash value), blinds rising on a timer. Bust and you are out; the survivors split a prize pool weighted to the top. Survival is everything in a freezeout.

Formats

Freezeout (one life), rebuy/re-entry (buy back early), bounty (cash for each knockout), sit & go (single table, starts when full), and MTTs (large multi-table fields).

Blinds, bubble, ICM

Rising blinds force action - you cannot wait forever. The 'bubble' just before the money tightens play sharply. ICM (Independent Chip Model) means chips are worth more or less depending on payout jumps, which is why big stacks can pressure medium ones near pay jumps.

Payouts

Usually only the top 10-15% cash, with the winner taking ~20-25% of the pool down to a min-cash near the buy-in. That top-heavy math rewards timely aggression over playing to merely min-cash.

Core strategy

Early/deep: patient, low risk. As blinds climb and stacks shorten: push-or-fold and steal with position. Adjust to stack sizes and protect against variance with bankroll discipline - even elites bust most tournaments they enter.

Stack depth changes everything

The single most important number in a tournament at any given moment is not your chip count in isolation but your stack measured in big blinds - your 'M' or effective stack depth - because it dictates which strategy is even available to you. Deep-stacked early, with 100 big blinds or more, you can play a nuanced game: speculative hands that flop big, small bets probing for information, folds that cost you little. As blinds rise and your stack shrinks toward 15-20 big blinds, that texture disappears; you no longer have room to see flops and fold, so the game collapses into a push-or-fold problem where your main weapon is all-in pressure. Below 10 big blinds you are in shove-or-die territory, and hand selection becomes almost mechanical - a chart of which hands to jam from which position. The mistake that busts recreational players is playing a deep-stack style with a short stack: calling raises to see flops, then folding, bleeding away the very chips they needed to shove. Match your strategy to your stack depth and you have solved half of tournament poker.

ICM, the bubble and pay-jump pressure

The Independent Chip Model is the concept that separates tournament thinking from cash-game thinking, and it explains behaviour that otherwise looks irrational. In a cash game a chip is worth its face value; in a tournament it is not, because chips convert to real money only through a top-heavy, stepped payout structure. Doubling your stack does not double your equity - each chip you win is worth slightly less than the last, and each chip you lose hurts more than the one before. This is why the bubble - the spot just before the money - warps the whole table. Short stacks tighten up desperately to survive into the cash, and big stacks exploit that fear by relentlessly attacking medium stacks who cannot afford to bust. The same logic returns at every pay jump, especially the final table, where a single ladder step can be worth more than several buy-ins. Playing ICM well means folding hands you would happily play for chip-EV, because survival past a pay jump is worth more than the chips at risk.

Managing variance and bankroll

Tournaments are the highest-variance form of poker, and no amount of skill exempts a player from that reality - even the best professionals cash in only a fraction of the events they enter and go through long, brutal stretches with no big result. The reason is structural: the top-heavy payout means most of your long-run profit comes from a small number of deep runs, so most tournaments end with nothing even when you play flawlessly. This is why bankroll management is not optional but foundational. A common guideline is to keep a bankroll of dozens of buy-ins for the level you play, so that an entirely normal downswing cannot bust you. Move down in stakes when the bankroll shrinks rather than chasing a big score at levels you cannot afford. Judge your play by the quality of your decisions over hundreds of tournaments, never by the result of the last one - the short-run gap between skill and outcome is enormous, and treating a single bust as evidence of bad play is the fastest route to tilt.

A practical game plan

Put the pieces together and a workable approach falls out. Early, when stacks are deep and blinds are tiny, play patiently and avoid marginal all-ins - there is no prize for early chips and plenty of tournament left. As the blinds climb, widen your aggression, steal blinds from position, and lean on your stack rather than calling and hoping. Track your stack in big blinds constantly and switch cleanly into push-or-fold as you shorten. Near the bubble and at each pay jump, respect ICM: pressure the players who fear busting, and refuse to gamble your tournament life on a thin edge unless the chip reward is large. Underneath all of it, protect yourself with bankroll discipline and the understanding that variance rules the short run. Tournament poker is a survival game with a math problem stapled to it - solve the math, weather the variance, and treat the buy-in as entertainment you can afford to lose. Play within your means and step away if it stops being fun. 18+ only.

Reading formats for their true cost

The format on the flyer changes far more than the name suggests, and choosing the right one for your bankroll and skill is a strategic decision in itself. A freezeout is the purest test - one buy-in, one life - and the cheapest to budget, since your maximum loss is fixed the moment you sit down. Rebuy and re-entry events look similar but carry a hidden cost: the disciplined player who buys in once competes against opponents willing to fire three or four bullets, which inflates the real price of contending and rewards deeper pockets. Bounty formats reshape strategy entirely, because a slice of every eliminated player's buy-in becomes an instant cash reward, which correctly incentivises calling all-ins slightly wider than pure chip-EV would suggest - the bounty is real money that changes the math of a marginal call. Sit-and-gos offer a controlled, single-table variance profile ideal for learning ICM in miniature, while large-field MTTs offer life-changing top prizes at the cost of brutal variance and long hours. Matching the format to what you can afford - in money, time and variance tolerance - is the first strategic choice, made before a single card is dealt.

Late-stage play: steals, re-steals and fold equity

As the blinds and antes swell relative to the stacks, tournaments transform into a fight over dead money, and the players who win consistently are the ones who understand fold equity - the value of getting opponents to fold rather than of winning at showdown. In this phase the blinds and antes in the middle of the table are worth attacking, and a well-timed raise from late position with a wide range picks them up often enough to keep your stack alive without ever seeing a flop. The counter-move is the re-steal: recognising when an aggressive opponent is stealing too often and shoving over the top with a range wider than your cards alone would justify, because the fold equity does the work. This is why position and stack awareness become everything late: shoving 12 big blinds from the button applies real pressure, while calling off chips to see flops out of position bleeds you dry. The mistake that ends most amateur runs is passivity here - folding into oblivion as the blinds eat the stack - when the correct adjustment is controlled aggression that uses fold equity while it still exists. Get short-stacked and your patience becomes your enemy; the math rewards decisive, well-chosen aggression over waiting for a hand that may never come. Weather the variance, respect ICM at the pay jumps, and treat every buy-in as entertainment you can afford to lose. 18+ only.

Turning ICM and structure into a simple plan

Everything that makes tournaments distinct from cash games - the escalating blinds, the pay jumps, the shrinking stacks - collapses into one practical idea: the value of a chip is not constant. Early on, with deep stacks and small blinds relative to your chips, a chip is worth roughly its face value and you can play a patient, skill-heavy game. As blinds climb and stacks shorten, each chip becomes more precious, survival gains value, and the correct strategy shifts toward decisive, all-in-or-fold aggression that captures blinds before they eat you. This is why a hand you would happily fold on day one becomes a profitable shove near the money: the changing structure has changed the math.

Independent Chip Model thinking sharpens this near the payouts, because tournament chips do not convert linearly into money. On the bubble and at every pay jump, the risk of busting carries a real dollar cost that a raw chip count hides, which is why correct play tightens up when a laddering opponent is under maximum pressure and loosens when you can exploit their caution. You do not need to compute ICM at the table to use it - you need the instinct that survival is worth more when a pay jump is imminent, and that fold equity is your sharpest weapon when stacks are shallow. Combine that with a bankroll deep enough to absorb the format's brutal variance, since even strong players cash a minority of the time, and you have the whole realistic plan: play the structure, respect the pay jumps, and let volume smooth the swings. Play only with money you can afford to lose. 18+ only.