📬 One free email a week: the weekend's freerolls, ranked by real value
Menu

← All format guides

Bubble play: the most profitable ten minutes of your tournament

Updated Nov 13, 2026

The bubble is the stretch right before the money — one or a few bust-outs away from everyone left getting paid. It is the highest-leverage phase of a tournament, because the payout structure stops rewarding chip accumulation and starts rewarding survival, and most of the field keeps playing the old game. The players who price it — who know what a lost chip costs versus what a won chip gains, in dollars — take money from both directions: from the scared players folding hands they should play, and from the careless players calling off hands they should fold.

By Roger Irwin (Rutland34), FindPokerTournaments — put together with help from other players and the training resources linked on this page.

📚 Keep learning

How it works

One number runs the whole phase: the bubble factor, from Lee Nelson, Tysen Streib and Steven Heston's Kill Everyone — the dollars of equity you lose when you lose an all-in, divided by the dollars you gain when you win it. It turns "should I call?" into arithmetic: you need bubble factor ÷ (1 + bubble factor) of the pot's equity, not the pot odds your chips quote you. This guide prices real tables seat by seat with the site's bubble factor calculator, works the hands players actually agonize over — QQ, JJ, AQs, AQo, AJs, AJo — against three kinds of shover, and finishes where the math meets real life: the same spot played on a $110 bankroll and on a $5,500 one.

Worked example

A $55 nightly, seven players left, six paid $1,100 / $700 / $460 / $320 / $220 / $150. You are second in chips with 9,000 — 18 big blinds — and the chip leader (15,000) shoves into you. Your stack is worth $488.48. Fold and it is worth $486.09; call and win and it is worth $703.05; call and lose and you are out for $0. You gain $216.95 by winning and lose $486.09 by losing — a bubble factor of 2.24 — so the hand you are holding needs 69.1% equity against his range, while the pot alone quotes you 48.0%. That 21-point gap is the bubble, priced. Whether you have 69.1% depends entirely on who he is — and that is the rest of this guide.

Why the bubble is the most profitable phase

Every tournament has a stretch where the money moves without cards being shown. It is the bubble, and it moves for a simple reason: the payout sheet changes what a chip is worth, and most players do not change with it.

Before the money, a chip lost and a chip won are close to the same animal. On the bubble they split apart. Lose your stack and you get $0 while six other players lock up at least the min-cash; win the same confrontation and you gain chips whose value is capped by a payout structure that pays the next man out exactly what it pays you for surviving. Losing costs more than winning gains — always, measurably, and by an amount you can compute. Kill Everyone's chart of the Sunday Million has the average player's bubble factor peaking at about 1.6 right on the money bubble, then falling once the bubble bursts. The average. At real final-table-shaped bubbles, against the wrong opponent, it runs far higher — this guide prices a seat at 2.24, another at 3.18, and a satellite seat at 13.52.

That asymmetry creates three populations, and money flows between them in one direction. The scared players feel the asymmetry but overdo it: they fold hands the math says to play, bleed their stacks down orbit by orbit, and donate their equity to whoever is willing to take it. The careless players ignore the asymmetry entirely: they call off at pot odds, bust on the bubble with hands that were never close, and donate their equity to the players who folded correctly and laddered past them. The pricers — the players this guide is training you to be — sit in the middle collecting from both: attacking the scared stacks relentlessly, because those players' own math forbids them to fight back, and folding without drama when the careless money actually has it, because the price of calling is printed before the hand starts.

None of this requires a solver or a HUD. It requires knowing three numbers for the spots your pools keep dealing you: what the pot offers, what the ICM math demands, and what your hand is actually worth against the range of the specific human doing the shoving. The rest of this guide is those three numbers, worked.

The bubble factor, priced at a real table

Here is the table from the top of this guide, priced seat by seat by the bubble factor calculator — a $55 nightly, seven left, six paid, blinds 250/500, $2,950 in the pool. The last line of each card is the seat's bubble factor against the chip leader and the equity it would need to call his shove. Read the column top to bottom and watch what the chip leader does to a table without playing a hand.

Chip leaderTHE SHOVER
15,000 chips · 30.0 BB
$EV now
$633.83
His price is everyone else's problemEvery seat below is priced against him.
You2ND STACK
9,000 chips · 18.0 BB
$EV now
$488.48
BF 2.24 vs the leaderYou need 69.1% to call him. Pot odds quote 48.0%.
Third stackMEDIUM
7,600 chips · 15.2 BB
$EV now
$443.37
BF 2.12 vs the leaderNeeds 68.0% to call him.
Fourth stackMEDIUM
6,800 chips · 13.6 BB
$EV now
$414.64
BF 2.04 vs the leaderNeeds 67.1% to call him.
Fifth stackMEDIUM
5,900 chips · 11.8 BB
$EV now
$379.11
BF 1.92 vs the leaderNeeds 65.8% to call him.
Sixth stackSHORT-MEDIUM
4,700 chips · 9.4 BB
$EV now
$324.91
BF 1.74 vs the leaderNeeds 63.5% to call him.
Short stackSHORTEST
3,600 chips · 7.2 BB
$EV now
$265.66
BF 1.51 vs the leaderNeeds 60.1% — the "freest" seat at the table, and it still needs 60%.

Your confrontation, in full. Fold, and the leader collects the 750 already in the middle: your stack is worth $486.09 (a touch under the $488.48 above — the pot you surrender was partly priced into your seat). Call and win, and it is worth $703.05: you gain $216.95. Call and lose, and you are the bubble boy: $0, a loss of $486.09. Bubble factor = 486.09 ÷ 216.95 = 2.24. Required equity = 2.24 ÷ 3.24 = 69.1%. The pot — your 9,000 to win a pot of 18,750 — only asks for 48.0%. Anyone playing this spot by pot odds is playing a different, cheaper game that is not on offer.

Two things to notice in the cards. First, the prison is not really about you: every seat needs at least 60% against the leader, including the short stack who has the least to protect. The leader can shove an enormous range into this table at a profit, because nobody's price lets them fight back without a monster. Second, your price depends on who is shoving, not just that someone is: against the 7,600 stack your bubble factor is 1.65 (you would need 62.3%), and against the 3,600 short stack it is 1.04 — 51.0%, essentially pot odds, because he cannot bust you. Same hand, same table, three different correct answers depending on the seat the shove came from. Bubble factor is a relationship between two stacks, never a property of your hand.

Checking the engine against the literature: the calculator's ICM core is the standard model, and it reproduces published worked examples — a PokerStrategy SNG example (nine players, 1,500 chips, 50/30/20 payouts) comes out at fold 11.1% of the pool, win 20.3%, required equity about 55%, matching their printed numbers; and it reproduces this site's own $11 example from the micro-stakes ICM guide to the cent ($29.27 fold value, $42.30 call-and-win, 69.2% required). The numbers in this guide are that engine's output, quoted exactly.

🫧 Run your own spot in the Bubble Calculator — your QQ is pre-loaded in the prison scenario. Flip the shover, drag the stacks, and watch the price move.

The torture hands: QQ, JJ, AQs, AQo, AJs, AJo against three shovers

Now the hands. A bubble factor tells you the price; it does not tell you what your hand is worth. That depends on the shover's range — and the shover's range depends on who he is, which in real pools means: how much bubble education he has. An educated reg shoves a structured range that respects the table. A losing player without that education shoves one of two wrong ranges: scared money shoves only hands he thinks are strong (far too tight), and a punter shoves anything he feels like (far too wide). Neither matches the textbook range a solver would give him. Price the same seven left / six paid table (you need 69.1%) against all three:

Your handVs educated reg (~14%)Vs scared money (~6%)Vs punter (~31%)
QQ68.8% CLOSE (−0.3)60.8% FOLD71.8% CLOSE (+2.7)
JJ63.1% FOLD53.4% FOLD68.0% CLOSE (−1.1)
AKs (reference)60.2% FOLD52.6% FOLD65.2% FOLD
AQs55.8% FOLD42.8% FOLD62.6% FOLD
AQo53.5% FOLD39.7% FOLD60.8% FOLD
AJs51.2% FOLD37.1% FOLD60.6% FOLD
AJo48.5% FOLD33.2% FOLD58.5% FOLD

The ranges, printed in full because the range IS the read: the reg shoves 55+, A8s+, ATo+, K9s+, KQo, Q9s+, J9s+, T9s, 98s (~14% of hands). Scared money shoves 88+, AJs+, AQo+, KQs (~6%). The punter shoves any pair, any ace, K7s+, K9o+, Q9s+, QTo+, J9s+, JTo, T9s, 98s, 87s (~31%). The equities are Monte Carlo estimates against exactly those ranges (250,000 simulated runouts per cell, ± about a point) — estimates, labelled as such, because a real player's range is never known to the decimal. The 69.1% they are compared against is exact.

The flip this table exists to show: QQ. Against the reg, queens have 68.8% — a fold by three-tenths of a point, close enough that the calculator marks it CLOSE and your read has to be very sure to override it. Against the punter, the same queens in the same seat against the same stack have 71.8% — the only cell at this entire table that clears the line at all, and it clears it by 2.7 points, still inside the close band. Nothing about the hand changed. Nothing about the stacks changed. The villain's education changed, and queens went from a fold to the table's only call. And look at scared money: against his 6% range, queens drop to 60.8% and nothing you can be dealt short of kings and aces is a call. Players call this man "tight" like it is an insult. On the bubble, tight is a weapon pointed at you: his range is so strong that correct play against him is to fold hands you would happily stack off with anywhere else.

And AKs — the hand from the micro-stakes guide. There, four-handed on an $11 bubble, AKs needed 69.2% and had about 55–60% depending on the range you put the shover on: a clear fold, and the same shape as this table (same 2.24-class bubble factor, same ~69% requirement — the $11 bubble and the $55 bubble price a medium stack almost identically). Here it never gets close against any of the three shovers. AKs is the hand bubble dreams are made of and bubble bankrolls are buried with: it looks like a premium, it plays like a coin, and on the bubble a coin is a terrible price. The hands that survive the bubble against real ranges are the ones already ahead — pairs — not the ones hoping to get there.

The CLOSE band is a doctrine, not a fudge. The calculator marks any cell within 3 points of the line as close, and means it: the equity side of the comparison is an estimate built on a range you assigned to a human. When the model says −0.3, the honest reading is "the math cannot decide this; your knowledge of this specific player has to." That is not a licence to call everything close. It is the opposite: it tells you exactly which spots your reads are worth money in, and which spots no read can save (AQs against the reg is not a read problem; it is 13 points short).

Work this matrix live in the calculator → Flip the shover between the three educations and watch the verdicts move; then drag the equity slider and find exactly where each hand crosses.

Figuring out the fold: the five-step workflow

Everything above compresses into a workflow you can run in the seconds a real decision allows — because you rehearsed it in study time until the steps are reflexes. Here it is, end to end, on the spot from this guide: you hold AQs, the chip leader has shoved, seven left, six paid.

Step 1 — Pot odds, the price in chips. There is 750 in the middle and 9,000 to call into a pot that will be 18,750. Chips say you need 48.0%. This is the number most players stop at, and on the bubble it is the least important of the three.

Step 2 — Who is he? Before ranges, identity. You have seen this screen name at your final tables for months, and in study time you did the unglamorous work: his SharkScope line says winning player, five figures of games, deep runs in exactly these nightlies. He is the educated reg. If the same line had said −30% ROI over 400 games, he is one of the other two — and his results shape plus how he has played tonight tells you which: the scared money who has folded for an hour, or the punter who has shoved three times already. The ROI and the games played are what turn "some guy shoved" into a range. This is the step that cannot be improvised at the table, which is why it lives in study time (the tagging section below).

Step 3 — Your equity against HIS range. The reg's range is the ~14% printed above. AQs against it: 55.8% (estimate). Note what just happened: you are ahead of the pot-odds price (55.8 > 48.0) — by chips, this is a call, and it is not even close by chips.

Step 4 — The ICM requirement. Your bubble factor against this stack is 2.24, so the real price is 69.1%. That number comes from the stacks and payouts only — it does not care about your hand or your read.

Step 5 — Compare, and obey the comparison. 55.8 against 69.1: fold, by 13.3 points. Not close, not a read spot, not a "feel" spot. The chips said call, the bubble said fold, and the 21-point difference between those two prices is the exact amount of money the average player donates in this spot over a year of nightlies. When the answer is this clear, the skill is not in finding it — it is in having rehearsed it often enough that folding AQs to a shove at 18 big blinds does not feel like surrender. It feels like what it is: declining a bad price.

That is the whole workflow: pot odds → who is he → equity vs his range → the ICM requirement → decide. Five steps, in that order, every bubble spot, until it runs itself. The calculator exists so you can run it a hundred times at home against the spots your pools actually deal — start with this one — because the version of you at the table, with the clock running and $486 of equity on the line, will not invent it on the spot.

When to just go for it

"Take the ICM line" is only half the education. The bubble factor cuts both ways: the same math that imprisons the medium stacks sets two other seats free — the shortest stack, who has almost nothing left to protect, and the chip leader, who cannot be hurt. Both of them should be going for it, and the guide prices both.

The short stack who must go. A $33 nightly, six left, five paid ($780 / $520 / $360 / $240 / $160), blinds 250/500. You are last: 2,900 chips, 5.8 big blinds, worth $207.64. The 7,200 stack shoves into your big blind. Fold and you keep $205.04; call and win and you jump to $364.96. Your bubble factor is 1.28 — you need 56.2%, against the 69.1% the medium stack needed in the prison. Why so gentle? Because folding is not safety for you: at 5.8 blinds with the ante-less orbits eating 750 a round, folding is the slow bust. The math knows it: your stack's value is mostly the chance to double, not the chance to ladder past five other stacks who are all healthier than you.

Look at the same table from the other seats and the prison reveals itself as a stack shape, not a phase: against this same shover, the 4,600 stack's bubble factor is 1.84 (needs 64.8%) and the 3,600 stack's is 1.63 (needs 61.9%). Only you, the shortest, are priced at 1.28. On any given bubble, exactly one seat is free, and it is the seat with the least to lose. Now the hands — you need 56.2%: against the reg's ~22% steal range (44+, A2s+, A7o+, K8s+, KTo+, Q9s+, QJo, J9s+, T9s, 98s), QQ (70.1%), JJ (65.9%) and AQs (60.5%) are calls, AQo (58.3%) and AJs (57.7%) are close calls, and AJo (55.4%) is a close fold. Against the punter's ~35%, everything listed is a call, AJo included (58.7%). Against scared money's 6%, the lesson flips one more time: QQ still calls (60.8%), JJ becomes a close fold (53.4%), and the ace hands all fold — even with 5.8 blinds, you fold AQs to the man who only shoves pairs. "Must go for it" never means "call anything." It means your price dropped far enough that your real hands clear it — against the ranges real people shove.

Run the short stack's table →

The big stack's playground. Same math, other end. A $55 nightly, six left, five paid ($1,400 / $950 / $640 / $430 / $290). You are the chip leader with 28,000 — worth $1,063.18 — and the 11,000 medium stack shoves into you. Call and lose and you still hold a stack worth $888.54: you do not bust, you do not even lose the lead. Your bubble factor is 1.09; you need 52.2%, close to the 47.2% the pot quotes. Against the shortest stack (2,400) your factor drops below one — 0.63, needing just 38.6% — because busting him gains you more, proportionally, than his double costs you. Now flip the chairs, because this is where the playground comes from: that same 11,000 stack, facing your shove, has a bubble factor of 3.18 and needs 76.1%. The 9,000 stack needs 74.9% (BF 2.98). They cannot call you without the top of their range, they know it, and you know it — so you shove hands that would be spew from any other seat, and the table's correct response is to let you. That is not bullying as a personality trait. It is arithmetic with a stack behind it, and it lasts exactly as long as the stack does: the moment you are no longer the coverer, your price changes back.

Structure decides how much of this applies. The flatter the payouts, the higher everyone's bubble factor and the more the ICM line rules — the satellite below is the limit case. The more top-heavy the structure (first place dwarfing the min-cash ladder), the lower the true price of going for it: when first pays ten min-cashes, surviving to fifth is a consolation, not a result, and the calculator will show it — enter a top-heavy payout sheet and watch every bubble factor at the table fall. And the coverer distinction is absolute: all of the freedom above belongs to the player who cannot be busted by the confrontation. The second-biggest stack tangling with the biggest is not "going for it." He is the prisoner from section two, volunteering.

Run the playground — then flip the "You" seat →

When the ICM line rules

The other half of the discipline: the spots where going for it is just donating. You have met most of them already — the medium stack against a bigger stack (every seat in the prison needed 60%+ against the leader), any stack against scared money's range (his tightness is doing ICM's work for it) — and one structure where the principle reaches its limit.

The satellite bubble. A $55 satellite, six players left, five identical $215 seats. You hold 11,000 (22 big blinds, third in chips) — a stack the model values at $197.61 of the $215 seat. The chip leader shoves. Fold and you keep $197.55 of it. Win and your stack is worth $212.15 — a gain of $14.61, because a seat is a seat: once you are nearly locked, more chips buy almost nothing. Lose and you are out for $0. Bubble factor: 13.52. Required equity: 93.1%, against pot odds of 48.4%. The calculator sets the equity slider to the best case for calling — QQ against the widest range a maniac would shove, about 71.8% — and it is 21 points short. In a satellite, folding is not a style. Above-average stacks fold everything but aces (and think hard about kings), medium stacks fold more than feels sane, and the only players who should be relaxed are the ones so short that the seat is slipping away anyway. If you play satellites one way and normal bubbles the same way, one of the two is costing you a fortune — the payout sheet, not the cards, decides which.

Run the satellite — try to find a call →

Pay jumps that matter. The same principle, one notch down: just inside the money, the first pay jumps are usually small and bubble factors fall fast (the Sunday Million average drops from ~1.6 toward 1 once the bubble bursts) — but at final-table bubbles and big structured jumps, the prison rebuilds itself. The habit to build is not "tight on bubbles, loose in the money." It is: read the next two rungs of the payout sheet, and price your stack against the man shoving, every time the shape changes. The calculator's table view prints your bubble factor against every seat for exactly this reason — the ladder, not your own row, is the map.

The bankroll layer: the same spot at $110 and at $5,500

Everything so far is the math answer, and the math answer has a property players find suspicious: it does not know what your bankroll is. ICM dollars are ICM dollars. But you are not an ICM dollar — you are a player with a finite roll, a next month of buy-ins to fund, and a life the roll has to survive. So the site's tools show the spot through a second lens, labelled as what it is: the roll answer. Same table, both lenses:

The spot: the $11 bubble from the micro-stakes guide — four left, three paid $55 / $33 / $22, you hold the second stack and QQ against a wide shover (about 71.8%, estimate). The math, identical on both cards: bubble factor 2.25, required 69.2%, pot odds 46.9%. Math answer: call — by 2.6 points, inside the close band.

💰 The $110 roll

The buy-in was already the story: $11 is 10.0% of this roll. The bankroll rules put a standard buy-in near 1% and a deliberate shot at 2–5% — at 10%, this event was a level jump before a card was dealt.

Busting here costs $29.27 of equity — 26.6% of everything you play with. The min-cash you are protecting, $22, is 20.0% of the roll.

Roll answer: the math's 2.6-point edge is real and it is not enough. On a call this close, at this share of the roll, folding the coin-flippy part of the range is not fear — it is the discipline that keeps a $110 roll alive to become a $5,500 one. What this roll actually needs is a smaller game, where its bubble spots are priced in cents of the roll, not quarters of it.

🏦 The $5,500 roll

The buy-in is a rounding error: $11 is 0.2% of this roll — this is a 500-buy-in bankroll for the $11 game, deeper than the standard rules even ask for.

Busting here costs the same $29.27 of equity — 0.53% of the roll. The min-cash, $22, is 0.4%: a line item, not a lifeline.

Roll answer: identical to the math answer, and it is not close. Take every +EV spot the calculator approves, at full size, forever. Variance at half a percent per spot is weather; the roll is built to stand in it.

The reconciliation, stated plainly, because this is where players lie to themselves in both directions. The roll never changes the math: QQ has 71.8% against that range on both cards, the price is 69.2% on both cards, and anyone who tells you the hand is "more of a call" because they are rolled for it is confusing two different questions. What the roll changes is which edges you can afford to take. A +2.6-point edge that risks 26.6% of your roll is a bad trade for the roll even though it is a good trade for the stack — you will face this spot a few times a year, and the roll that survives to face it fifty times beats the roll that takes it twice and busts. A +2.6-point edge that risks 0.53% is a rounding error you take without thinking. The rules from the bankroll guide are the bridge: keep a single confrontation's downside near 2% of the roll and the two answers almost always agree; drift toward 5% and you are shot-taking, where the roll answer is allowed to veto close math answers; past 5% in one hand, you are not taking a shot any more, you are gambling, and no bubble factor fixes the game selection that put you there.

One more honesty note, in the other direction: "my roll is small" is not a blanket excuse to fold every close bubble spot for a year. Folding +EV spots has a price too — it is paid in the tournaments you never win, the final tables you ladder to instead of through. The $110-roll play is not "fold more, forever." It is "fix the buy-in, so the spots go back to being math problems instead of rent problems." Run the $110 version, then the $5,500 version, and watch the ICM columns refuse to move while the lens underneath them changes completely.

Study time: tag the folders and the punters

Step 2 of the workflow — who is he? — is the whole edge, and it is built entirely away from the table. Here is the habit, concretely.

Read the line that matters. A villain's SharkScope profile answers the two questions the matrix needs. ROI tells you his direction: a winning player over a real sample plays the educated ranges — structured shoves, disciplined folds. A losing player plays one of the two uneducated ones. Games played tells you how much to trust the line and how set his habits are: a −30% ROI over 6,000 games is a permanent population member; over 60 games it is a coin you cannot read yet. Match the buy-in band while you are at it — a player can be a crusher at $11 and a punter at $55, and it is the $55 line that prices his shove in your $55 nightly.

Then split the losers. ROI alone cannot tell scared money from a punter — both lose, one by folding his equity away, the other by lighting his on fire. The split comes from watching: scared money shows you an hour of folds and a VPIP you can count on one hand; the punter shows you three shoves nobody made him make. Both reads get a tag in your notes, in study time, the moment you form them: folder or punter, with the stakes and the date. The next time that screen name shoves on your bubble, step 2 takes two seconds and the matrix column is already chosen.

Know what the tags buy you. Against your tagged folders, widen everything: shove into them mercilessly (their calling range is the scared column — QQ and JJ fold to you), and fold your AQs when they finally shove, because their 6% range is the one range in poker that means it. Against your tagged punters, tighten your shoves (they call) and widen your calls (their 31% range is why QQ becomes a call on the prison table). Against the regs, play the textbook column and respect the close band. Three tags, three columns, and the bubble stops being a personality test and becomes a lookup.

Why study these spots at all — the repetition argument. You will face a money bubble in nearly every tournament you play deep, and the same shapes recur with a small cast: the medium stack against the leader, your short stack against a steal, your big stack against the table. A football team does not improvise its two-minute drill. Price your ten most common bubble shapes with the calculator, decide your answers per villain tag, and the at-the-table workflow shrinks to steps 1 and 5 — the price, and the decision you already made. Players who improvise the bubble are paying tuition every night for a lesson that costs one study session. (And the standing rule, here as everywhere on this site: all of this happens between sessions. Study tools stay closed while a poker client is open — rooms prohibit in-play calculation, and the entire point of the habit is that the answer is in your head when it matters. SharkScope links on this page are affiliate links, disclosed as everywhere on the site; the tags work in any notes app you already keep.)

Strategy tips

  1. Run the five steps in order — pot odds, who is he, equity vs his range, the ICM requirement, decide — and never let step 1 answer the question alone on a bubble.
  2. Your bubble factor is a relationship, not a hand strength: re-price it against every stack that can shove on you, not just the one that did.
  3. The shortest stack is the freest seat at the table, not the tightest. If folding is your plan at 6 BB, the plan is to bust slowly.
  4. Tag folders and punters in study time. The scared-money column and the punter column differ by 30+ equity points with hands like AJo — that gap is your edge, pre-computed.
  5. Treat anything within 3 points of the required equity as a read spot, and be honest that it is one. Outside the band, obey the math without negotiating.
  6. When the roll answer and the math answer disagree on a close spot, fix the buy-in next time rather than the fold this time. Game selection is the durable answer.

Common mistakes

🔎 Know who you're up against

On the bubble, who is shoving matters more than what you hold: the same QQ is a fold against a winning reg's 14% range and the table's only call against a punter's 31%. Read the SharkScope line in study time — ROI for his direction, games played for how set it is — and tag the folders and the punters in your pools. When the spot arrives, the column is already chosen.

How to use SharkScope to pick tournaments · SharkScope ↗affiliate link (affiliate link)

SharkScope covers ACR, CoinPoker, BetOnline, Ignition/Bodog, GG, WSOP and more, not ClubWPT Gold. Check each room's rules on third-party tools.

🎯 Sharpen your game

Hand-picked places to get better at this format. Free stuff first. Community marks working players, coaches and smaller creators worth supporting. Links checked Nov 5, 2026.

📖 Free articles & videos

  • What is the Bubble Factor in poker tournaments? ↗ Article
    GTO Wizard · The framework straight from the solver era: bubble factor defined, required equity derived, and Kill Everyone's Sunday Million chart (average factor peaks at about 1.6 on the money bubble). Our engine reproduces its worked examples.
  • The Bubble Factor in different SnG formats ↗ Article
    PokerStrategy · The classic worked examples — the 9-handed 50/30/20 case (fold 11.1% of the pool, win 20.3%, needs about 55%) is the published check our calculator's test suite reproduces.

📚 Books & audiobooks

  • Kill Everyone ↗
    Lee Nelson, Tysen Streib & Steven Heston · 2007 · The book that invented the bubble factor (Nelson, Streib & Heston). The required-equity math in our guide and calculator is theirs; this is the source, with the charts for every stage of a tournament.
  • Endgame Poker Strategy ↗ Community
    Dara O'Kearney & Barry Carter · 2021 · Dara O'Kearney's ICM book — the endgame and satellite thinking behind the satellite section, from the player who literally wrote the satellite book too.

🔎 Tools

  • SharkScope ↗ (affiliate link)
    Where the villain columns come from: a shover's ROI and games played tell you whether he plays the reg's range, the scared range or the punter's — before the spot, in study time. Tag the folders and the punters in your pools and the matrix picks its own column. How to use it

🎓 Training sites & coaches

  • ICMIZER ↗
    Preflop solver & ICM trainer · Shows bubble factors inside its paid ICM solves — the deep-study version of this math. Equal-skill ranges, though: it will not price a punter, which is the gap our free matrix exists to cover.
  • HoldemResources Calculator ↗
    ICM solver (HRC) · The other pro ICM engine; its future-game simulation is the deepest bubble modeling available for study. Paid, desktop, equal-skill — pair it with this guide's player-type layer.
  • Push Fold Trainer ↗
    Push/fold trainer · Free reps for the shoving side of the bubble: drill your push ranges until the short-stack decisions in this guide are reflexes, not calculations.
  • GTO Lab ↗
    MTT ICM trainer (browser) · Its ICM trainer drills the exact spots this guide prices — bubble, two and three tables left, final table — from pre-solved solutions, and its coaches (Petrangelo, Dvoress) literally wrote a masterclass on ICM. NLH MTT only; ICM access is the $89/mo Tools tier. A between-sessions study tool, never alongside a client.

𝕏 Players to follow on X

  • Dara O'Kearney @daraokearney ↗
    Player on X · The ICM and satellite author whose books anchor this shelf; posts endgame and satellite spots at exactly these stakes.

Find these tournaments

The bubble factor calculator → Skill-adjusted ICM chop calculator → ICM for micro-stakes players → Bankroll management →

FAQ

How tight should I play on the bubble?

The bubble is the stretch right before the money — one or a few bust-outs away from everyone left getting paid. It is the highest-leverage phase of a tournament, because the payout structure stops rewarding chip accumulation and starts rewarding survival, and most of the field keeps playing the old game. The players who price it — who know what a lost chip costs versus what a won chip gains, in dollars — take money from both directions: from the scared players folding hands they should play, and from the careless players calling off hands they should fold.

What is a good bubble factor?

There is no good or bad value — it is a price, not a grade. Around 1.0 you can play close to pot odds (cash games, heads-up, winner-take-all). Around 1.3–1.6 is normal approaching the money in a big field. Above 2 a covered medium stack needs genuinely premium equity to call, and satellites run into double digits. What matters is knowing yours against the player shoving, before the decision arrives.

Should I ever call when the math says fold?

Inside the close band (within about 3 points of the required equity), yes — that is what the band is for, and your read of the specific player is the tiebreaker the model cannot supply. Outside the band, a call is not a read; it is a donation with a story attached. If you find yourself overriding the model weekly, the problem is not the model.

Does ICM apply before the bubble too?

Yes, with smaller numbers. The same engine prices the first hand of a 50/30/20 sit-and-go at a bubble factor of about 1.21 — you need roughly 55% where a cash game needs 50%. The bubble is where the effect peaks and where the money concentrates, which is why it gets its own guide; but "ICM only matters near the money" is wrong at both ends of the tournament.

Can I check a villain's SharkScope during the hand?

No — study tools stay closed while you play; that is the rooms' rule and this site's, and it is also impractical. The lookups and the tagging happen between sessions, so that at the table "who is he?" is a memory, not a search.

Related guides

ICM chops and the skill tax · ICM for micro-stakes players · Bankroll management · Session planning

Play responsibly. Tournaments are entertainment, not income. Set a budget and time limit, and never chase losses. 18+/21+ depending on site and state. Help: 1-800-GAMBLER.