Home › Learn

Poker bankroll management: how many buy-ins you actually need

Cash games, online and live. Every number on this page is calculated rather than quoted, so you can see why the advice you have read ranges from 20 buy-ins to 100 — and which end of that range applies to you.

The answer table

Start here, read the reasoning afterwards. Minimum means a 10% risk of ruin — one player in ten with these numbers loses the whole roll. Standard is 5%, Conservative is 1%. All figures are buy-ins of 100 big blinds, rounded up, calculated from the win rate and standard deviation in the first column.

σ (sigma) is standard deviation in big blinds per 100 hands — how much your results bounce around. Run your own numbers in the bankroll calculator.

The same numbers in money

Online rows assume 5 bb/100 and σ 100; live rows assume 10 bb/100 and σ 85. Most live players should add a margin on top — see the live section below.

What bankroll management is

A bankroll is money set aside for poker that you can lose entirely without changing how you live. Bankroll management is the set of rules that decides which stake that money lets you play and what you do when it shrinks or grows. That is all it is. It does not make you win; it decides whether you are still in the game by the time your edge shows up.

The mechanism is a formula called risk of ruin: the probability you lose the whole roll if you keep playing the same stake all the way down. It needs exactly three inputs — your win rate, your standard deviation, and your bankroll in big blinds. The output is a percentage, and the shape of the relationship is what matters:

• Your win rate is on the bottom of the fraction. Halve it and you double the bankroll you need. This is why bankroll arguments are really win-rate arguments.
• Your standard deviation is squared. A loose aggressive style at σ 140 instead of σ 100 needs roughly twice the roll at the same win rate. Your style is a bankroll decision.
• Safety is cheap at first, then expensive. Going from a 10% risk to 5% costs about 30% more money. Going from 5% to 1% costs another 50%.

Why variance sets the number

Here is the picture every bankroll rule is trying to survive. Each grey line below is the same winning player — 5 bb/100, σ 100 — playing the same 100,000 hands. The green line is the expected result, and the shaded bands hold the middle 70% and 95% of outcomes.

The slope is the edge: 5 bb/100 over 100,000 hands is 50 buy-ins of profit. The spread is the variance, and it is far bigger than people expect. By the end of this sample the middle 95% of outcomes runs from about -12 to 112 buy-ins. The same player, with the same skill, can finish either of those and nothing went wrong.

Worse, the path matters more than the destination. Over these 100,000 hands the typical worst drop from a peak is 22 buy-ins, one run in ten falls 36 buy-ins or more from its high point, and one in a hundred falls 53. That is for a genuinely good player. Your bankroll is not sized for your expected result; it is sized for the hole you have to climb out of on the way there.

Which is what the risk-of-ruin table is really saying. Read across this row and the usual advice starts to make sense:

Recreational, serious, professional

The same maths produces three different answers because the cost of going broke is different in each case.

• Recreational. You have income elsewhere and poker is entertainment. Your real bankroll rule is a deposit limit you can refill without wincing — 20 buy-ins of the game you enjoy is plenty, and the 10% risk of ruin column is genuinely fine, because ruin means depositing again rather than losing your job.
• Serious amateur. You are trying to move up and you hate reloading. Use the Standard column (30 buy-ins online 6-max) and treat the roll as a closed system: no deposits, no withdrawals below target, move down when the rule says so.
• Professional. The roll pays the rent, so ruin means a job hunt. Hold the Conservative column or more — 50 to 100 buy-ins for cash — and keep at least six months of living expenses in a separate account that the bankroll rules never touch. A pro who takes a monthly draw from a 50 buy-in roll does not really have a 50 buy-in roll, because the draw keeps arriving during a downswing.

Moving up and moving down

Most bankroll damage is not done by choosing the wrong number; it is done by ignoring the number you chose. Write the two lines down before you need them. The move-up line is the full requirement for the new stake. The move-down line is where your risk of ruin has roughly tripled.

Two practical notes. First, moving up is not a promotion you award yourself for running well — the requirement is measured in buy-ins of the new stake, which at most online ladders means roughly doubling your roll each step. Second, moving down is the skill, not the shame. A player who drops a level at the line and rebuilds almost never actually busts, which is why real-world bust rates are far lower than the risk-of-ruin column suggests. The formula assumes you keep firing at the same stake into the abyss. Do not be that assumption.

Taking a shot

A shot is a deliberate, limited attempt at a stake your roll does not yet support. It is a perfectly reasonable thing to do, as long as every term is set in advance.

• Size it so losing does not change your main game. Two buy-ins of the higher stake is the usual shape, and never more than about 5% of your bankroll on the table at one time.
• Set the give-up point before you sit. Two buy-ins lost, or one session, whichever comes first. Write it down; the version of you two buy-ins deep is not a reliable negotiator.
• Set the stay-up condition too. If the shot wins enough that you now hold the full requirement for the bigger game, you are not on a shot any more — you have moved up.
• Only shoot a game you have studied. A shot at a tougher table with no preparation is a donation with a story attached.

Stop-losses and session rules

Risk of ruin is a long-run number and says nothing about today. Session rules cover the part of the game the maths cannot see: you, tired, three buy-ins down, convinced the next hand owes you something.

• A stop-loss is a tilt control, not a bankroll control. Quitting at two or three buy-ins down does not change your expected value if you are playing well — it protects you from the hours when you are not.
• Set a time limit as well as a money limit. Most people's play degrades well before their bankroll does.
• Never move up to recover a loss. This is the single most expensive habit in poker, and it is the exact opposite of what the move-down rule tells you to do.
• Leave a good table only for a good reason. A stop-win is usually a bad rule: the game being profitable is not a signal to stop.

Cashing out

Almost no bankroll guide covers withdrawals, and it is where a lot of careers quietly stall. Two failure modes: never withdrawing, so the roll becomes money you have mentally spent already and every downswing feels like a personal loss; or withdrawing everything, so you never reach the requirement for the next stake.

Pick a target bankroll for the stake you want to play, then set a standing rule for profit above it — for example, withdraw half of anything above target at the end of each month and leave the rest working. If you are playing for a living, pay yourself a fixed monthly draw rather than a percentage of results, and size it so it still works in a losing month. A variable draw during a downswing is how a 50 buy-in roll becomes a 25 buy-in roll without a single bad session.

Record-keeping

You cannot manage a bankroll you are not measuring, and all three inputs to the maths come from records you have to keep yourself.

• Online: a tracker such as PokerTracker 4 or Hold’em Manager 3 imports your hands and reports your win rate and standard deviation directly. Those are the two numbers every table on this page needs.
• Live: log every session — date, stake, hours, result, and a one-line note on the game. A phone app (Poker Bankroll Tracker, PokerStack, or a spreadsheet) is enough. Hours matter as much as results, because your win rate per hour is the only way to compare a live session with an online one.
• Both: record the date you moved up or down and why. When you look back at a bad stretch, the question is always whether it started when something changed.

Honest record-keeping also protects you from the most common self-deception in poker: remembering the sample that suits the story. A tracker has no opinion about whether you are running badly.

Downswing checklist: variance or a leak?

When the graph turns down, there are only two explanations, and they need opposite responses. Variance punishes people who change their game to chase it. A leak punishes people who wait for it to pass. Work through this in order.

1. Measure it, do not feel it. Put your result and your hand count into the variance calculator. Inside the middle 70% of outcomes, there is nothing to explain. Outside the 95% band, variance alone rarely stretches that far.
2. Check your sample honestly. At σ 100 it takes around 960,000 hands to know your win rate within ±2 bb/100. A 20,000-hand losing stretch is compatible with being a solid winner, and a 20,000-hand winning stretch is compatible with being a loser.
3. Check the conditions, not just the cards. Did you move up a level, add tables, change sites, start playing later at night, or start playing while doing something else? Most "downswings" have a date attached.
4. Check your bankroll line. If you are below the move-down number in the table above, drop a level now. Moving down is how you keep the option of playing at all.
5. Review hands, not results. Pick the spots where the money actually went — big pots you lost with second-best hands, river calls, blind defence — and check your decisions against a solver-derived baseline rather than the outcome.
6. Fix one thing at a time. Rebuilding your whole strategy during a downswing guarantees you will not know what helped.

Related

• Poker bankroll calculator — enter your roll and get the stake, the buy-ins and the move-down line.
• Poker variance calculator — downswing odds, risk of ruin and the variance-or-leak test.
• Poker variance explained — standard deviation by format and how big a downswing is normal.
• Poker equity calculator — hand versus range equity.
• Preflop charts — the ranges your win rate is built on.

Frequently asked questions