You put in $125 you can afford to lose. One buy-in at this stake is $5.
That's 30 buy-ins of $10, so you can move up to $0.05/$0.10.
You've dropped below 20 buy-ins at this stake, so you go back down. Nothing went wrong. That's what the rule is for.
Updated November 2026
The usual rule of thumb is 20 to 30 buy-ins for no-limit cash games and 100 or more for tournaments. How much you really need depends on how much you win and how much your results swing, and the calculator below shows what each choice risks.
All of this assumes the money is money you can afford to lose. If it isn't, the right bankroll is zero.
It's the money you've set aside only for poker, kept apart from rent and groceries. You measure it in buy-ins: one buy-in is what it costs to sit down with a full stack, usually 100 big blinds.
So in a game with $0.50 and $1 blinds, one buy-in is $100, and 25 buy-ins is $2,500.
This chart shows the chance of losing your whole bankroll, for three players with different win rates. A win rate is written in big blinds per 100 hands.
Look at how quickly the lines drop. A player winning 5 big blinds per 100 hands goes from about 3 chances in 10 of going broke at 10 buy-ins to about 1 in 40 at 30. A small winner needs far more.
Here's the money for 25 buy-ins of 100 big blinds, at the stakes most players start with.
One buy-in is 100 big blinds
It's the chance that you lose your entire bankroll before your edge has time to show. It depends on three numbers: your win rate, your bankroll, and how much your results swing.
The formula is on the right. Swing is measured as standard deviation, and for six-player no-limit a figure around 90 big blinds per 100 hands is a common estimate. You can change it in the calculator.
Risk = e^(−2 × win rate × bankroll / swing²)
Win rate and swing are per 100 hands, and bankroll is in big blinds.
Most entries win nothing, and a few win a lot, so your results swing far more than in a cash game. Going without a real win for 50 tournaments in a row is ordinary.
That's why the usual advice is at least 100 buy-ins. Fields of thousands need more than that.
More than most people guess. Say a live table takes $6 from each pot and deals 30 hands an hour. That's $180 an hour leaving the table, or $20 an hour for each of nine players, shared out over time.
Compare that with a $3 an hour win rate at the same game. That's why a player needs a real edge at low stakes, and why you should know the rake before you sit down.
Set a line, then stick to it. For example, only take money out above 40 buy-ins at your current stake. That keeps enough behind for a bad run and still lets you enjoy some of what you've won.
Whatever you decide, decide it before a good month changes your mind.
You put in $125 you can afford to lose. One buy-in at this stake is $5.
That's 30 buy-ins of $10, so you can move up to $0.05/$0.10.
You've dropped below 20 buy-ins at this stake, so you go back down. Nothing went wrong. That's what the rule is for.
They swing more than cash games and less than big tournaments. A common guideline is 50 to 100 buy-ins. Fast formats with big prize multipliers need more, because most of the winnings come from rare results.
Poker should be something you can walk away from. If you're betting money you need, hiding how much you play or playing to escape something, stop and talk to someone. In the US, the National Council on Problem Gambling runs a free confidential helpline. Most countries have something similar.
Twelve spots in three rounds.
Whatever you can afford to lose. Start at a stake where 20 to 30 buy-ins fits in that amount.
Less strictly. If the game is for fun with friends, one buy-in you're happy to lose is plenty.
It depends on the stakes and the players. Anything clearly above zero over a large number of hands is good, and it takes a lot of hands to know.
Keep only what you plan to play with, and don't mix it with money you need for other things.
Yes, at the very lowest stakes, if you treat it as money you can lose. Just don't move up until the numbers say so.
Yes. It's all one bankroll, as long as you can really move it between them.
Risk of ruin uses the standard formula above, which treats results as a random walk with a fixed win rate and spread. It's a simplification, and real results depend on game conditions. The 20 to 30 and 100 buy-in figures are common rules of thumb, not calculations.
Send them this page.