September 1, 2026
Every bankroll chart you have seen gives you a number of buy-ins. Twenty for cash. Thirty if you are cautious. A hundred for tournaments. The figures get passed around because they are easy to pass around, and because for a lot of players they are roughly right.
The trouble starts when you ask what they are roughly right for.
Where the numbers come from
A bankroll requirement is the answer to a risk of ruin question: how much do I need so that a bad run doesn’t end me. Feed in a win rate, a standard deviation and the chance of going broke you are willing to accept, and it hands back a number of buy-ins.
The figures people repeat for live no limit cash sit somewhere around twenty to thirty. Run the arithmetic for a solid winner and you land in the same neighborhood, which is why the folklore has survived. For a five big blind per hundred winner accepting a five percent risk of ruin, the answer is twenty-four buy-ins.
Why one chart cannot be right
Halve that win rate and the requirement doubles to forty-eight. Same game, same swings, same stake. The only thing that moved is how hard you are beating it.
That is the assumption nobody checks. A bankroll chart isn’t telling you how many buy-ins a poker player needs. It is telling you how many buy-ins a player with one particular edge needs, and it has no idea what yours is. Neither do you, precisely, which is a separate and more uncomfortable problem.
In practice it means the chart is most wrong for the people leaning on it hardest. A confident new regular who is beating 1/2 by a little rather than a lot needs more than the number on the chart, not less.
What live play changes
Live is slow. Twenty-five hands an hour, against several hundred for someone running four tables online. That cuts both ways.
The swings per hour are smaller, so a live session rarely produces the sort of graph that makes online players quit. But the sample builds so slowly that a bad stretch is measured in months instead of days, and a roll has to survive the calendar and not just the arithmetic. Ten thousand hands is most of a live year.
Live games also tend to be softer, which raises win rates and lowers the requirement for the same risk. That is the one live specific factor working in your favor, and it is why the folk numbers feel conservative to players who mostly sit in soft rooms.
Rules that survive contact
Three that hold regardless of which number you settle on.
Keep the roll separate from money you need. A bankroll sharing an account with rent isn’t a bankroll, it is a balance. The whole calculation assumes you could lose all of it without changing how you play, and nobody manages that with rent in the same pot.
Move down before you have to. Coming down a stake at forty percent of your roll is a decision. Coming down at ten percent is a rescue, and by then you are playing scared, which costs you the edge the calculation was built on.
Set the threshold when you are nowhere near it. A rule written on a Tuesday afternoon is worth more than one improvised at midnight after a three buy-in night.
Tournaments are a different animal
The hundred buy-in figure for tournaments isn’t caution, it is arithmetic. Tournaments pay a small number of large prizes, so the distribution is a long thin tail on the good side and a great many zeroes everywhere else. You can play well for a year and show a loss.
Which is why running both out of one roll and one number goes wrong. The cash roll and the tournament roll answer different questions and they move at different speeds.
The part that makes any of it work
Every number above depends on inputs you have to supply. Your win rate at a stake. Your actual hours. Your real results, including the sessions you would rather not count.
Without those you aren’t managing a bankroll. You are reading a number off a chart written for a player who might not be you, and hoping.
That is why StackXo has a Bankroll Advisor instead of a chart. It starts on textbook assumptions when you have nothing to go on, then leans on your own numbers as sessions accumulate. And when your results at a stake aren’t clearly positive it declines to give a figure at all, because a losing game isn’t a bankroll problem and no amount of roll will fix it.
Whichever app you keep those results in, and there are several kinds worth telling apart, the requirement is the same: enough sessions, honestly recorded.
The right number of buy-ins is the one that matches an edge you can show. Everything before that is a chart.