Poker Bankroll Management: Buy-Ins, Variance, and Moving Up
Poker bankroll management is a different discipline from casino bankroll management. In house-edge games the bankroll only decides how long the ride lasts; in poker a skilled player can hold a genuine positive edge — but that edge arrives wrapped in variance brutal enough to bankrupt winners who under-roll themselves. Your bankroll is the engine that converts skill into money. Blow the engine and the skill is worthless.
Play within your roll →Poker bankroll is not casino bankroll
Against roulette or any fixed-house-edge game, no bankroll strategy changes the expected outcome — management there is purely about pacing losses and protecting entertainment budget. Poker is categorically different because you play against other people, not the house, so a positive win rate is achievable. That changes the bankroll's job: it is no longer a spending allowance but working capital, sized so that normal statistical noise cannot destroy it before your edge expresses itself. The professional framing is risk of ruin — the probability that variance busts you despite a positive expectation. Skill sets your expected earn; bankroll sizing sets whether you survive long enough to collect it. One honest caveat before the math: most recreational players are not long-term winners, and until your tracked results prove otherwise, treat poker money as entertainment spend you can fully afford to lose.
How many buy-ins do you actually need?
The conventional guideline for no-limit cash games is 20 to 40 buy-ins for your regular stake — 20 as an aggressive floor for proven winners in soft games, 40-plus for thinner edges, tougher lineups or anyone who cannot easily reload from outside income. PokerNews' bankroll guidance sits in this range, and the logic behind the spread is simple: the required cushion scales with the ratio of variance to win rate. A big winner in a soft game rebuilds from downswings quickly and can run lighter; a marginal winner needs a far deeper buffer because downswings dig deeper relative to the earn rate. Tournament players need dramatically more — often 100-plus buy-ins — since most entries cash nothing. Practical version: at a table with 100bb buy-ins, a 30-buy-in roll means your bankroll should be about 3,000 big blinds for that stake. If it is not, you are not rolled for the game; you are gambling on variance behaving.
Downswing math: winners lose, a lot, for a long time
Here is the part every player underestimates until it happens to them. A solid winner earning 5 big blinds per 100 hands with a typical standard deviation of around 80bb/100 will, over a 100,000-hand stretch, routinely experience 10-buy-in downswings; 20-buy-in downswings are not rare, and break-even stretches of tens of thousands of hands are entirely normal — all while playing well. That noise is exactly the variance concept from finance applied to poker: short samples are dominated by dispersion, not expectation. The implications are behavioral. First, a downswing is not evidence you have become bad, and a heater is not evidence you have become great — only large samples speak. Second, the 20-to-40-buy-in rule exists precisely because these swings are guaranteed, not merely possible. Third, if a 15-buy-in downswing would wreck your finances or your emotional state, you are playing too big regardless of what the guideline says.
Moving up: roll, confidence, and sample
Move up when three conditions align, not one. Roll: you hold the full buy-in requirement for the new stake — 25 to 40 buy-ins at the higher level, not at your current one. Confidence: you are playing your best game and genuinely want the tougher challenge, not fleeing boredom or chasing losses. Sample: your win rate at the current stake is proven over a meaningful volume of hands, not a 5,000-hand heater. Then take the shot with rules written in advance: allocate a fixed shot budget, commonly 3 to 5 buy-ins, and commit to dropping back down if the shot money is gone. Pre-committed rules are the entire trick — they convert an emotional mid-session decision into a mechanical one you made while calm. Shot-taking with defined stop-losses lets you probe higher stakes repeatedly without ever endangering the core roll that funds your regular game.
Moving down without ego — and tracking honestly
Everything above is worthless if you cannot execute the unglamorous half: moving down. Set the trigger in advance — for example, drop a level whenever the roll falls below 20 buy-ins for the current stake — and treat it as automatic, because in the moment your ego will negotiate hard. Moving down is not demotion; it is the mechanism that makes aggressive shot-taking safe. Pair it with honest tracking: record every session's stake, hours, and result, because memory is a flattering liar that recalls the hero calls and forgets the Tuesday bleed. Untracked players systematically overestimate their win rate, then size their bankroll for the imaginary number. On Llamabet the ledger is effectively kept for you — every Hold'em hand is chained and verifiable at /poker/verify, buy-ins start from a 1 SUI minimum so moving down always has somewhere to go, and instant on-chain payouts mean your bankroll figure is always real, current, and self-custodied rather than an IOU on a casino database.
Tilt: the bankroll's real enemy
Most bankrolls are not destroyed by variance operating on good play — they are destroyed by tilt converting a normal downswing into a catastrophic one. The sequence is depressingly standard: bad beat, emotional escalation, looser calls, bigger stakes to 'win it back', and a 6-buy-in bad night becomes a 20-buy-in crater that the bankroll math never modeled. Chasing losses is the same fallacy that ruins casino gamblers, and no buy-in formula survives it. Defenses are mundane and effective: a hard stop-loss per session (say 3 buy-ins), a rule against moving up while stuck, mandatory breaks after big pots lost, and quitting the moment you notice yourself playing hands you would fold on a normal night. Treat tilt control as part of bankroll management, not a separate soft skill — the roll only works if the person managing it is the calm version of you. And always: play with money whose loss changes nothing about your life.
Frequently asked questions
How many buy-ins should you have for poker cash games?
The conventional guideline is 20 to 40 buy-ins for your regular stake: 20 for proven winners in soft games who can reload, 40 or more for smaller edges, tougher games or anyone whose roll is hard to rebuild. Tournament players need far more — often 100-plus entries — because of higher variance.
How big can a downswing get for a winning poker player?
Larger than intuition suggests. A solid winner at 5bb/100 will routinely see 10-buy-in downswings, occasionally 20-plus, and can run break-even for tens of thousands of hands while playing well. This is normal variance, and it is exactly why the multi-buy-in bankroll rules exist.
When should you move up in stakes in poker?
When three things align: a full bankroll for the new stake (25 to 40 buy-ins at that level), a proven win rate over a real sample at your current stake, and a calm reason to move. Take the shot with a pre-set budget of 3 to 5 buy-ins and a firm rule to move back down if it is lost.