Bankroll management for online poker — the math

Why buy-in-based bankroll rules exist

Poker is a high-variance game even for skilled players. A winning NL100 cash-game player with a 4 bb/100 win rate experiences a standard deviation of approximately 100 bb/100 — meaning that over a 10,000 hand sample (about a month of part-time play), the player's actual result has roughly a 95% confidence interval of -200 bb to +600 bb. The downside of that interval is two full buy-ins lost despite genuine skill edge.

Over a year of serious play, downswings of 20-30 buy-ins are not unusual even for winning players. A bankroll that cannot absorb a 20-buy-in downswing without forcing a stakes drop is structurally fragile.

The standard heuristics

Cash games: 20-30 buy-ins for the stake you primarily play.

For NL100, that's $2,000-$3,000 dedicated bankroll (where one buy-in is $100 = 100 BB at NL100 — actually NL100 is $0.50/$1 with $100 max buy-in).

The 30-buy-in figure is the conservative end. Players with stable income outside poker (recreational players) can play closer to 20 buy-ins. Players who depend on poker income should hold 30-50 buy-ins. The variance is not lower for professional players; the consequence of a downswing is more severe.

Tournaments (MTT): 100-200 buy-ins for the average buy-in size you play.

MTT variance is much higher than cash-game variance because the prize-pool distribution is top-heavy — most cashes are small (~2× buy-in), large cashes are rare (~100× buy-in for a final-table finish in a 1,000-entry event). A winning MTT player can go hundreds of tournaments without cashing for more than 5× buy-in, and the long-run win rate only manifests over thousands of entries.

The 100-buy-in figure is barely adequate for serious MTT play. Professional MTT grinders typically hold 300-500 buy-ins for their primary buy-in size.

Sit & Go (SNG): 50-100 buy-ins. Variance is lower than MTTs because field sizes are smaller and prize-pool distribution is flatter.

Spin & Go / jackpot SNGs: Variance is structurally extreme due to randomized prize pools. 1,000+ buy-ins is not unreasonable for serious volume. Most players treat these as recreational entertainment rather than a serious income source.

When to move up

The standard rule: move up in stakes when your bankroll reaches 30-50 buy-ins at the new stake (in addition to maintaining the lower-stake bankroll), AND you have a verifiable win rate at the lower stake based on at least 50,000-100,000 hands of cash play or 1,000+ tournaments.

Skipping the win-rate verification is the most common mistake. A player who reached the new stake's bankroll requirement via running hot rather than via genuine win-rate is moving up into stakes their skill level does not justify. The down-move that follows is usually larger than the up-move.

When to move down

The standard rule: move down when your bankroll falls below 15-20 buy-ins at your current stake.

The move-down trigger is editorially harder to follow than the move-up trigger. Players resist moving down because it feels like an admission of failure. The math is the opposite: continuing to play at a stake where the bankroll is structurally fragile is the failure. Moving down preserves capital for future moves up.

How online poker's characteristics change the math

Rake. Online poker rake is materially higher than the equivalent live game (5% with 3 BB cap on cash games at most rooms, versus typical live cardroom $5 max rake on $1/$2 hands which is effectively 1-2% of pot at typical pot sizes). Higher rake means lower win rates and higher variance relative to win rate. Conservative buy-in requirements correspondingly increase.

Multi-tabling. Playing 8 or 16 tables simultaneously is standard at HUD-permitting rooms. Multi-tabling increases hands-per-hour but does not increase win rate per hand — and at high table counts, attention dilution reduces win rate per hand. Players new to multi-tabling typically need to drop one stake while learning the multi-table workload.

Fast-fold (Zoom, Rush & Cash, Boost). Fast-fold variants increase hands-per-hour dramatically (300+ hands/hour vs 60-80 at full-ring tables) but flatten the variance distribution because pre-flop fold-equity considerations dominate over post-flop play. Win rates at fast-fold are typically 30-50% of the equivalent stake's full-ring win rate. Bankroll requirements are similar.

Anonymous tables (Ignition, BetOnline, partypoker). Anonymous-table rooms have lower variance for skilled players because opponents cannot adjust to specific player tendencies via HUD profiling. Win rates can be higher despite the absence of a HUD-driven edge because the recreational pool is genuinely soft. Bankroll requirements may be slightly lower than at HUD-permitting rooms.

A practical framework

For a recreational online poker player with stable income outside poker:

• NL10 or below: 15 buy-ins is fine. Bankroll growth comes from making sound decisions, not from bankroll management.
• NL25 / NL50: 20-25 buy-ins.
• NL100 / NL200: 25-30 buy-ins.
• NL400 and above: 30-40 buy-ins. At these stakes, professionals dominate the player pool and variance per session is larger.
• MTT play at any stake: 100 buy-ins minimum, 200+ for serious volume.

These figures are conservative ends of the standard recommendations and account for the typical online poker player's actual edge being smaller than they believe.

Tracking and discipline

The discipline part of bankroll management is harder than the math part. Three habits:

1. Separate bankroll from operating funds. A poker bankroll lives in the poker account (or a dedicated wallet). Money won at poker that's moved out for living expenses is not bankroll. Money deposited as recreational entertainment is also not "real bankroll" in the disciplined sense.

2. Track results in hands, not in dollars. Win rates per 100 hands and BB/100 are the meaningful metrics. Dollar swings within a session or week are noise.

3. Pre-commit to stakes rules. Decide in advance at what bankroll level you'll move up, at what level you'll move down. Following pre-commitments removes emotional decision-making at the moments when emotional decisions are most likely to be wrong.

Bankroll management is not glamorous. It is the discipline that separates winning players who survive variance from talented players who go broke.