Bankroll Management Poker: Cash + Tournament Math (2026)

Bankroll Management Poker: Cash + Tournament Math (2026)

Picture this: you sit down at your $2/$5 home game with $800. Three buy-ins later, the math you skipped before sitting down has decided the night for you. Not because you played badly. Not because the deck hated you in some cosmic way. Because$800 was never enough bankroll for $2/$5 — it covered the buy-in but not the variance, and variance always finds the player who showed up under-rolled.

Bankroll management in poker is the boring part of the game nobody wants to read about. It is also the only thing standing between a player who lasts five years at the tables and one who busts in three months. Whether you grind $0.25/$0.50 NLHE online, fire $109 multi-table tournaments on Sundays, or run live$5/$10 once a week, the size of your roll relative to your buy-in — not your reads, not your GTO study — controls how long you stay in action.

In this 2026 guide we will show exactly how much bankroll you need for both cash games and tournaments at every stake level, why MTT players need 3 to 5 times what cash players do at the same buy-in, how your win rate (BB/100 for cash, ROI for tournaments) shifts the required number, and where the 30-buy-in graduation rule and the half-Kelly stake-sizing rule actually come from. Two interactive components — a stake-by-stake bankroll chart and a full poker bankroll calculator — let you plug in your own win rate and hours played to see what your real number should be. By the end you will know whether the bankroll you have actually supports the game you are playing.

TL;DR — Quick Bankroll Reference

Key Numbers You Need to Know

The pattern: tournament bankrolls run roughly 2 to 3 times bigger than cash bankrolls at the same buy-in, because finish-distribution variance dwarfs hand-to-hand variance. Doubling your win rate (1 bb/100 → 2 bb/100) cuts required bankroll by roughly 40%.

Why Poker Bankroll Math Is Different in 2026

Poker bankroll math is not the same conversation as casino bankroll math. Casino games have a fixed house edge — you lose at a known rate, and bankroll math is about delaying the bust. Poker has a player edge (if you are a winner) but the variance is so large that the edge can stay invisible for tens of thousands of hands. Bankroll math in poker is about surviving the noise long enough for the edge to show up.

Cash Game vs Tournament Variance

Cash games have hand-to-hand variance: every hand is a small bet against villain, with standard deviation of roughly 75 bb per 100 hands for a typical 6-max NLHE player. Over 100,000 hands, the worst 5% of outcomes for a 3 bb/100 winner is roughly 30 buy-ins down. That sounds bad until you compare it to tournament variance.

Tournament variance comes from finish-distribution skew. In a typical online MTT, you cash 12 to 18% of the time, final-table 1 to 3%, and win 0.5%. Most of your prize pool concentrates in the top 5% of finishes. Run 200 MTTs in a row without a final table and you are down 60 to 80 buy-ins on a perfectly normal sample for a winning player. Tournament variance is so heavy-tailed that you can be a long-term winner and still spend 60% of your sessions losing.

Live vs Online Volume Effects

Live poker runs 25 to 35 hands per hour. Online single-table runs 70 to 100. Online multi-tabling 6+ tables runs 400 to 600. More hands per hour means more variance per calendar week, which compresses both upswings and downswings into shorter time windows. A live grinder hitting a 30-buy-in downswing might experience it over six months. The same downswing for an online multi-tabler hits in three weeks.

This is why online cash bankroll requirements run 50 to 100% above live at the same nominal stake. The math is the same — the time scale is different. Volume also amplifies skill differentials: tighter online pools compress win rates from the 8 to 15 bb/100 of recreational live games down to 2 to 5 bb/100 even for strong online regulars. Lower edge plus higher volume = bankroll math that punishes anyone playing under-rolled.

Cash Game Bankroll Math

Cash games are the easier format to bankroll because variance is lower per session and you can quit any time. The math reduces to one inequality: does your stack cover the expected loss in your worst 5% of 100k-hand stretches? If yes, your risk of ruin is under 5%. If no, you are gambling with your bankroll on top of gambling with the cards.

The 20/50/100 Buy-In Rule by Skill Level

The traditional 20 buy-in rule was set in the early 2000s when typical online win rates ran 8 to 15 bb/100 at the micros. Modern games — even at NL2 and NL5 — have tightened to 1 to 5 bb/100 for typical winners. Lower edge means higher variance relative to the edge, which means the 20 buy-in floor is now too aggressive for most non-pro players.

The modern tiering looks like this:

• Recreational (poker is a hobby, you can reload): 20 to 30 buy-ins. You absorb downswings by waiting for next paycheck. The 20 buy-in floor still applies as the absolute minimum to avoid going broke from one cold session.
• Winning regular (poker is supplementary income): 40 to 50 buy-ins for live, 50 to 75 for online. You cannot reload casually because that would mean dipping into life money. Variance buffer must be real.
• Full-time pro (poker pays bills): 100 buy-ins minimum, 150+ for tournaments. Bankroll funds living expenses on top of variance — going broke means not paying rent.

Stake-by-Stake Cash Bankroll Table

Here is the modern recommendation for online NLHE cash by stake level, assuming you are a winning regular at 3 to 4 bb/100:

Live cash is roughly 60 to 70% of these numbers because hands-per-hour are 4x lower. Live $1/$2 with $200 max needs$4,000 (20 BI) recreational, $8,000 (40 BI) serious,$20,000 (100 BI) pro.

How Win Rate (BB/100) Affects Required Bankroll

The relationship between win rate and required bankroll is non-linear. Cutting your win rate in half doesn't double the bankroll — it more than doubles it, because variance stays roughly constant while edge shrinks. The standard 75 bb/100 standard deviation barely moves whether you win at 2 or 6 bb/100.

This is why micro-stakes grinders need bigger bankrolls than mid-stakes regulars. Once you reach NL100+ where pool skill differential is wider, win rates concentrate around 3 to 5 bb/100 for solid winners and 30 to 50 buy-ins becomes adequate. At NL2 to NL10 where rake eats half your edge, you need 75 to 100 buy-ins or you are bleeding through downswings faster than you can build the roll.

Tournament Bankroll Math

Tournaments need much bigger bankrolls than cash games because the win-rate distribution is fundamentally different. In cash, your big-blind-per-100 number compounds smoothly across hands. In tournaments, your ROI compounds across events with extreme right-tail concentration — most events you bust before the money, a few you cash for 1 to 3 buy-ins, and rare deep runs deliver 20 to 100 buy-ins in a single session.

Why MTTs Need 100+ Buy-Ins (Variance Amplification)

Run 100 simulated $33 MTTs for a winning player at 30$33 × 30% × 100 = $990. The 90$1,500 to +$3,500. That is the math. Variance dwarfs expected value over 100 events.

Now stretch to 1,000 events. Expected $9,900, confidence interval -$2,000 to +$22,000. The interval gets relatively narrower (variance grows as √N while EV grows as N), but you still need to absorb the downside before the upside arrives. A 20-buy-in downswing in MTTs is a normal week. A 50-buy-in downswing happens to every winning player at least once per year.

MTT vs STT vs Hyper-Turbo: Bankroll Differences

Different tournament formats have different variance profiles:

• Single-table SnGs (9 to 10 players, top 3 paid): variance is moderate. 50 buy-ins is enough for a winning regular at 5 to 10% ROI. Field is small, payout structure is flat (50/30/20 split), so finish distribution is closer to normal than MTT.
• MTT (200 to 1,000 entries): variance is high. 100 buy-ins is the floor, 200 is comfortable. Top-heavy payouts (winner gets 18 to 25% of prize pool) plus low cash rate (12 to 15%) creates the heavy right tail.
• Large-field MTT (1,000+ entries): variance is extreme. 200 to 300 buy-ins minimum for a winning regular. Sunday Million types (5,000+ entrants) push the bankroll requirement even higher because deep-run probability is roughly 0.5% even for winners.
• Hyper-Turbo MTT (5-minute blind levels, near-shove play): variance is enormous because of fast-blind structures forcing all-in equity races. 300 to 600 buy-ins, even for a +ROI player.
• Spin & Go / Jackpot SnG: variance is dominated by the jackpot multiplier roll, not the play. 50 to 100 buy-ins works because most variance comes from the multiplier distribution, not the heads-up play.

Stake-by-Stake Tournament Bankroll Table

For online MTTs with 200 to 1,000 entrants and a winning regular at 30 to 50% ROI:

Live MTTs need 50 to 70% of these numbers because field sizes are smaller and overlay (added prize money from rake structure) is sometimes positive at smaller live tournaments, which boosts effective ROI.

Bankroll Requirements by Format: Visual Comparison

Reading the Chart: What "95% Survival" Means

The chart shows recommended bankroll for each format scaled to the same survival probability — 95% chance of not going broke over 100,000 hands (cash) or 1,000 events (tournaments) at typical winning rates. Notice how a $33 MTT needs roughly the same bankroll as$5/$10 NLHE cash even though the buy-in is 100 times smaller. That is variance amplification in action: tournament finish distribution is so top-heavy that the same dollar of EV requires far more buffer than a cash game dollar.

The lime tier (cash games) sits below the dashed line at roughly 50 buy-ins. The green tier (small-field MTTs and SnGs) sits at 100 buy-ins. The yellow tier (large-field MTTs) sits at 200+ buy-ins. The red tier (hyper-turbos) is the variance killer — 300 to 600 buy-ins required, which is why most pros avoid this format unless they have a clear, measurable edge.

Worked Examples: Real Player Scenarios

Numbers in tables are abstractions. Here is what the math looks like for two typical players plugging real win-rate, volume, and bankroll data into the equations.

Example 1: $5/$10 NLHE Cash Game Player at 3 BB/100

The Setup and Hourly EV

Player A grinds $5/$10 online NLHE 6-max. Win rate is 3 bb/100, std dev 90 bb/100. Plays 6 tables at 80 hands per table per hour = 480 hands per hour. Bankroll is $50,000.

Hourly EV = 480 hands × 0.03 bb/hand × $10/bb =$144/hour. Over 30 hours per week, expected weekly profit is $4,320. That is the expected number — actual swings will dominate any individual week.

Risk of Ruin Over 100k Hands

100,000 hands at 480 hands/hour takes 208 hours of play. Expected return: 100,000 × 0.0003 × $1,000 =$30,000. Std dev over 100k hands: √(100,000) × 90 × $0.10 =$2,846 in big blinds, or $28,460 in dollars.

The 5% worst case: $30,000 − 1.645 ×$28,460 = -$16,827. Player A's$50,000 bankroll covers this comfortably with 33 buy-ins of remaining cushion. Risk of going broke before the 100k hands play out is below 1%. This player is properly bankrolled — for a live tracking workflow, they could plug their session results into our poker bankroll tracker to monitor real-time stack-vs-buy-in ratios.

If Player A had only $20,000? Same math, different verdict. The 5$16,827 leaves $3,173, which is 3 buy-ins — they would be forced to drop down to$2/$5 long before recovery. Bankroll at 20 buy-ins for$5/$10 is gamble, not management. Same logic applies if you are running our universal bankroll calculator on any other variance game — the inputs change but the survival math doesn't.

Example 2: $33 MTT Grinder at 130% ROI

Expected Monthly EV and Variance

Player B fires $33 online MTTs 5 nights per week, 4 events per session = 20 events/week, 80/month. ROI is a strong 30$33. Bankroll $5,000.

Expected monthly profit: 80 × $33 × 0.30 =$792. That is the EV. The variance is where it gets brutal.

For typical online MTTs, std dev per event is roughly 1.8 buy-ins (about $59). Over 80 events: total variance = 80 × 1.8² = 259, so std dev =$59 × √80 = $528. The 5$792 − 1.645 × $528 = -$76. Yes, you can be a 30% ROI MTT regular and lose money in a normal month.

Stretch to a year: expected $9,500, monthly std dev sums to roughly$1,830 annually. The 5% worst annual outcome is +$6,500 — still positive over a year, but with monthly swings ranging from −$1,200 to +$3,500. Player B's$5,000 bankroll ($5,000 /$33 = 150 buy-ins) is adequate but tight. Pro level would be 200+ buy-ins ($6,600+).

If Player B drops to $3,300 (100 BI), they are at the floor. One bad month and they are below 80 buy-ins, which forces a drop to$22 events. Use the bankroll calculator to test different ROI/BI combinations against your own variance tolerance — the math generalizes across formats once you know your true win rate.

Moving Up, Moving Down, and Shot-Taking

Stake movement is where most bankroll plans die. Players move up too fast on a heater, refuse to move down on a downswing, and end up bouncing between stakes they cannot beat or surviving in. Three rules handle 95% of the cases.

The 30-Buy-In Graduation Rule

To move up to the next stake, hit 30 buy-ins for that new stake while still maintaining 50 buy-ins for your current stake. Example: at $1/$2 with $10,000 bankroll, you have 50 BI for$1/$2 ($10,000 / $200 = 50). To move up to$2/$5, you need 30 BI for$2/$5 ($15,000) plus the safety net for $1/$2. So move up at $15,000, not$10,000.

This is more conservative than the 20 BI shot-take rule because moving up means staying up. Shot-taking is one session at the higher stake; graduation is permanent commitment. The 30 BI floor at the new stake plus the safety net at the lower stake gives you about 6 months of runway to learn the level before bankroll math forces a drop.

The 20% Decline Downswing Trigger

Drop down when your bankroll falls 20% from peak. If you peaked at $20,000 grinding$1/$2, drop to$0.50/$1 the moment you hit$16,000. This is more aggressive than the traditional 25% rule because modern variance is wider — by the time you have lost 25%, you might be in the middle of a 35% trough.

Moving down is not failure. It is the correct mathematical response to bankroll erosion. The faster you drop down, the more bankroll you have at the lower stake, the more comfortably you can grind back. Refusing to move down ego-tilts you into doubling the loss before the math forces a drop anyway.

The 15/2 Shot-Taking Strategy

Shot-taking is different from graduation. A shot is a planned 2 to 5 buy-in attempt at the higher stake to test your skill-edge there. Rule: shot-take only when you have 15 buy-ins for the higher stake AND 2x your normal bankroll for current stake.

Example: $0.50/$1 grinder with $5,000 bankroll wants to shot$1/$2. Need 15 BI for$1/$2 ($3,000) plus 2x normal $5,000 =$10,000. Take 5 buy-ins ($1,000) and play at$1/$2. If you lose all 5, drop back to$0.50/$1 immediately. If you win, keep playing the higher stake until your bankroll naturally graduates by the 30 BI rule.

The same shot-taking logic applies across formats. If you are tempted to mix bankroll between live poker and other casino-style variance games — say, occasional video poker bankroll strategy sessions where the math is fundamentally different — keep separate bankrolls. Mixing two different variance distributions inside one stack means the higher-variance game eats the lower-variance game's buffer.

The Kelly Criterion Myth in Poker

Kelly Criterion sizing — bet a fraction equal to your edge divided by your variance — is famous in finance and sports betting but breaks down in poker. The formula assumes you know your edge precisely. In poker, you don't. Win rates fluctuate ±50% across different game pools, table compositions, and skill drift over time.

Why Half-Kelly Is Standard for Pros

Full Kelly applied to a 3 bb/100 winner with 90 bb/100 std dev says you should risk 3/(90²) = 0.037% of bankroll per hand. That works out to maximum bet of $37 on a$100,000 roll, which fits exactly the 50 to 100 buy-in tier.

But the formula assumes your 3 bb/100 estimate is exact. If your actual win rate is 1.5 bb/100 (half what you thought), Full Kelly produces 100% bankroll variance — you'll experience drawdowns that feel like ruin even when math says you survive. Most poker pros use Half-Kelly or Quarter-Kelly: bet 50% or 25% of what Full Kelly says, accepting slower bankroll growth in exchange for narrower drawdown ranges. The same framework applies to other strategy games — see how bankroll management blackjack handles the equivalent tradeoff for card counters, where bet-spread variance exceeds the simple Kelly assumption.

Stop-Loss Rules That Actually Work

Stop-loss in poker is not the same as stop-loss in trading. In trading, you exit a position at a defined number to limit downside. In poker, you exit a session because tilt makes future decisions worse than the variance you have already absorbed. The math behind a stop-loss in poker is psychological, not statistical.

Practical stop-loss rules pros use:

1. 3 buy-in cash session stop: lose 3 max buy-ins in one cash session, end the session. Period. Variance happens; tilt-spew on top of variance is the killer.
2. 2-hour MTT cool-off: bust 2 MTTs in a row in the bubble or early stages with the same hand pattern, take a 2-hour break before re-entering. Most early-stage MTT bust-outs are tilt patterns, not variance.
3. Weekly profit-protect: if you are up 30+ buy-ins on the week, withdraw 50% of profits from your poker site to bankroll-locked storage. This prevents over-reinvestment if you crash from the heater.

The point is not to limit losses mathematically — that is the bankroll math itself. The point is to remove emotional variance from compounding on top of card variance. A bankroll built for variance is not built for tilt-spew.

Interactive Poker Bankroll Calculator

Plug your format, stake, win rate, and starting bankroll into the calculator below. It returns the bankroll size needed for 95% survival, expected weekly EV, downswing depth estimate, and hours/events until your next stake-up — based on the same formulas used in the worked examples above.

The calculator handles cash games (NL Hold'em with bb/100 input), MTTs (with ROI input), and STTs (with ROI input). Adjust the volume slider to see how more hours per week shrinks the time-to-double but does not change the underlying bankroll requirement — bankroll math is volume-invariant in expectation, only calendar time changes.

FAQ