Bankroll management is the system of rules that determines which stakes you are allowed to play at a given bankroll size, when to move up, when to move down, and how much money you can safely take “for life.” Without this system even a winning player with a 5bb/100 winrate will sooner or later go broke — variance in poker is high enough to eat any realistic bankroll if you play too high relative to its size.
In 2026 bankroll requirements have noticeably tightened compared to the 2010s. The fields are tougher, winrates are lower, and variance in cash and MTT is consistently high. If ten years ago 20 buy-ins were considered enough for NL cash, today the minimum safe level for a regular is 30–50 BI depending on format and winrate.
Why Bankroll Management Is Needed
The main task of bankroll management is to protect you from going completely to zero during a losing stretch. The standard deviation of winrate in NLHE 6-max at NL100 is 80–100 bb/100. That means within one standard deviation over 10,000 hands you can deviate from expectation by 800–1000 bb, or 8–10 buy-ins. Over 100,000 hands — by 25–30 buy-ins. Without a cushion of several dozen buy-ins, even a strong reg risks going broke purely statistically.
The second task is psychological. When you play on a bankroll that is 20 BI at your stake, every buy-in feels like 5% of your total capital. That pressure inevitably affects decisions: you start folding more often in marginal spots, shift toward a defensive style, lose part of your theoretical edge. Playing “on your own money” is a scenario where your real winrate is always lower than your calculated one.
The third task is career planning. Clear rules for moving up and down give you a clear development path: you know how many hands you have to grind at your current stake, what winrate to show and what bankroll to build before moving up. Without this plan, a player has no clear success criterion and progress becomes chaotic.
Formulas: How Many BI You Need for Each Format
The minimum safe bankroll is the product of the number of buy-ins and the specific buy-in of the stake. The numbers differ by format because variance is fundamentally different.
Cash NLHE 6-max
The base recommendation in 2026 is 30 BI for a reg with a winrate of 4–6 bb/100, 40 BI for a reg with 2–4 bb/100, 50 BI and up for those who barely beat the stake (1–2 bb/100). A regular at NL50 with a 3 bb/100 winrate must have a minimum of $2,000 on the bankroll, otherwise the risk of a 20+ BI downswing over half a year turns from theoretical into real.
Cash Zoom / Rush & Cash / Snap
Fast formats provide more hands per hour but a lower winrate due to the tighter play of the pool. Realistic winrate values in 2026 on Zoom NL50 are 1.5–3 bb/100, versus 3–5 bb/100 in regular cash. This means relative variance is higher and the bankroll should be 10 BI larger: 40–60 BI at the stake.
PLO 6-max
Pot-Limit Omaha is the most “swingy” of the mass formats. Standard deviation here is 150–200 bb/100, 1.5–2 times higher than in NLHE. The corresponding bankroll is a minimum of 50 BI, optimally 75–100 BI at the stake. For PLO100 that means a $5,000–$10,000 bankroll.
MTT and SnG
MTT has very high variance due to the lottery nature of top payouts. The standard recommendation is 100–200 buy-ins in the bank. If you mostly play $11 and $22 MTTs, the bankroll should be in the $2,200–$4,400 range. For a $33–$55 regular — already $5,000–$10,000.
SnG is simpler: variance is roughly half that of MTT. 50–100 BI is enough for safe play. Hyper-SnG requires an expanded bank — 100–200 BI due to short structures and hyper variance.
Spin & Go and Hyper Turbo MTT
Spin & Go and their equivalents are the most variance-heavy format in modern poker. A ×1000 multiplier hits once per 100,000 tournaments, which creates a huge tail in the distribution. The realistic recommendation is 200+ buy-ins at the stake, and for a serious Spin $7–$15 grinder that means a minimum $1,500 bankroll for $7 Spins.
Buy-In Table by Stakes and Formats
A baseline cheat sheet for a regular with a winrate “above average for the pool” (roughly the 50th percentile among winning players of the format). The numbers are minimum safe values; for shot-takes and temporary expansion you can dip 5–10 BI lower.
| Format | Typical winrate | Min BI | Bankroll example |
|---|---|---|---|
| NLHE 6-max NL10 | 5–8 bb/100 | 30 BI | $300 |
| NLHE 6-max NL50 | 3–5 bb/100 | 40 BI | $2000 |
| NLHE 6-max NL200 | 1.5–3 bb/100 | 50 BI | $10000 |
| Zoom NL50 | 2–4 bb/100 | 50 BI | $2500 |
| PLO50 | 2–4 bb/100 | 75 BI | $3750 |
| PLO200 | 1–3 bb/100 | 100 BI | $20000 |
| MTT $5–$22 | 5–15% ROI | 100–150 BI | $1100–$3300 |
| MTT $55–$109 | 3–10% ROI | 150–200 BI | $8000–$22000 |
| SnG $5–$15 | 5–10% ROI | 50–100 BI | $500–$1500 |
| Hyper SnG $7–$30 | 2–7% ROI | 200 BI | $1400–$6000 |
| Spin & Go $3–$7 | 5–15% ROI | 300 BI | $900–$2100 |
Variance and Standard Deviation
Any bankroll management without an understanding of variance is magical thinking. The standard deviation in poker tells you how much the actual result can deviate from the expected one over a finite sample.
In NLHE 6-max the typical standard deviation (σ) is around 90 bb/100. This means 68% of all 10,000-hand stretches fall within “expectation ± 90 bb.” With a winrate of 3 bb/100, expectation over 10,000 hands = +300 bb (3 BI). Standard deviation in this window = 900 bb (9 BI). So in 16% of cases over 10,000 hands you’ll get a result worse than −6 BI, and that’s normal.
This is mathematics. Not a curse, not bad form, not “the cards aren’t running” — just statistics. That is why the 20-buy-in “safe” bankroll from the 2010s in modern poker leads to ruin in 25–30% of scenarios.
Shot-Takes and Move-Up Rules
A shot-take is an attempt to play a stake above your current one without waiting for the full bankroll. It is a risky but sometimes necessary move — for example, if a specific table opened with an obvious fish, or if you want to test yourself against a new structure.
- Start shot-takes only with 40+ BI on the current stake (the reserve that compensates for a partial loss of the shot).
- Shot size — maximum 3–5 BI on the new stake. If you lose more than 50% of that volume — return back.
- Shot-takes require pure mode: only this one table above the current stake, with no parallel play at the old one. Otherwise you are psychologically scattered and play worse at both stakes.
- After a successful shot-take (plus 3–5 BI on the new stake), play it in mixed mode 1–2 tables together with the main stake until the bankroll reaches a full 40 BI on the new one.
Moving down is mandatory when you reach 25 BI on the current stake. This is not shame or a signal about the player’s level, it is the mathematically correct decision for protection from going to zero. The strongest regulars in the world climb up and down the ladder dozens of times over a career.
Psychological vs Mathematical Bankroll
The mathematical bankroll is the amount that is statistically sufficient for long-term play with an acceptable Risk of Ruin. The psychological bankroll is the amount you can play on without losing emotional balance after every lost buy-in.
These numbers often differ by a factor of 2–3. Mathematically, 50 BI ($5,000) is enough at NL100 with a 2 bb/100 winrate. Psychologically, many players are comfortable starting NL100 only with $8,000–$10,000, because otherwise every drop even of 5 BI feels like a catastrophe and triggers tilt.
If you know you are tilt-prone, add 20–30% to the mathematical minimum. This is an investment in decision stability that pays off through a higher real winrate.
What to Do During a Long Downswing
A downswing of 15–25 BI over 20,000 hands is routine statistics, not a reason to overhaul your life. Actions during such a downswing:
- Confirm the facts: check your winrate graph in the tracker for the last 30 days. If it deviated from the expectation line by no more than 2σ — it’s variance, and the strategy stays as is.
- Move down one step. This isn’t humiliation, it is a technical measure. Most regulars do this 2–3 times a year.
- Cut volume by 1.5–2 times. Fatigue amplifies technical mistakes, and during a downswing the psychological cost of each mistake is higher.
- Run a game review: 20–30 lost pots from the past month for analysis through a solver, a minimum of 3 disputed spots a week in a group or with a coach.
- Don’t try to “win it back” — an attempt to recover the loss in a forced single session leads to an even bigger downswing in 80% of cases.
How Many Hands You Need to Estimate Winrate
Most players overestimate the accuracy of their own winrate. Real confidence intervals over typical samples are much wider than they look on the tracker graph. Because of this, many make decisions about moving to a new stake or changing format based on a falsely high winrate, not understanding that the deviation from the true value can be 2–3 bb/100.
Base rule: to estimate winrate to within ±1 bb/100 you need a minimum of 100,000 hands on a stable stake without interruptions. For ±2 bb/100 accuracy — 25,000–30,000 hands. Smaller samples give only a rough indicative estimate that cannot justify career decisions like turning pro.
This means a practical conclusion: you move to a new stake not “after a good month,” but after a full cycle of at least 25,000 hands on the current stake with a winrate above 2 bb/100 and a normal distribution (without spikes from 1–2 lucky sessions).
Risk of Ruin (RoR) and How to Calculate It
Risk of Ruin (RoR) is the mathematically grounded probability that you will run your bankroll to zero given parameters of winrate, variance, and starting bank. RoR is the main argument for or against any bankroll decision.
The base RoR formula for poker: RoR = exp(−2·WR·BR / σ²), where WR is winrate in bb/100, σ is standard deviation in bb/100, BR is bankroll in big blinds. With WR=3 bb/100, σ=90 bb/100, BR=5000 bb (50 BI) we get RoR ≈ 0.7%. With BR=3000 bb (30 BI) — already 4.5%, six times higher.
Practical conclusion: the difference between a “sufficient” and “insufficient” bankroll is not a matter of personal preference but specific percentage points in the probability of going broke. 50 BI instead of 30 BI at NL100 reduces the risk of ruin by a factor of six — that’s an objective difference that justifies every invested dollar.
The Link Between Bankroll and Liferoll
The liferoll is the money with which you pay for life outside the game: rent, food, communications, leisure. Professional players separate bankroll and liferoll with a strict boundary: no withdrawals from the poker bank for household expenses outside of the schedule.
Standard scheme for a cash or MTT regular: a 6-month liferoll on hand (personal financial reserve), plus a full bankroll by the rules of 30–50 BI on the main stake. This means that to live at NL200 as the main stake, a regular must have $10,000 of bankroll plus $15,000–$25,000 of liferoll depending on the cost of living in their city. A minimum of $25,000 in total reserve.
If you cannot separate these two banks (for example, your only source of income is poker, and you have no half-year financial cushion), the right path is to keep an extended reserve in the poker bank (70–100 BI instead of 30–50) and regularly withdraw part of the profits to the liferoll. Without this discipline, at the first serious downswing the player falls into the trap of “I’ll have to drop the stake, but then I can’t cover rent” — and makes bad decisions out of desperation.
Bankroll for Different Playing Styles
Not all players are equally exposed to variance. Playing style substantially affects the required bankroll size even at the same winrate.
A tight-aggressive regular with a standard VPIP/PFR of 22/19 and a low 3-bet level (around 7%) usually has a standard deviation of 80–90 bb/100. A loose-aggressive regular with VPIP/PFR 28/24 and 3-bet aggression of 12%+ often shows variance of 110–130 bb/100. The difference in required bankroll is about 30%: where a tight reg needs 40 BI, a loose-aggressive one needs 55.
Hyper-aggressive cash players and LAG style create the maximum variance. At NL200 with a 2 bb/100 winrate and σ=120 bb/100, the safe bankroll is 60–70 BI. This means switching from a tight style to a high-variance style requires not just a strategic rework but also an increased bankroll. Many players make the mistake of changing style without increasing the bank — and fall into a 30 BI level downswing their bankroll can’t withstand.
For an MTT regular, a similar effect comes from a bet on high-variance structures: PKO, Mystery Bounty, Hyper Turbo. All of them give a winning player a higher ROI, but variance in them is 1.5–2 times higher than in regular Freezeout MTTs. The bankroll must be increased proportionally.
Any bankroll management strategy relies on honest data about your own winrate. Mining hands from your pool at a specific stake and a database of opponent statistics in the tracker let you separate “I’m running bad” from “the field has objectively gotten tougher”: comparison with the average numbers of a regular at the level shows whether your downswing is pure variance or a signal for a real strategy review.