Bankroll Requirements for Live Cash Games ($1/$2, $1/$3, $2/$5)

Ask five forums how much bankroll you need for $1/$2 live and you'll get answers between $2,000 and $20,000, delivered with equal confidence. Both camps have a point, because the question is incomplete. Bankroll requirements for live cash games depend on your win rate, your game's variance, and what going broke would actually cost you. The good news: all three fit into a formula you can run on a napkin, and this post runs it for the three stakes most live players actually play.

The short answer: buy-in tables

Standard 100bb buy-ins and the rolls behind them:

• $1/$2 ($200 buy-in): 20 buy-ins = $4,000 · 30 = $6,000 · 40 = $8,000
• $1/$3 ($300 buy-in): 20 = $6,000 · 30 = $9,000 · 40 = $12,000
• $2/$5 ($500 buy-in): 20 = $10,000 · 30 = $15,000 · 40 = $20,000

Two adjustments before you trust these. First, check your room's cap: many $1/$2 games cap the buy-in at $300 (150bb), and uncapped games play much deeper, which raises variance. Second, count what the game actually costs after rake. A $1/$2 game raking 10% to $6 plus a $1–2 tip per won pot takes roughly 2–3 big blinds per hour off your win rate — the difference between a 10bb/hour crusher and a 7bb/hour solid winner.

What hourly variance looks like at these stakes

Live poker deals 25–33 hands per hour. That's the number everything else hangs on, because it makes live variance a per-hour problem rather than a per-hand one.

A strong $1/$2 player wins about 8bb/hour — $16 — with an hourly standard deviation around $180. Read that again: a normal one-hour swing is eleven times the hourly expectation. Over a 40-hour month, the expectation is +$640 with a standard deviation near $1,150, which means this clearly winning player books a losing month about three times a year. Not because they played badly. Because 40 hours is only 1,200 hands, and 1,200 hands is noise.

At $2/$5 the same profile scales up: roughly $40/hour expectation, $450/hour standard deviation. The shape doesn't change — the zeroes do. This is why the buy-in rules exist at all, and why a deeper look at what variance actually is pays for itself the first time you lose $900 in a session where you played well.

Compute your own number in two minutes

The classic risk-of-ruin approximation, in plain terms: risk of ruin ≈ e raised to the power of (−2 × win rate × bankroll ÷ variance), with win rate and standard deviation measured in big blinds per 100 hands, and bankroll in big blinds.

Solving for a 1% risk of ruin gives a usable formula:

bankroll needed (in bb) ≈ std dev² × 4.6 ÷ (2 × win rate)

Worked example, live $1/$2. Say you measure 7.5bb/hour over 30 hands/hour — that's about 25bb per 100 hands — and your standard deviation runs around 150bb/100, typical for a full-ring live game. Then:

150 × 150 × 4.6 ÷ (2 × 25) ≈ 2,070bb ≈ 21 buy-ins.

That's where the "20–30 buy-ins" rule comes from: it's a 1%-ruin number for a *strong* winner in a *normal* live game. Now watch what happens when the inputs move:

• 25bb/100 winner, calm game (std dev 150): ~21 buy-ins
• 15bb/100 winner (std dev 150): ~35 buy-ins
• 10bb/100 winner (std dev 150): ~52 buy-ins
• 5bb/100 winner (std dev 150): ~104 buy-ins
• 25bb/100 winner, wild straddled game (std dev 200): ~37 buy-ins

The table explains most forum arguments. The player saying "$4,000 is plenty for $1/$2" is a 25bb/100 winner. The player saying "I went broke with $8,000" was winning at 8bb/100 and didn't know it. Both are telling the truth about their own game.

One more honesty adjustment: you don't know your true win rate under 500 hours, and your measured rate is probably optimistic because we remember our best stretches. Use half your measured win rate in the formula until you have 1,000+ hours logged. A bankroll built on an inflated win rate is a sandcastle.

When 20 buy-ins is not enough

The 20-buy-in rule fails you when any of these is true:

• Your edge is thin. Below ~15bb/100, the formula demands 35–100+ buy-ins. Most live players are thinner than they think after rake.
• The game is deep or straddled. Bigger stacks and straddle pots inflate standard deviation, and ruin risk grows with the square of it.
• The rake is brutal. Some rooms effectively cap your win rate. A 10% rake to $7 with a bad-beat jackpot drop can turn a 20bb/100 game into a 12bb/100 game.
• You can't reload. If the roll is your entire poker future and there's no salary behind it, take the conservative column and add a third.
• It's your income. Paying rent from the roll means withdrawing during downswings — exactly when the math needs you to stay whole. Pros who live off the roll typically hold 50+ buy-ins *plus* six months of expenses outside it.

The replenishability question

There's one legitimate shortcut, and it's worth stating plainly: if you have a stable job and can genuinely reload $500 a month without pain, then your real bankroll is your income, and $2,000 is a perfectly rational $1/$2 hobby roll. Just be honest that this is what you're doing. The danger isn't the small roll — it's believing you have a $6,000 bankroll when you have $2,000 and a credit card.

Whatever your number, it only works if you know where you stand after every session. Log stake, hours, buy-in and cash-out after every game, and review monthly; the main bankroll guide covers the tracking discipline, and a dedicated poker bankroll tracker turns the formula's inputs — your real win rate and real variance — into measured numbers instead of guesses. When the roll finally covers 30 buy-ins of the next stake, the shot-taking rules tell you how to test it without risking the climb.

FAQ

Is $2,000 enough to play $1/$2? It's 10 buy-ins. For a recreational player with a job who can reload, it's a rational hobby budget. As a professional-style bankroll, no — the math says even a strong winner faces real ruin risk at 10 buy-ins.

How many hours before I can trust my live win rate? Treat anything under 500 hours as a rough hint, anything under 1,000 as an estimate. Until then, plug half your measured win rate into the risk-of-ruin formula and size the roll for the pessimistic case.

Does the buy-in cap at my casino change the requirement? Yes. A 100bb cap lowers variance versus deep or uncapped games, so capped games need fewer buy-ins for the same ruin risk. If your regular game plays 200bb+ deep or straddles every orbit, move up a column in the table.

Should my bankroll include money I could reload from salary? Keep two numbers in your head: the roll (what you can lose without touching your life) and your access to money (the roll plus what you could replenish). Make decisions from the first number; sleep at night knowing the second.

Can I run one bankroll across live and online? You can, but track the two separately because the variance profiles differ — online needs more buy-ins for the same stake because the pools are tougher and the pace is faster. One roll, two ledgers.