Playing Too High: Bankroll and Game Selection
How to size a poker bankroll so normal variance can't break you, pick the right stakes for your roll, and choose tables where the money actually flows toward you.
Assumptions: Examples assume online cash games with 100 big blind buy-ins, at stakes matched to the bankroll under discussion; rake is ignored unless stated.
You can play flawless poker and still go broke. Not through bad luck in some cosmic sense — through a completely predictable failure: betting money you couldn't afford to lose on outcomes that were always going to swing. More beginners quit poker because they played too high than because they played badly. The fix costs nothing and requires no talent. It's a number, a rule, and the discipline to follow both.
What a bankroll actually is
Your bankroll is money set aside for poker and nothing else. Not your checking account, not "whatever's on the site right now," not money that has a rent payment's name on it. A real bankroll has two properties:
- Losing all of it would hurt no part of your life except your poker. If losing it changes what you eat this month, it isn't a bankroll — it's tuition you can't afford.
- It's sized so that normal downswings can't destroy it. This is the part with math in it.
The second property is the one beginners reject, because it implies something uncomfortable: even winning players lose money over alarmingly long stretches. A solid winner earning 5bb per 100 hands with a typical standard deviation of 80bb/100 still has about an 8% chance of being a loser after 50,000 hands — that's months of regular play. Downswings of 20 to 30 buy-ins happen to good players, routinely, forever. Your bankroll's job is to be deep enough that one of those ordinary downswings doesn't end your poker career.
The guideline: 30 to 50 buy-ins
For online cash games, the beginner rule is simple: keep 30 to 50 buy-ins for the stake you play, where a buy-in is 100 big blinds. Use 30 as the bare minimum for moving up and 50 as the comfortable cruising depth — and lean toward 50 while you're learning, because a still-leaky game makes downswings deeper and more frequent.
Concretely, a $500 bankroll supports:
- $0.01/$0.02 ($2 buy-in): 250 buy-ins — overkill, but a fine starting lab
- $0.02/$0.05 ($5 buy-in): 100 buy-ins — comfortable
- $0.05/$0.10 ($10 buy-in): 50 buy-ins — the highest stake this roll properly supports
- $0.50/$1 ($100 buy-in): 5 buy-ins — a coin toss away from broke
That last line is the trap. $0.50/$1 is where the "real money" feels like it starts, and five buy-ins feels like five chances. It isn't. Five buy-ins is one ordinary bad night.
Move up when you have 30+ buy-ins for the next stake and you're beating your current one. Move down — without drama, without shame — whenever a downswing leaves you under about 25 buy-ins for your level. Moving down is not failure; it's the system working. The players who go broke are almost never the ones who moved down too eagerly.
The math of ruin: five buy-ins at $1/$2
Meet a player every poker room knows. He has $1,000, he's read a book, and the $1/$2 game ($200 buy-in) is where the action is. Five buy-ins. He even might be a small winner in that lineup — call it 2bb/100 with a standard deviation of 90bb/100, which is realistic for a tight beginner.
Run his numbers. With a 500bb bankroll (his $1,000 at $2 blinds), a 2bb/100 winrate, and 90bb/100 variance, his risk of ruin is about 78%. Read that again: a winning player in this game, with this roll, goes broke roughly four times out of five — not because the strategy fails, but because the swings are far larger than the cushion. For that winrate-and-variance profile, a 5% risk of ruin requires about 6,100 big blinds (roughly 61 buy-ins, $12,200 at $1/$2), and 1% requires about 9,300bb. His $1,000 isn't a bankroll at this stake. It's a fuse.
The same player with the same $1,000 at $0.05/$0.10 holds 100 buy-ins and is, for practical purposes, unbreakable by variance. Properly rolled players at 5bb/100 with 40 buy-ins behind them carry a risk of ruin around 0.2%. The difference between 78% and 0.2% is not skill. It's the decision made before sitting down.
What being under-rolled does to your decisions
The ruin math actually understates the damage, because it assumes you keep playing well — and you won't. Watch the same hand at two stack-to-bankroll ratios:
Both heroes made the same correct play — QQ against AK is roughly 56/44 in your favor, and you take that bet all day. Both lost. The difference is what happens next. The $1/$2 hero is now staring at $800 where $1,000 used to be, and every decision for the rest of the session is contaminated: he passes on a profitable flip because "I can't afford another one," he chases a bad draw to "get back to even," he quits a good game early or stays in a bad one to recover the loss. Scared money plays scared, and scared play is losing play even when the underlying strategy is sound.
The $0.05/$0.10 hero feels a flicker of annoyance and moves on. That emotional flatness is not a personality trait — it's purchased, in advance, by correct bankroll sizing. When any single stack is around 2% of your roll, you can make every decision on its merits. When it's 20%, you can't, and knowing the math doesn't save you.
Game selection: the other half of playing the right game
Stakes are only one axis. At any given stake there are softer and tougher tables, and beginners should be ruthless about choosing — table selection is worth more than most strategy adjustments you could make this year. You're not being predatory; you're declining to volunteer as the table's worst player, which is what sitting at random guarantees often enough.
Online lobbies hand you the data:
- Players per flop (%). The single best softness indicator. High numbers mean players are calling preflop with junk — exactly the opponents whose mistakes pay you. As a rough compass at the micros: a table seeing flops 35%+ is promising; one in the low 20s is likely a grid of tight regulars.
- Average pot size. Big average pots at a loose table mean the money is moving — loose-passive callers paying off value bets. Big pots can also mean aggressive maniacs, which is fine too: as the tight, disciplined player from this track, you're positioned to profit from both.
- Stack sizes at the table. Several full or deep stacks of casual players is good. A table of identical 100bb stacks that never top up and play 19% of hands is a regular convention. Multiple short stacks sitting with 30-40bb often signals shove-bots and grinders — less profit, more variance, fewer postflop mistakes for you to collect.
The lineup you want is loose-passive: lots of callers, little raising, pots that grow because nobody folds rather than because anybody fights. Against that table, the entire strategy this track has taught you — tight starting hands, raise don't limp, bet your strong hands, price your draws — converts directly into money. The lineup to avoid is six tight regulars trading the same blinds back and forth; even if you're their equal, nobody at that table is making the mistakes that generate profit, and the rake quietly beats everyone.
And keep selecting after you sit. Tables change: the fish busts and leaves, two regs replace the limpers, the game tightens. The question "is this still a good table?" deserves an answer every 30 minutes, and "no" means standing up — mid-session, without ceremony. Leaving a bad game is the same skill as folding a bad hand.
Shot-taking: breaking the rules on purpose, safely
Eventually you'll want to try the next stake before the full 30 buy-ins are banked, and there's a disciplined way to do it. A proper shot has three parts agreed with yourself in advance: a budget (say, three buy-ins of the higher stake), a trigger (you're beating your current stake over a meaningful sample, not just running hot this week), and an exit (lose the budget, return to your old stake immediately, no second shot until the roll regrows). Written down, before you sit. The shot either wins and accelerates you, or loses a pre-paid, pre-sized amount that your roll absorbed by design.
What separates a shot from degeneracy is the exit actually firing. The player who budgets three buy-ins, loses them, and takes "just one more" hasn't taken a shot — he's begun a sequence whose ending you already know from the $1/$2 example. If you don't trust yourself to honor the exit, don't take shots at all; the 30-buy-in escalator gets you there anyway, just slower and with certainty instead of variance.
Two boundary notes while we're being precise about the roll itself. Money on the site is bankroll; money in your wallet earmarked for poker is bankroll; winnings you "mentally spent" are still bankroll until withdrawn — decide what you withdraw on a schedule, not after wins. And going the other way: topping up from your living money after a downswing is not a bankroll strategy, it's the absence of one. The roll is a closed system. That closure is what makes every number in this lesson mean something.
The rules, stated once
- Keep a poker bankroll that is only a poker bankroll.
- Hold 30-50 buy-ins for your stake; while learning, stay near 50.
- A $500 roll plays $0.01/$0.02 up to $0.05/$0.10. It does not play $0.50/$1, and it definitely does not "take a shot" at $1/$2.
- Move down at ~25 buy-ins without negotiating with yourself.
- Pick tables by players-per-flop and average pot, prefer loose-passive lineups, and re-evaluate every half hour.
- If a single buy-in loss would change how you play the next hand, you are playing too high — regardless of what the chart says.
None of this makes you a better card player. It makes you something more important first: a player who is still in the game next month, with a bankroll intact and a thousand more hands of experience, while the $1,000-at-$1/$2 guy is explaining on a forum how he ran bad. He didn't run bad. He ran normal, with no room for normal to happen.