Bankroll and Risk

WeBall3 min readUpdated Aug 15, 2026Reviewed Aug 15, 2026

A bankroll is money ring-fenced for betting that you could lose entirely without it changing anything. Stake a small fixed percentage of it, expect long losing runs even with a real edge, and judge yourself on closing line value rather than profit until you have several hundred bets.

This is the least glamorous part of betting and the part that decides whether anyone survives long enough for their edge to matter.

The bankroll

A bankroll is money you have already decided you can lose in full, kept separate from money you need for anything else. The separation is what makes every other discipline possible — unit sizing, drawdown tolerance and honest record-keeping all require a fixed denominator.

The right size is not a formula. It is whatever amount you would be genuinely untroubled to lose entirely, because a losing run long enough to threaten it is not unusual — it is expected.

Units and stake sizing

A unit is a standard bet size expressed as a percentage of bankroll, conventionally 1%. Using units rather than dollars lets results be compared across bankrolls and stops a good week from quietly inflating your stakes.

The critical property is that it stays constant. A "unit" that grows after wins and shrinks after losses is not a staking plan, it is a mood.

Three approaches, in order of how much they demand of you:

Flat betting — the same stake on everything. Impossible to get wrong, makes your results interpretable, and the correct default for anyone whose probability estimates are not yet proven.

Percentage staking — a fixed percentage of the current bankroll, so stakes scale naturally with results.

Fractional Kelly — stake proportional to edge, at a quarter or half of what the formula says. Theoretically superior and genuinely worse than flat betting if your probability estimates are overconfident, which most people's are.

Variance is larger than you think

52.4%

Breakeven at −110

The bar, not 50%

~55%

A very strong win rate

Professional territory

1,000+

Bets to prove it

Before profit is distinguishable from luck

A bettor with a genuine 3% edge will have losing weeks, losing months and occasionally a losing season. That is not the edge disappearing; it is the ordinary behaviour of a small advantage applied to a noisy process.

The consequence is that short-run results are not evidence in either direction. Judge the decision instead: closing line value tells you whether your prices were good after a few hundred bets rather than a few thousand.

Risk of ruin

Even a profitable bettor can go broke if stakes are large relative to bankroll. Doubling your unit size does not double your growth — it multiplies your chance of never reaching the long run in which your edge pays.

This is why conservative staking is arithmetic rather than timidity, and why every serious staking method includes a hard cap.

When it stops being about money

If stakes are rising to chase losses, or betting has started to feel compulsory rather than chosen, no amount of bankroll discipline fixes that. The responsible gambling section covers what to look for, and free confidential help is available at 1-800-GAMBLER.

Every guide in this section

Flat BettingWhy staking the same amount on every bet beats clever staking for most bettors, what it gives up, and when you have earned the right to move on.Start here3 min

Common questions

How big should my bankroll be?

An amount whose total loss would change nothing about your life. Not an amount you expect to lose — an amount you could lose entirely without consequence. If no such amount exists right now, that is a clear answer about whether to be betting.

How many bets before I know if I'm any good?

For profit, well over a thousand. For closing line value, a few hundred. That gap is the entire reason to track CLV instead of your record.