Blog
Sep 15, 2026·4 min read

What is bankroll management?

Deciding what a bet can cost you before you know how it ends. It's the part of betting that has nothing to do with picking, and the part that decides who's still here in March.

bankroll management

Short answer: bankroll management is deciding, in advance, how much of your money a single bet is allowed to lose.

It is the part of betting with nothing to do with picking winners, and it is the part that decides whether a genuine edge ever gets the chance to show up.

Start with the bankroll

A bankroll is money set aside for betting and not needed for anything else. Rent is not a bankroll. Next month's bills are not a bankroll. The test is whether losing all of it changes your life in any way beyond disappointment. If it does, the number is too big.

It should live somewhere you can see it as a total, and you should know what that total was when you started. Almost every bad outcome in betting begins with someone who cannot answer that question.

Fixed stakes

The core discipline is unglamorous: bet the same amount every time, regardless of how you feel about the game.

That amount is your unit. Bettors who last generally use somewhere around 1 to 2% of their bankroll per bet. Small enough that a bad run is survivable, large enough that a good one is worth having. A unit is one standard bet size, which is also why results here are reported in units rather than currency: the same record means the same thing whether your unit is $20 or $500.

What fixed staking prevents is the thing that actually empties accounts. Not a losing streak, an escalating one. The bettor who doubles up on the game they are sure about, then doubles again to get even, does not lose to variance. They lose to sizing.

Why the streaks are worse than you expect

Here is the number worth sitting with.

Suppose you are a genuinely good bettor winning 55% of your bets, better than most professionals manage. Across a season of around 300 bets, your longest run of consecutive losses will typically be about six, and runs of four or five will happen repeatedly.

That is not bad luck. That is the expected behaviour of a winning strategy. A 55% bettor loses 45% of the time, and losses arrive in clumps because randomness is lumpy rather than evenly spread.

So the question bankroll management actually answers is this: when that six-loss run arrives, and it will, does it cost you 6% of your bankroll or 60%? At one unit per bet, it is an unpleasant fortnight. At "whatever felt right," it can be the end.

The same arithmetic explains why a short record tells you almost nothing. Six weeks of results cannot distinguish a good bettor having a bad month from a bad bettor having a good one.

Three rules that do most of the work

Never chase. Increasing stakes to recover a loss converts a normal drawdown into a serious one. The bets that follow a loss are not more likely to win, and the maths does not know or care what you have just been through.

Size by the bet, not by the feeling. Confidence is a poor guide, and if your stake varies at all, it should vary with the edge and the price rather than with enthusiasm. That is the principle behind how our published unit sizes work: a bigger mispricing is not automatically a bigger bet, because a large edge at a poor price can be worth less than a small edge at a good one.

Keep it separate. A bankroll mixed into your current account is not a bankroll, it is a vague intention. Separation is what makes the total visible, and the total is what every other rule depends on.

What good bankroll management feels like

Boring, mostly. That is the point.

Stakes that do not change, a balance that moves slowly in both directions, and weeks where nothing much happens. The reward for doing it properly is not excitement. It is still being solvent when the edge you are relying on finally has enough sample behind it to show.

If any of that reads as an argument for betting more carefully rather than betting more, it is. And if you find the fixed stake genuinely hard to stick to, if the pull to raise it after a loss is strong, that is worth taking seriously as information rather than as a failure of willpower. Our responsible gambling page lists free, confidential help, and the honest version of a betting site says so somewhere other than the footer.


These are model outputs, not guarantees, and not financial advice. If betting has stopped being entertainment for you, our responsible gambling page lists free confidential help.

Published September 15, 2026 · For information only · 21+

Read the theory, then watch it run.

The model's highest-edge play lands in your inbox at 9 AM tomorrow. Free, every day.

Get today's free pick