Sports Betting Bankroll Management — The Essential Guide 2026

Bankroll management decides how long your money lasts and how violently it swings on the way. It is the most overlooked part of betting and the only part you fully control.

The honest version first, because everything below depends on it: no staking plan can turn a negative edge into a positive one. Bookmaker margin gives the average bet a negative expected return before you place it, and a staking plan changes only the rate and shape of the outcome, never its sign. Anyone selling a staking system as a route to income is selling arithmetic that does not exist.

18+. Only bet money you can afford to lose entirely.

Define your bankroll

Your bankroll is a dedicated fund, separate from your current account and from money committed to anything else. There is no correct starting figure, and any guide quoting one is guessing at your circumstances. The only test that matters: would losing the whole amount change anything about your month? If so, the number is too high.

Never borrow to fund a bankroll, and never top one up from money earmarked for something else. If a bankroll reaches zero, that is the end of the exercise, not the moment to reload.

Unit size, flat staking and percentage staking

Flat betting means staking the same amount every time regardless of how confident you feel. That single constraint removes the two decisions people get wrong under pressure — staking up on a strong feeling, and staking up to recover — and confidence is a poor guide anyway, since your sense of certainty is not calibrated to actual probability.

The common range is a unit of 1-3% of bankroll — £10, £20 or £30 on £1,000. Unit size decides how long a bankroll survives an ordinary run of bad luck, not how much you win: at a 2% unit it takes fifty consecutive losing units to clear the bankroll, at 10% it takes ten. That is the whole trade.

There are two ways to hold a unit constant:

• Fixed flat staking sets the unit in currency once — £20 a bet — and leaves it there until you deliberately review it.
• Percentage staking recalculates the unit from the current bankroll each time, so 2% is £20 at £1,000 and £14 at £700. This is the safer of the two: stakes shrink automatically during a losing run, so a percentage-staked bankroll cannot be cleared the way a fixed-stake one can. The cost is slower recovery, because stakes only grow back as the bankroll does.

Neither is a strategy. Both are containers for one.

The Kelly Criterion

The Kelly Criterion comes from J. L. Kelly Jr's 1956 paper *A New Interpretation of Information Rate* (Bell System Technical Journal, vol. 35, pp. 917–926) — written at Bell Labs about information transmission, not gambling. It gives the stake that maximises a bankroll's long-run growth rate for a given edge.

Kelly % = (bp − q) / b, where b is decimal odds minus 1, p is your estimated win probability and q is 1 − p. At odds of 1.70 with an estimated 65% chance: b = 0.70, p = 0.65, q = 0.35, so Kelly = (0.455 − 0.35) / 0.70 = 15% of bankroll.

Almost nobody stakes that, for two reasons.

Kelly is optimal only if your probability is correct. The formula takes *p* as given. In betting *p* is your own estimate, so an overestimated edge makes Kelly systematically over-stake — its growth-maximising guarantee is a guarantee about a number you do not actually have.

Even with a correct *p*, full Kelly is brutally volatile. Maximising the growth rate says nothing about the path taken to get there, and that path includes very deep drawdowns.

Hence fractional Kelly — usually a half or a quarter of the Kelly number. The trade is favourable and can be shown directly: under the standard continuous approximation, staking fraction *f* of full Kelly delivers *f*(2 − *f*) of the maximum growth rate while scaling volatility by *f*. At half Kelly that is 0.5 × 1.5 = 0.75 — three quarters of the growth for half the volatility, plus a buffer against estimation error. Quarter-Kelly on the example above gives 3.75%, back in ordinary flat-staking territory. Not a coincidence.

A negative Kelly number means the bet is negative expected value and the correct stake is nothing. That arithmetic is in expected value betting explained and how to calculate value in a bet.

Variance, and what a normal losing run looks like

Most bankrolls are not lost to bad judgement. They are lost because nobody planned for ordinary variance.

Losing streaks are not anomalies to be explained — they are the arithmetic of independent events. At a 55% strike rate, five straight losses come up about 1.8% of the time (0.45 to the power of 5), which sounds small until you place 200 bets, at which point such runs are routine. Ten-bet runs happen too, and at a 2% unit that is a 20% drawdown produced by nothing but sequencing.

Two consequences worth absorbing in advance:

• A drawdown is not evidence your method is broken, and a winning run is not evidence it works. Both are far shorter than the sample needed to say anything.
• Recovery is asymmetric. Losing 20% needs a 25% gain to get level; 50% needs 100%; 75% needs 300%. Limiting the depth of a drawdown matters far more than chasing it back.

That asymmetry is why chasing is fatal rather than merely unwise: doubling after three losses turns an ordinary run into a bankroll-ending one in a handful of bets. Leave the unit exactly where it is — see the mistakes that cost bettors most.

Record keeping

Records are the only way to know what is actually happening rather than what you remember happening. Log the date, sport and market, selection and odds taken, stake, result, and — the field everyone skips — your reasoning at the time. That last one is the only defence against rewriting your own history once you know the result.

Patterns appear over a few hundred bets, but treat them as leads rather than conclusions: at that sample size the swings above are easily large enough to manufacture a pattern that is not there. Records will not make you profitable; they will tell you honestly whether you are, which is more than most bettors ever establish. Apply the same standard to anyone selling picks — see how to spot a trustworthy tipster.

What bankroll management cannot do

Every price carries the bookmaker's margin, which gives the average bet a negative expected return before your judgement is applied. Staking rules govern the *distribution* of outcomes — drawdown depth, how long the money lasts, whether one bet can end you. They do not touch the expectation itself.

Sizing well therefore makes an unprofitable habit last longer and hurt less. It does not make it profitable. Where genuine edges exist they come from pricing — see how to find value bets and the market-by-market breakdown in our Premier League markets guide — and they are rarer and smaller than the people selling them suggest.

When to stop

Bankroll management also means noticing when betting has stopped being entertainment. The warning signs: betting money meant for rent, bills or savings; borrowing to fund it; hiding the amount from people close to you; chasing losses with rising stakes; anxiety when you cannot bet; work or relationships slipping.

Every licensed sportsbook must offer deposit limits, loss limits, session reminders and self-exclusion, and in the UK a single GAMSTOP registration blocks every UK-licensed operator. Set deposit limits on day one, before you think you need them — they cost nothing and they are the one control that still works when your judgement is impaired.

If any of those signs apply to you, stop and get support. Tools and helplines are on our responsible gambling page, and BeGambleAware.org provides free, confidential help.