Bankroll management is the least exciting part of sports betting and the part that decides most outcomes. Two people can back exactly the same selections and finish in completely different places, purely because of how much each one staked and when they changed it.
This article covers how to set a bankroll, the two staking methods worth using, why normal losing runs are longer than most people expect, and the arithmetic that makes doubling systems fail. It also covers the honest limit of all of this: staking discipline controls how you lose, not whether you lose.
What a bankroll actually is
A bankroll is a fixed sum of money, decided in advance, that is set aside for betting and nothing else. The definition contains three requirements, and dropping any one of them removes the point of having a bankroll at all.
- It is decided before you place a bet, not adjusted while you are betting.
- It is money you can lose entirely without affecting anything you need.
- It is separate from the account you pay rent, bills, food and debts from.
The practical test is simple. If losing the whole amount would change a single decision you make next month, the amount is too large. A bankroll is best thought of as the cost of the entertainment, in the same way as a concert ticket, rather than as capital that will be recovered.
Unit sizing: turning a bankroll into stakes
A unit is one standard stake, expressed as a percentage of the bankroll. Most guidance settles between 1% and 3% per bet, and there is a reason it is that low.
| Unit size | Stake on a 1,000 bankroll | Consecutive losses to zero |
|---|---|---|
| 1% | 10 | 100 |
| 2% | 20 | 50 |
| 3% | 30 | 33 |
| 5% | 50 | 20 |
| 10% | 100 | 10 |
| 25% | 250 | 4 |
At 10% per bet, a losing run of ten ends the bankroll. As the next section shows, a run of ten is unusual but nowhere near impossible, which means a 10% unit is not a strategy so much as a short countdown.
Losing runs are longer than you expect
This is the part most people underestimate, and it is pure arithmetic rather than luck or judgement.
| Consecutive losses | Probability of that exact run | Reads as |
|---|---|---|
| 3 | 12.5% | One time in eight |
| 4 | 6.3% | One time in sixteen |
| 5 | 3.1% | One time in thirty-two |
| 6 | 1.6% | One time in sixty-four |
| 7 | 0.8% | One time in one hundred and twenty-eight |
| 10 | 0.1% | One time in one thousand and twenty-four |
Those numbers look reassuring in isolation and are not. Across two hundred bets there are nearly two hundred places where a run of seven could begin, so seeing one over a season is unremarkable rather than unlucky. A staking plan that cannot survive seven consecutive losses is not a staking plan.
The same logic applies in the other direction: a run of eight wins does not mean anything has been solved. Both tails of the distribution are normal, and treating a hot streak as evidence of skill is how unit sizes creep upward.
Flat staking and percentage staking
There are two methods worth using, and the difference between them is whether the stake responds to the bankroll's current size.
- Flat staking
- Every bet is the same fixed amount, calculated once from the starting bankroll. Simple, transparent, and easy to hold to. The drawback is that as the bankroll shrinks, a fixed stake becomes a larger share of what remains.
- Percentage staking
- Every bet is the same percentage of the current bankroll, recalculated as it changes. Stakes shrink automatically during a losing run, which makes the bankroll much harder to exhaust. The drawback is the recalculation and the smaller stakes when recovering.
| Bet | Flat stake | Flat balance | Percentage stake | Percentage balance |
|---|---|---|---|---|
| 1 | 20 | 980 | 20.00 | 980.00 |
| 2 | 20 | 960 | 19.60 | 960.40 |
| 3 | 20 | 940 | 19.21 | 941.19 |
| 4 | 20 | 920 | 18.82 | 922.37 |
| 5 | 20 | 900 | 18.45 | 903.92 |
Over five bets the difference is small. Over two hundred it is substantial, and it is the reason percentage staking is more resilient. For most people betting recreationally, flat staking is easier to follow honestly, and following a simple plan beats abandoning a sophisticated one.
Some bettors vary unit size by confidence, using one to three units per bet. This is defensible if the scale is defined in advance and used sparingly. It becomes a problem when 'confident' starts meaning 'losing and want it back'.
Why doubling systems fail
The Martingale system says: double your stake after every loss, so that the eventual win recovers everything plus one unit. It is the most persistent bad idea in betting, and its failure is arithmetic rather than bad luck.
| Loss number | Next stake | Cumulative staked | Target profit |
|---|---|---|---|
| 1 | 20 | 30 | 10 |
| 2 | 40 | 70 | 10 |
| 3 | 80 | 150 | 10 |
| 4 | 160 | 310 | 10 |
| 5 | 320 | 630 | 10 |
| 6 | 640 | 1,270 | 10 |
| 7 | 1,280 | 2,550 | 10 |
After seven losses you are risking 1,280 to win 10, and you have already staked 2,550. The target profit never changes while the exposure doubles every step. Three things then finish the system off: bankrolls are finite, sportsbooks apply maximum stake limits, and the underlying probability of each bet is unchanged by what happened before it.
That last point is the fundamental one. A losing run does not make the next bet more likely to win. Every variation on the theme, including reverse Martingale, Fibonacci and d'Alembert, shares the same flaw: they redistribute when you lose without changing how much you lose on average.
No staking plan changes the odds. It only changes how quickly you meet them.
P77SPORT editorial
Keeping records that tell you the truth
Memory is unreliable about betting in a specific, consistent direction: wins are recalled vividly and losses are rounded down. A simple record removes the argument.
- Date, event and market.
- The price taken and the stake.
- The result and the net outcome.
- A one-line reason for the bet.
That last column is the valuable one. After fifty bets it will show you which reasons keep appearing in your losses, and quite often the answer is 'I had already lost earlier that day'. A spreadsheet is sufficient. Running totals matter more than individual results.
Bankroll and promotions
Promotions interact with staking in a way that is easy to miss. A bonus with a turnover requirement obliges you to stake a multiple of the amount before any withdrawal, and that obligation can push you into betting more, faster, and at minimum odds you would not otherwise choose.
Before accepting an offer, work out the total turnover required and ask whether you would place that volume of bets anyway. If not, the promotion is changing your staking behaviour rather than adding value to it. How sports betting promotions work covers the conditions in detail, and our promotions page explains the terms to look for.
Knowing when to stop
A bankroll defines a financial limit. It does not define a time limit or an emotional one, and both matter as much.
The honest summary of everything above: sound bankroll management makes losing slower and more predictable. It does not make betting profitable, because the margin built into every price means a typical customer loses over time. If betting has stopped feeling like entertainment, or if the limits above are becoming difficult to hold, our responsible gaming resource covers the warning signs and where to look for support.
GOOD TO KNOW
Frequently asked questions
Common guidance is 1% to 3% of the bankroll per bet. At 3%, a bankroll survives thirty-three consecutive losses; at 10% it survives ten, which is well within the range of a normal losing run.
Percentage staking is more resilient because stakes shrink automatically during a losing run. Flat staking is easier to follow consistently. For recreational betting, the plan you will actually stick to is the better one.
No. Doubling after a loss requires exponentially larger stakes to win a fixed amount, and bankrolls and sportsbook stake limits are both finite. It also assumes a losing run makes the next bet more likely to win, which is untrue.
Percentage staking increases stakes naturally as a bankroll grows, which is fine. Deliberately raising unit size because of a winning streak is not, since a streak carries no information about the next result.
No. Staking discipline controls how fast a bankroll depletes and protects against ruin from a single bad run. It cannot overcome the margin built into every price, which makes betting a negative-expectation activity over time.
