Fixed-Percentage Staking: Bankroll Calculations and Risk

Fixed-percentage staking changes the cash stake as the bankroll changes. At a hypothetical rate of 1%, a bankroll of 10,000 produces a stake of 100, while a bankroll of 8,000 produces a stake of 80. The fraction stays constant; the amount does not.

This method describes exposure relative to a changing balance. It does not increase the probability that a prediction will win. The percentages below illustrate calculations rather than personal recommendations or universal safety limits.

What counts as the bankroll?

A bankroll is a separately defined sum used to account for betting activity. It is not the person’s entire bank balance. Money required for housing, food, debt or other essential expenses should not become a reserve for extending a losing sequence.

The examples here use sequential bets: the previous bet has settled before the next one is placed. Current bankroll therefore means initial funds plus realised net betting results. Deposits and withdrawals must be recorded separately so that newly deposited money is not mistaken for profit.

When bets overlap, available cash and total betting capital differ. Some funds are already committed to open positions. That situation requires a stated calculation base and a separate measure of total exposure.

The fixed-percentage staking formula

The basic formula is stake = current bankroll × selected fraction. A rate of 1% is written as 0.01, and 0.5% as 0.005.

With 10,000 and a hypothetical fraction of 1%, the initial stake is 100. A loss reduces the bankroll to 9,900 and the next stake to 99. If that bet wins at 1.80, its net profit is 79.20. The new bankroll is 9,979.20, and the next calculated stake is 99.792 before rounding.

The returned stake is not additional profit. Counting the entire payout as income would inflate the bankroll and all subsequent stake calculations.

For a losing bet, bankroll is multiplied by (1 − f), where f is the fraction. For a winning bet at decimal odds k, it is multiplied by [1 + f × (k − 1)]. These are accounting formulas, not forecasts of the next outcome.

What happens during a losing streak?

Recalculating each stake from a smaller bankroll reduces the amount of the next loss. At 1%, ten consecutive losses leave approximately 9,043.82 from a starting balance of 10,000. Twenty losses leave approximately 8,179.07.

After n consecutive losses, the theoretical balance is initial bankroll × (1 − f)ⁿ. The following comparison excludes intermediate rounding, fees, minimum stakes and deposits.

This table neither predicts a twenty-loss sequence nor defines the worst possible outcome. It illustrates how sensitive the balance is to the chosen fraction. A seemingly modest change in percentage can produce a substantial difference over a sequence.

With infinitely divisible money and a fraction below 100%, this mathematical model does not reach exactly zero after a finite number of full losses. Real betting has minimum stakes, rounding and levels of loss that are unacceptable long before zero. Describing the method as impossible to lose with is therefore misleading.

Fixed percentage versus a fixed flat stake

Flat staking keeps the cash amount unchanged during a defined period. Fixed-percentage staking follows the current balance. Starting at 10,000 with an initial stake of 100, twenty losses leave 8,000 using flat stakes or about 8,179.07 when recalculating 1%.

During a winning sequence, the percentage-based stake increases. This also increases the cash size of a later loss. The method changes the path of stake sizes; it does not eliminate uncertainty.

Consider one win at 2.00 followed by one loss, both at 1% of the current bankroll. The win takes 10,000 to 10,100. The next stake is 101, so its loss leaves 9,999. One win and one loss therefore produce a loss of one unit. The same outcomes at a flat stake of 100 would produce zero net profit.

For exactly the same set of odds and outcomes, the order of multiplicative factors does not change the final theoretical balance in this simplified model. However, interim drawdown, stopping rules, overlapping positions and rounding can change the practical experience and which bets are ultimately placed.

How to handle several open bets

“One percent per bet” does not mean a maximum daily exposure of 1%. Ten simultaneous stakes of 100 calculated from a 10,000 bankroll commit 1,000 in total.

If each next stake is instead calculated only from uncommitted cash, stakes shrink before any earlier result has settled. That is a different definition of the calculation base. Alternating between the two definitions without recording it makes the staking record inconsistent.

Both approaches require tracking total open exposure and exposure per match. A quarter over and a team over within that quarter may depend on the same possessions. A small fraction attached to each does not make them independent.

Expected winnings from an unfinished event should not be treated as available bankroll. A favourable current score is not the same as a settled winning bet.

Common mistakes with percentage-based staking

• Increasing the percentage after a loss. This adds a recovery progression to the original rule.
• Treating gross payout as profit. The returned stake must not be counted twice.
• Ignoring deposits. A larger balance may reflect new funding rather than successful selections.
• Ignoring correlated markets. Several positions can lose because of one match scenario.
• Assuming a small fraction proves the selection is good. Stake size does not improve the available odds.
• Choosing the fraction only from the best historical period. Future results may have a different distribution.

Comparing staking policies on new fixtures

Fix the starting balance, percentage, rounding rule, treatment of simultaneous bets and refund policy before testing. The same stream of selections can then be evaluated using a constant unit and a percentage-based stake.

Compare more than final balance. Maximum drawdown, turnover, number of bets, open exposure and dependence on a few large wins are all relevant. A method that finishes ahead in a short test can still have a much worse interim drawdown.

Prediction evaluation should remain separate: record the actual available line, odds, decision time, market and model version. A changed staking policy does not turn retrospectively selected winners into a valid forward test.

Fixed-percentage staking FAQ

What percentage is right for everyone?

There is no universal fraction. The numerical examples show consequences, not an individual’s acceptable loss limit.

Does the method improve win rate?

No. If selections, markets and prices remain unchanged, their outcomes remain unchanged. The stake amounts and financial path change.

Can the percentage be raised to recover losses faster?

That changes the policy and increases risk after an adverse result. The separate guide to Martingale and loss recovery explains the arithmetic of escalating stakes.

When should betting stop?

A precommitted spending limit remains a limit regardless of the formula. GambleAware recommends sticking to a budget and avoiding loss chasing. A smaller next calculated stake is not a reason that gambling must continue.