Strategies and signals
Flat Staking in Sports Betting: Profit, ROI and Drawdown
Flat staking means using the same monetary amount for every qualifying selection during a defined period. If one unit is 100, the next stake remains 100 whether the previous bet won or lost. It is a useful way to examine prediction quality without large changes in stake size dominating the results.
Consistent stakes can still produce a loss. The outcome depends on the selections, the available odds and the accuracy of any probability estimates. This guide explains flat staking, break-even win rates, drawdown and the records needed for a meaningful comparison.
What exactly does flat staking mean?
Here, flat staking means a fixed cash amount throughout a stated evaluation period. Calling this amount one unit makes it possible to compare results without relying on a particular currency.
With a hypothetical starting bankroll of 10,000, a stake of 100 is 1% of the initial bankroll. This is an arithmetic example, not a recommendation that 1% is suitable for everyone. Acceptable exposure depends on circumstances, the amount a person can afford to lose and how many bets are open together.
If the stake remains 100 as the balance changes, it is flat staking. If it is recalculated as 1% of the new balance before every bet, it is fixed-percentage staking. These names are sometimes used loosely, so an audit should state the actual calculation.
Why equal stakes help evaluate a betting model
With a recovery system, a small number of large bets may determine the final result. Equal initial stakes make it easier to examine whether the selection policy itself is useful, although different odds still produce different potential profits.
This is relevant when evaluating basketball signals. If the forecasting model, selection criteria and staking plan all change at once, the cause of an improvement is unclear. Reporting a constant-unit benchmark helps separate prediction changes from money allocation.
Equal stakes do not remove correlation. A match over and two team overs can share the same game scenario. Three one-unit positions represent three units of exposure even when each individual amount appears small.
Calculating flat-stake profit and ROI
At decimal odds k, a winning bet produces net profit of stake × (k − 1). A losing bet loses its stake. A fully refunded bet produces zero betting profit, subject to the market’s settlement terms.
Consider 100 completed bets of 100 units each, with 60 wins and 40 losses. Assume that every selection has the same odds. Changing only that price produces the following results:
| Decimal odds | Net result | ROI on 10,000 turnover |
|---|---|---|
| 1.65 | −100 | −1% |
| 1.70 | +200 | +2% |
| 1.80 | +800 | +8% |
| 1.90 | +1,400 | +14% |
At odds of 1.70, the wins earn 60 × 70 = 4,200. The losses cost 40 × 100 = 4,000. Net profit is 200, and ROI is 200 divided by 10,000, or 2%, before any additional costs.
When prices differ, calculate the result for each selection. An overall average price and a win rate may be insufficient because the prices attached to the winning bets matter. Subscription costs and other charges should be reported separately when assessing the user’s total outcome.
The win rate needed to break even
For equal stakes at a constant price, with only full wins and full losses, the break-even win rate is 1 / decimal odds. This is a price threshold, not an estimate of how likely the event is to happen.
| Decimal odds | Break-even win rate |
|---|---|
| 1.60 | 62.50% |
| 1.70 | 58.82% |
| 1.80 | 55.56% |
| 1.90 | 52.63% |
| 2.00 | 50.00% |
A headline such as “70% winners” therefore leaves several questions unanswered. Which prices were available? How large was the sample? Were selections recorded before the outcomes? Could subscribers actually obtain the quoted lines?
How a flat stake behaves during a drawdown
Starting with 10,000 and using a flat stake of 100, twenty consecutive losses leave 8,000. The next stake is still 100, but that amount now represents 1.25% of the remaining bankroll rather than 1%.
A fixed cash amount does not imply a fixed percentage risk. If the unit is reviewed periodically, the review rule and the date of each change should be recorded. Quietly increasing it after a winning run or modifying it during a loss sequence makes comparisons less reliable.
Drawdown should also be measured from the previous bankroll peak. Falling from 10,000 to 8,000 is a 20% decline. Returning from 8,000 to 10,000 requires a 25% gain. Recovery percentages and loss percentages are not symmetrical.
Open positions and correlated markets
Ten simultaneous bets of 100 put 1,000 at risk. With a bankroll of 10,000, that is 10% of the initial capital, even though every individual selection is described as a 1% stake.
Within one basketball game, a first-half over, match over and favourite’s team over may all depend on a fast, efficient scoring environment. If that assumption is wrong, the positions can lose together. The number of market names does not measure diversification.
A useful record therefore includes total open exposure and exposure per match. The size of one flat unit does not, by itself, determine how many positions should be open at once.
What to include in a flat-staking journal
- Decision timestamp, fixture, market, selection and exact line.
- Available odds and the stake actually used.
- The rule or model version responsible for the signal.
- Settlement status and net result in cash and units.
- Bankroll after settlement and total outstanding exposure.
- Separate treatment of refunds, voids, partial settlements and cash-outs.
Win rate normally uses WIN and LOSS, with PUSH and VOID shown separately. ROI needs an explicit denominator. For example, a report may include settled PUSH stakes in turnover and exclude VOID stakes. Other conventions are possible, but they must be aligned before reports are compared.
A half-win, half-loss or cash-out should be represented by its actual financial result. Replacing it with a full WIN or LOSS can distort both accuracy and profitability figures.
Testing changes without fitting the past
Selection criteria should be fixed before evaluating subsequent fixtures. Every qualifying result must remain in the record, including losses. Retrospectively changing a line, selecting a different market or removing an unsuccessful period invalidates the comparison.
Prematch and live selections, totals and handicaps, and different game formats may behave differently. Small segments also produce unstable percentages: three wins are three observations, not evidence of a perfect system.
A filter can improve win rate while removing many profitable opportunities. Monitor the number of wins, unique matches, turnover and drawdown alongside the percentage. Equal stakes make this comparison easier to interpret; they do not guarantee that a better rule exists.
Flat staking FAQ
Does flat staking guarantee a profit?
No. It defines the amount staked. Negative-expectation selections remain negative-expectation selections at a fixed size.
Can the unit be increased after a winning streak?
That creates a change in the risk policy. It should be explicitly defined and recorded rather than justified only by recent success.
How is it different from Martingale?
Flat staking keeps the amount unchanged after a result. Martingale increases it after a loss, producing a different distribution of potential losses.
Should every available signal be followed?
A signal does not ensure that the same price is still available or that adding another position fits an existing exposure limit. GambleAware recommends setting a spending limit before gambling. An accounting method does not require using that limit in full.