An 80% binary options payout does not mean an 80% chance of winning, or an 80% return on your account. It describes the profit on a winning stake under a particular payout convention. If a losing trade costs the entire stake, winning half your trades still loses money.
The distinction matters because profitability depends on three separate figures: the amount gained when right, the amount lost when wrong, and the probability of each outcome. A large advertised payout answers only one of those questions.
For UK readers, the regulatory position comes first. The FCA’s permanent retail binary options ban took effect on 2 April 2019, prohibiting their sale, marketing and distribution to retail consumers by firms carrying out those activities in or from the UK. The examples below explain the mathematics, not a route around that restriction.
What a binary options payout actually means
Consider a hypothetical contract with a £100 stake and an advertised 80% profit payout. Assume the original stake is returned on a win and lost completely on a loss, with no fees, rebates or ties.
- Winning outcome: £180 comes back, consisting of the £100 stake plus £80 profit.
- Losing outcome: nothing comes back, producing a £100 loss.
- Maximum profit: £80, not £180.
This distinction between cash received and net profit is the starting point for every calculation. Counting the returned stake as earnings makes the result look much better than it is.
The word “payout” alone is not precise enough. Establish whether a quoted percentage means profit above the stake or the total settlement amount. If a contract costs £60 and pays £100 in total when successful, its winning profit is £40. The return on the £60 outlay is therefore 66.67%, not 100%.
The calculations here use the first format: a stake, a stated profit percentage on a win, and full loss of the stake otherwise. For the underlying settlement conditions, including strikes and expiry, see how binary options contracts work.
Calculating expected return
Expected return combines the possible gains and losses, weighted by their probabilities. It is a mathematical average across hypothetical repetitions under the stated assumptions, not a promise about the next trade.
Let S be the stake, r the winning profit rate expressed as a decimal, and p the probability of winning. With a full loss on an unsuccessful trade:
Expected net profit = S × [p × r − (1 − p)]
For a £100 stake, an 80% profit payout and a genuinely 50% winning probability:
Expected net profit = £100 × [(0.50 × 0.80) − 0.50] = −£10
No individual trade loses exactly £10 in this example. Each produces either £80 profit or a £100 loss. The minus £10 is the probability-weighted average.
A sequence containing exactly 50 wins and 50 losses at £100 per trade would generate £4,000 in winning profits and £5,000 in losses. The net result would be a £1,000 loss on £10,000 of cumulative stakes. That is a 10% loss relative to turnover, not necessarily a 10% loss relative to the starting account balance.
This asymmetric structure can create negative expected returns even with an even chance of winning. The SEC’s warning about overstated binary options returns identifies this same problem: the loss on an unsuccessful contract can exceed the profit on a successful one.
The break-even win rate
To find the winning probability needed to break even, set expected profit to zero and rearrange the formula:
Break-even win rate = 1 ÷ (1 + winning profit rate)
At an 80% payout, that becomes 1 ÷ 1.80, or approximately 55.56%. A positive expected return requires a probability above the exact threshold, before allowing for any extra costs.
| Profit payout on a win | Profit on a £100 winning stake | Break-even win rate | Expected result per £100 at 50% probability |
|---|---|---|---|
| 60% | £60 | 62.50% | −£20 |
| 70% | £70 | 58.82% | −£15 |
| 80% | £80 | 55.56% | −£10 |
| 90% | £90 | 52.63% | −£5 |
| 95% | £95 | 51.28% | −£2.50 |
| 100% | £100 | 50.00% | £0 |
These thresholds are calculated requirements, not evidence that a trader can achieve them. An 80% payout tells you what probability you would need. It does not tell you what probability you have.
Winning more often than losing is therefore insufficient. At an 80% payout, 55 wins and 45 losses on £100 stakes produce £4,400 of profits against £4,500 of losses: a £100 deficit despite the majority of predictions being correct.
Two outcomes do not mean equal probabilities
A yes-or-no question can have two possible answers without each being equally likely. Whether an asset finishes above a strike depends on where its price starts, the distance to that strike, the time remaining and the assumed distribution of future price movements.
For a simple illustration, compare two hypothetical contracts on an asset priced at £100. One asks whether it will finish above £100 in five minutes. The other asks whether it will finish above £110 over the same period. Their possible answers are both yes or no, but there is no mathematical reason to assign them identical probabilities.
Nor does a higher payout automatically make a contract better value. Using assumed probabilities, an 80% payout with a 60% winning chance produces an expected return of 8% of the stake. A 95% payout with a 50% winning chance produces an expected return of minus 2.5%.
The larger headline number loses that comparison. The difficulty is establishing a credible probability rather than choosing an attractive assumption.
Why a recorded win rate is not a known probability
Suppose a hypothetical record contains 60 wins from 100 trades. Its observed win rate is 60%, but that does not establish a true future winning probability of exactly 60%.
Using the NIST method for Wilson confidence intervals, the calculated 95% interval for that example is approximately 50.2% to 69.1%. This calculation assumes independent trials with a constant winning probability. The interval includes values below the 55.56% threshold required for an 80% payout.
The practical point is not that the strategy must lose. It is that 100 observations leave substantial uncertainty about whether its probability clears the required threshold.
Before using any track record as an estimate, ask whether it contains every qualifying trade, whether the rules were fixed before testing, and whether the recorded payouts match those available when each decision was made. A selection of winning screenshots cannot answer those questions.
Overlapping trades also need care. Ten positions exposed to the same price move should not automatically be treated as ten independent tests of a forecasting method. A narrow statistical interval built on the wrong assumptions offers false reassurance.
Claims about exceptional accuracy deserve separate scrutiny from the arithmetic. The guide to signals, account managers and guaranteed return claims covers that assessment without treating promotional results as verified evidence.
Changing payouts can reverse the result
A win rate cannot be evaluated separately from the payouts attached to the trades. Consider two hypothetical sets of 100 trades, each with £100 stakes and exactly 60 wins.
With an 80% winning payout, the result is £4,800 in profits minus £4,000 in losses: an £800 gain. With a 60% winning payout, the same winning record produces £3,600 minus £4,000: a £400 loss.
The prediction accuracy has not changed. The economics have.
Where payouts differ across trades, calculate each result using its actual terms. An unweighted average payout across all opportunities can mislead if larger stakes were placed on lower-paying contracts, or if successful trades received different payouts from unsuccessful ones.
For a completed record, the clean calculation is total winning profits minus total losing stakes and charges. For a forecast, calculate each trade’s expected result using its own stake, payout and estimated probability, then add those expected results. Do not replace missing information with the highest advertised payout.
Fees, refunds and ties change the calculation
The basic formula assumes just two net outcomes. Different settlement terms require different arithmetic.
Additional charges
If a fixed charge of c applies to every trade, expected net profit becomes:
S × [p × r − (1 − p)] − c
For a hypothetical £100 stake, an 80% profit payout and a £2 charge on every trade, the break-even probability rises to 1.02 ÷ 1.80, or approximately 56.67%. Charges applying only to particular outcomes must instead be assigned to those outcomes.
Partial refunds on losses
Suppose a losing £100 trade returns £10 in withdrawable cash. The net loss is then £90, not £100. With an £80 winning profit, the break-even probability becomes £90 ÷ (£80 + £90), or approximately 52.94%.
This calculation assumes a genuine cash refund. A promotional credit with conditions should not be inserted into the formula as though it were freely available money.
Expiry at the strike
If an exact tie returns the stake, it creates a third outcome with zero net profit before costs. Expected profit is then the winning probability multiplied by winning profit, minus the losing probability multiplied by the loss. The tie probability occupies the remaining share.
Count ties separately and state whether a reported win rate includes them. Otherwise, two apparently identical accuracy figures may measure different things.
Expected profit does not remove losing streaks
Even a genuinely positive expected return does not guarantee a profitable session or account survival. The sequence of results and the size of each stake still matter.
Assume, purely for illustration, independent trades with a constant 60% winning probability. The probability that the next five trades all lose is 0.40 raised to the fifth power: 1.024%. That is the probability for one specified block of five, not the probability of encountering such a run somewhere in a much longer record.
Increasing stakes after losses does not improve the underlying payout or winning probability. At an 80% payout, losing £10 and then winning with a doubled £20 stake produces £16 profit on the second trade, leaving just £6 after the earlier loss.
After losses of £10, £20 and £40, a win on an £80 stake earns £64 against £70 already lost. The sequence remains £6 down. Doubling has not even recovered the losses.
Recovering every previous loss plus a target profit would require still larger stakes, increasing the amount exposed when another loss arrives. The relationship between expected value, account size and survival is covered in trading expectancy and risk of ruin. A staking pattern cannot turn an unchanged negative expectation into a positive one.
A profitable calculation is not proof of a payable balance
Every example above assumes the contract is settled honestly and the money shown as available can actually be withdrawn. Those assumptions are separate from forecasting accuracy.
The FCA’s binary options scam warning describes firms faking prices and payouts, closing accounts and refusing to return money. In those circumstances, an account statement showing profits is not evidence that the customer has received a return.
Keep two questions separate: would the stated contract have positive expected value under defensible assumptions, and are the stated terms and payment obligations reliable? A spreadsheet can answer the first only as well as its inputs allow. It cannot establish the second.
How to assess a payout claim
Reduce any claim to four figures: the cash committed, the net profit on a win, the net loss on a loss, and an evidence-based estimate of the winning probability. Then account for charges, refunds and settlement rules.
At an 80% profit payout with full stake loss, the central hurdle remains a winning probability above 55.56% before costs. Neither a large advertised return nor a short winning record establishes that advantage.
For UK retail consumers, this analysis does not change the regulatory restriction described at the start. The UK retail binary options ban guide addresses its scope. Payout mathematics is useful for checking claims; it is not evidence that an offer is lawful, profitable or safe.