“One more trade and I’ll win it all back.” At that moment martingale seems almost like bookkeeping: you increase the amount, make a profit and start over. But in a binary option, the profit is usually smaller than the amount you risk losing. So even a successful doubling can leave the whole cycle in the red. Let’s break down the martingale method with numbers, and then show how to test its promises on intrade.bar without turning the urge to win back losses into a trading plan.
What martingale changes, and what it can’t do
Classic martingale increases the amount after every loss. Starting from 10, the sequence looks like this: 10, 20, 40, 80, 160. After a winning trade, the size goes back to the original. In a short example it all looks reassuring: you just need to wait for one correct forecast.
But the direction of the next price does not depend on the size of your trade. Doubling doesn’t explain why a level should hold, a pullback should end or a candlestick pattern should play out. It distributes money across the outcomes of a strategy you have already chosen. If the signal is weak, a larger amount makes it more expensive, not more accurate.
Other progressions may increase the size by half, add a fixed amount or calculate it to hit a target profit. They should all be judged the same way: how much money will be needed after several losses, what will be left in the account and what result the whole cycle will give. A nice name doesn’t exempt a sequence from arithmetic.
How to read the payout before you calculate
Take a notional payout of 85% of the trade amount. With a size of 10, a win gives 8.50 profit: the original 10 comes back to your balance along with the 8.50 earned. A loss means losing 10. Here and below, the percentage is constant and there are no extra costs or refunds. This is a teaching model, not a promise of the same payout on every trade.
On intrade.bar, the percentage is shown before you open a trade. That is the number to plug into the calculation: a figure from an old screenshot doesn’t describe the current contract. Be clear whether you are comparing net profit or the total amount returned. Plugging in 18.50 instead of 8.50 as the income of the last step is an easy way to get a profit that exists only in the spreadsheet.
If the opening and closing quotes are equal, intrade.bar refunds the trade amount. Such an outcome brings neither 85% nor a full loss. If you are studying a progression, decide in advance what happens after a refund: the size stays the same or the cycle ends. Without that, two calculations of the same scheme may quietly describe different rules.
Where ordinary doubling breaks down
After losing 10, the next trade of 20 will bring 17 profit if the forecast is right. The result of the two trades is plus 7. After two losses, 10 and 20, a win on 40 gives 34: a net plus 4. After three losses, 70 has already been spent. A win on 80 brings 68, and the cycle ends with a loss of 2.
| Number of the winning trade | Its amount | Losses before it | Profit at 85% | Cycle result |
|---|---|---|---|---|
| 1 | 10 | 0 | 8.50 | +8.50 |
| 2 | 20 | 10 | 17 | +7 |
| 3 | 40 | 30 | 34 | +4 |
| 4 | 80 | 70 | 68 | −2 |
| 5 | 160 | 150 | 136 | −14 |
| 6 | 320 | 310 | 272 | −38 |
You can verify the pattern with a formula. With a starting amount b and the first win at step n, the previous losses equal b × (2^(n−1) − 1). The profit of the winning step at 85% is 0.85 × b × 2^(n−1). The cycle result is b × (1 − 0.15 × 2^(n−1)). The longer the streak, the bigger the gap between a full loss and a partial profit.
The last line of the history will still be profitable. That is exactly why looking only at it isn’t enough. Green is a pleasant color, but it is no accountant. Add up the results of all trades from the start of the cycle and compare the balance before the first trade with the balance after the last one, accounting for deposits and withdrawals separately.
What if you increase the amount “correctly”?
To cover accumulated losses L with one win and still make a cycle profit G, with a payout rate r you need an amount of (L + G) / r. Starting from 10, with a target of 8.50 and a payout of 0.85, after the first loss you will need about 21.76. After the second, about 47.37, then about 103.10. This is a calculation of capital requirements, not a reason to fund them.
The next size grows roughly 2.176 times. A mathematically exact payout adjustment fixes the flaw of ordinary doubling, but makes risk grow faster. And if the percentage drops before the next trade, the same target will need even more money. A constant recovery target doesn’t mean constant conditions for reaching it.
There is also practical rounding. If only whole amounts are allowed, the calculated value has to change: round down and it won’t be enough for the target, round up and the risk increases. A maximum trade amount or a daily limit you have chosen can also stop the chain. The formula doesn’t guarantee that the trade you need can be opened at all.
Why a large balance doesn’t make the chain infinite
Say the balance is 1,000 notional units and the first amount is 10. Six losses with doubling will take 10 + 20 + 40 + 80 + 160 + 320 = 630. The remainder is 370, and the next amount under the rules is 640. There is nothing left to continue the original sequence with, even though the account isn’t empty yet.
A bigger balance with the same starting 10 gives extra steps. But each new step is noticeably more expensive than the last, and the finite limit doesn’t go away. If you scale up the starting size along with the balance, relative resilience stays the same. “I’ll just deposit more” changes the number of attempts available, not how the method works.
The chain also depends on time. After a loss, a suitable signal may not appear, the expiry you need may be unavailable or payout conditions may change. Opening any trade just because it’s the turn of the next amount means abandoning your original trading hypothesis. Then you are no longer testing a signal, but your own willingness to keep going.
A losing streak doesn’t have to end right now
For independent outcomes with a constant probability of loss q, the probability of a pre-chosen block of k losses is q^k. With a notional q = 0.5, six losses in one specific block give 1/64, or about 1.56%. Real trades need not be independent, and this example does not define the probability for your strategy.
Besides, one block and the whole history are different things. Over hundreds of trades there are many places where an unfavorable sequence can begin. Neighboring blocks overlap, so you can’t mechanically treat them as independent. But you also can’t declare a streak practically impossible based on the probability of one chosen block.
The thought “after five losses a win is due” sounds convincing only to someone tired of losing. In an independent model, past outcomes don’t change the next probability. In the market, dependence may exist, for example because a trend continues, but it has to be tested. The number of previous losses on its own is no substitute for that test.
Why 95% winning cycles still settle nothing
Imagine nineteen cycles of plus 8 and one with a loss of 300. That’s 95% winning cycles, yet the total result is 152 − 300 = −148. Progressions often create exactly this profile: small, frequent profits and rarer, larger losses. The percentage of winning cycles hides the difference in their money scale.
For comparison, take the same signals and calculate two versions: the same amount on every trade, and the progression. Keep all outcomes, the maximum capital requirement, the drawdown and the final result. Then you will see what the hypothesis itself delivered and what the position sizing changed. Don’t exclude an unfinished cycle just because it spoils a nice report.
In intrade.bar statistics you can see the number of winning, losing and refunded trades. To analyze a money chain, add a CSV export and your own table of amounts to those counters. WinRate alone won’t show that one big loss outweighed several small wins. Refunds are not included in the WinRate denominator.
Practice on intrade.bar: check the math and stop chasing losses
Start on a demo account with one hypothesis described in advance. Before each qualifying signal, record the percentage shown, the same practice amount and the chosen expiry. In a separate table you can calculate what would have happened with a progression on the same outcomes. That way, checking the arithmetic doesn’t require actually ramping up amounts, even on a practice account.
After several losses, ask yourself a simple question: would you open the next trade with the same amount if the previous losses hadn’t happened? If the answer is no, your decision is based on the balance, not the chart. Go back to the signal conditions; if they aren’t met, there is no new trade. Skipping here keeps the test meaningful.
intrade.bar has voluntary risk management. You can leave No limits or set your own daily loss and profit limits in advance. For example, with a notional balance of 1,000, a 2% loss limit sets a boundary of 20. This is only an example of how it works. If 10 has already been lost and there are no open trades, an attempt to place the next 20 will exceed the chosen boundary and the new trade will not open.
The check takes into account the amount of the new trade, the daily loss already incurred and the amounts in open trades. Once the set limit is reached on closed trades, new trades can’t be opened until the end of the day. The base for the limits is fixed when you set them; they can only be changed the next day, Moscow time, which starts at 21:00 UTC. No limits is not a way to lift a restriction you have already chosen on the same day.
At the end of the test, compare two things: the result of your hypothesis and your ability to stick to the original amount. If the idea only looks attractive after adding ever larger attempts, go back to testing it. On intrade.bar you can start this review on a demo, see the terms of each contract in advance and set your own hard daily limit before the urge to “win it all back in one trade” ever appears.
Questions and answers
Does martingale work with an 85% payout?
Doubling doesn’t ensure a constant cycle profit. Starting from 10, a win on the fourth step leaves a result of −2. Calculating the exact amount needed to recover requires risk to grow even faster and doesn’t remove the balance limit.
Can a daily limit stop doubling on intrade.bar?
Yes, if you have turned on voluntary risk management in advance: a new trade won’t open when its amount, together with the daily loss and open amounts, exceeds the set limit. The chosen limits stay in force until the end of the day and can only be changed the next day.
What should I look at instead of the percentage of winning cycles?
The money result of all trades, the maximum drawdown and the capital needed to keep the chain going. It helps to compare the same signals with a constant amount: that way the effect of the progression is visible separately from the quality of the signals.
Fanya biashara yako ya kwanza ya mazoezi
Akaunti ya demo hufunguliwa mara tu baada ya usajili
