Martingale in Forex: The Risks Explained

Why does martingale keep coming back?

The martingale idea is old and simple. After a loss, increase the next bet so that a single win recovers everything lost before, plus the original profit. Each cycle ends with the same small gain, and the method appears to be a sure way to win. Many people who try it see exactly that pattern for a long time, and that long run of small wins is the reason the idea keeps returning in new forms.

The pattern is also the trap. A method that wins most of its trades looks reliable on a chart, and its rare large losses are easy to overlook in a short history. The martingale does not fail often. When it fails, it fails completely, because the size it has built up to recover the losses is larger than the account can support. The losses that matter are the ones that have not yet happened.

This guide explains what a martingale strategy does, why the doubling destroys accounts, how grid expert advisors hide the same risk behind a clean equity curve, and how to stress test any such approach before it is used. It treats martingale as a high-risk method, and it does not present any version of it as a safe way to trade.

What exactly does a martingale strategy do?

A martingale strategy sets a base position and increases it after every losing trade. The usual multiplier is two, so the position doubles each time a trade loses. When a trade wins, the position returns to the base size, and the cycle starts again. The win covers the total of the previous losses plus the base profit.

The arithmetic is clear. Suppose the base position risks 10 money and the multiplier is two. After one loss, the next position risks 20. After two losses, 40. After three, 80. After four, 160. The total risked across four losses is 310 money, and the fifth trade must win 160 to bring the cycle back to a small profit. The method works only as long as the fifth trade wins, and the size grows with every loss.

Some martingale variants use a smaller multiplier, or they add to the position in fixed steps rather than doubling. The principle is the same: the exposure increases after losses, so the recovery depends on a win that becomes larger with each step. Any strategy that increases its size into a loss is using a martingale principle, whatever its name.

Why does the doubling destroy accounts?

The doubling is exponential, and exponential growth is what makes the method dangerous. A run of ten losses at a multiplier of two requires a position more than a thousand times the base size on the tenth trade. The total money risked across the run is more than the base position multiplied by one thousand and twenty-three. A trader who begins with a modest base can reach a position far larger than any account they hold.

Brokers impose margin requirements that stop this growth long before the theory reaches its limit. When the position grows too large for the account, the broker closes it, often at the worst price, and the loss is locked in. The method then fails in practice at a much shorter run than the theory suggests, and the trader is left with a large loss and no recovery. Margin is not a protection against martingale; it is the point at which the method is closed out.

The combination of exponential exposure and a finite account means that martingale is not a method with a small chance of ruin. It is a method whose ruin is certain over a long enough period, because the run that ends it will eventually arrive. The only question is when, and whether the account is still there when it does.

How does a high win rate hide the risk?

A martingale method that wins nine trades in ten, each for a small amount, can show a steadily rising equity curve for months. The chart looks clean, the drawdowns look small and the strategy appears to be working. The risk is in the one trade in ten that loses, and specifically in the run of losses that the method has not yet met.

A short history cannot show the worst run, because the sample may not contain one. A backtest of one year may show no run longer than four losses, and the trader concludes the method is safe. The test is misleading because it did not look far enough. The worst run is a property of the process, and a sample that does not contain it has not tested the risk.

The lesson is that a high win rate and a small average loss are not evidence of safety when the method increases its size into losses. They are evidence that the method has so far avoided its worst case, which is a different thing.

How do grid EAs hide the same risk?

A grid expert advisor places orders at fixed price intervals, usually adding positions as the price moves against the first one, and closing the whole grid when the price returns to a target. The method is a martingale in practice when it adds exposure into a loss, because each new position increases the total risk while the price is moving in the wrong direction.

Grid EAs often look excellent in a short test. The price tends to oscillate around a range, the grid closes repeatedly at a small profit, and the equity curve climbs. The drawdown appears only when the price moves in one direction for a long period, and the grid keeps adding positions without closing any of them. A grid’s maximum drawdown can be far deeper than its history suggests, because the history may not contain a strong trend.

The drawdown tools guide explains how to measure drawdown and how to stress test it. For a grid, the stress test must assume a trend long enough to fill the grid to its limit, and it must check whether the account survives that trend.

What does a stop-loss do to a martingale?

A stop-loss on a martingale position limits the loss on that position, but it does not limit the loss on the whole sequence. If the sequence has grown to several large positions, a stop placed on the last one protects only that one. The combined loss across the sequence can still be very large, and it can exceed the planned risk by a wide margin.

Gaps make this worse. A large position that gaps through its stop, perhaps at the weekly open or around news, can lose several times its planned loss in a single step. The stop then has done nothing except confirm the loss. Many martingale EAs have no stop at all, relying on the recovery to come, and those methods carry the full risk of the run.

When assessing any martingale or grid approach, ask two questions. What is the maximum combined exposure the method can reach? What loss would that exposure produce if the price gapped past the stop? The answers, not the stop itself, determine whether the approach is survivable.

How should a martingale or grid EA be stress tested?

Start with the longest losing run in the sample, measured in trades and in money. Then assume a longer run, perhaps twice as long, and calculate the exposure and the loss at its end. Check the account balance, the margin level and the stop-out level in the broker’s contract at that point. If the account would be closed out, the approach cannot survive the scenario.

Repeat the test with a gap. Assume that the price moves a large distance past the last stop in a single step, and calculate the loss. The gap size should be based on the largest move observed in the instrument’s history, not on the planned stop distance. The grid drawdown stress test included with this guide sets out these scenarios in printable form.

A martingale approach that passes the stress test only by assuming the run never reaches the scenario has not passed. It has merely avoided the test. The stress test is valuable precisely because it asks what happens when the rare event occurs.

What are the common mistakes with martingale?

The first mistake is to test a martingale on a short sample and to read the result as proof of safety. The second is to assume that the broker’s margin will not close the account. The third is to increase the base size after a winning run, which raises the exposure of every later step.

The fourth mistake is to trade a grid or martingale on an account that is too small for its worst run. The fifth is to rely on a stop that cannot protect the combined position. The sixth is to add funds to an account that is losing on a martingale, which increases the amount that can be lost rather than recovering it. Each of these mistakes follows from the same belief, that the recovery will come.

How do you recognise a martingale in a product description?

Many products do not use the word martingale, but describe the same behaviour in other terms. A product that promises to “recover” losses, to “average down”, to “add on dips” or to “close the grid when the market returns” is describing a martingale or grid method. A product that describes its position size rising after losses, without saying so, is doing the same.

Read the logic, not the marketing. Ask what the advisor does after a losing trade. If the answer is that it opens a larger or additional position, the method is a martingale in its risk, whatever it is called. Ask also what the maximum exposure is, and whether it has a limit. An answer of “no limit” is a clear warning.

What does a worked martingale sequence look like?

Imagine an EA with a base position risking 10 money per trade and a multiplier of two. The trader’s account is 2,000, and the EA has won 47 trades in a row, each for a modest amount, since the account opened. The equity curve is smooth and the drawdown has never exceeded 40 money. The trader sees no reason for concern.

Then a run of losses begins. The first loss is 10, the second 20, the third 40, the fourth 80, the fifth 160, the sixth 320 and the seventh 640. The total of those seven losses is 1,270 money, which is 63% of the account. The eighth position would risk 1,280, which is more than the entire remaining balance. At that point the broker’s margin rules close the position, usually at a worse price than the one planned, and the account is left with its last funds.

The example shows what the smooth equity curve concealed. Forty-seven wins had produced a small, steady gain, and the method was tested against none of the seven losses that ended it. A backtest of the same EA over a year that contained no run longer than four losses would have shown the same clean curve. The example also shows that the fatal step was not the first loss or the second, but the step at which the size had outgrown the account, which the trader could only have seen by calculating the sequence in advance.

How much does a losing run cost in lots?

The cost of a losing run in lots is easy to underestimate, because the lot size is small at the start. At a base of 0.01 lots and a multiplier of two, the sequence reaches 0.64 lots after seven losses, and 1.28 lots after eight. A trader who sees the first two positions as trivial may not notice that the fifth is already the size of a meaningful position on a small account.

Convert each step to money before starting, using the pip value and the stop distance. Write the table in the journal with the exposure at each step, and compare the last step with the account balance. If the last step is larger than the account, the method cannot survive the run the table describes, however rare that run may be.

Why do recovery targets make the risk worse?

A recovery target is the price move that returns the whole sequence to break-even. The further the sequence has grown, the larger the move needed to reach that target, and the longer the exposure must be held. A long hold brings a risk of overnight financing, news and gaps, all of which add to the loss.

The target also encourages the trader to hold through adverse moves in the belief that the target will be reached. Each additional step increases the exposure and moves the target further away, so that the position becomes harder to exit the longer it is held. A martingale with a fixed target can therefore be trapped in a losing position for a period the sample never contained.

What does a responsible decision on a martingale look like?

A responsible decision begins with the worst case, not the average. Write the sequence table for the planned base size and multiplier, calculate the exposure at each step, and find the step at which the account would be closed out. If that step is reached by a run that could plausibly occur in the instrument’s history, the method is too risky at that size.

The decision then has three options. The base size can be reduced until the closeout step lies beyond any plausible run, which often means a size so small that the method is no longer worth running. The method can be abandoned in favour of an approach whose worst case is bounded. Or the account can be enlarged so that the closeout step is survivable, which means committing more money to a method whose failure mode is still complete loss. Each option should be written in the journal with the calculation behind it.

How do you explain a martingale to someone who is tempted by it?

The most effective explanation is a table, not an argument. Write the sequence for the base size, the multiplier and the account balance, and show the exposure at each step beside the balance. Most people who are tempted by a martingale accept the logic of recovery in principle, and then change their mind when they see the last few rows of the table, where the position is larger than their savings.

Pair the table with the fact that the table is built from the method’s own rules, not from a worst case invented to frighten. The step where the account is closed out is a direct consequence of the stated multiplier, and nothing about the market needs to change for it to arrive. That point is usually the one that persuades, because it removes the argument that the method is unlucky rather than flawed.

Why is a martingale harder to stop than to start?

A martingale is easy to start, because the first trades are small and the early results look encouraging. It is hard to stop, because the stop decision is made at the moment the account is most damaged and the trader most wants to recover. The sunk cost of the losses pushes toward one more step, and the belief that the next win will make everything right is strongest when the position is largest.

For that reason, the decision to stop should be made in advance and written down, with a specific trigger. A trigger such as a loss of a fixed percentage of the account, or a sequence of a fixed length, removes the need to decide under pressure. When the trigger is reached, the method stops, whatever the trader believes about the next trade.

What does a martingale do to the account’s mental state?

A martingale changes the trader’s relationship to the account. Each loss is followed by a larger position, which means each loss must be watched more closely, and the trader is under pressure for longer after every adverse move. That pressure affects sleep, attention and the quality of other decisions, including decisions outside the martingale itself.

Many traders who run such methods describe a sense of being trapped, of holding a position they would close if they could, and of waiting for a recovery that does not come. That state is a warning in its own right. A method that produces it is not a method that the trader can run calmly, and the record of that feeling belongs in the journal alongside the numbers.

Why do small accounts face the greatest danger?

A small account reaches its closeout step sooner than a large one, because the same sequence of losses consumes a larger share of the balance. A trader with a small account may therefore see the method fail within the first few losing runs, while a trader with a large account may see it fail only after a long period of apparent success. The small account is not safer; it is simply closer to the point of failure.

For a beginner, this is an important point. The temptation to use a martingale is often strongest when the account is small and the wish to grow it is urgent. That combination is the one in which the method is most likely to end the account entirely, and it is the situation a careful trader should avoid most.

How do you keep a martingale out of a diversified plan?

Some traders run a martingale on one account while running a sound plan on another, and they treat the two as separate. That separation is not a safeguard. Both accounts belong to the same person, the same attention and the same balance sheet, and a loss on one can change the decisions made on the other. A martingale that is carried alongside a careful plan can also undo the plan’s discipline, because the habit of doubling after a loss tends to spread.

The honest approach is to decide whether the trader will run a martingale at all, and to write that decision into the plan. If the decision is no, the plan should say so and name the rule that prevents it. If the decision is yes, the plan should give the maximum exposure and the trigger for stopping, and it should account for the total risk across all accounts. Review that decision after every month of trading, and record any moment when the rule was tempting to break, because those moments show whether the decision is holding, and a decision that is never tested against temptation is not really a decision at all, and the test should be written down in the same place as the rule it protects, so that the two are read together at every review and no one can later argue that the rule and the test were separate matters.

What does this guide not promise?

This guide does not promise that any martingale or grid approach will fail on any particular date. Some runs of such methods end in profit for a long time. The point is that the failure is certain over a long enough period, and its size is not bounded by the history a trader has seen.

The guide also does not recommend any martingale product, grid EA or recovery method. The methods described here are presented so that their risk can be seen and measured, and the advice is to avoid them with money that cannot be lost.

What is the offer in this guide?

The offer is the grid EA drawdown stress test, a printable worksheet for testing the worst case of an approach that adds to losing positions. It covers the settings audit, the failure scenarios and the margin and stop-out arithmetic. Use it before any martingale or grid approach is considered, and keep the completed copy with the record of the decision.

The worksheet does not make a martingale safe. It shows the size of the risk, so that the decision to use or avoid the method is made with the figures in front of you.

How do you start this week?

If you use a martingale or grid approach now, start by measuring its worst drawdown in money from the account history. Then calculate the exposure at the end of a run twice as long as the longest in the record, and check whether the account would be closed out. Record the result.

If you are considering one, do the same test before opening any account, using the worksheet. If the account would not survive a plausible run, do not use the approach at that size. For the arithmetic of leverage and margin that underlies this test, read the leverage guide, and for the statistics that help judge a sample, the Monte Carlo testing guide.

For more free tools and checklists, visit the resource library.