Why the Martingale Strategy Fails in Forex

Written by Dominic Walsh · Published 12 Aug 2026 · Last updated 17 Aug 2026

The martingale strategy doubles the position size after every loss, on the theory that one winning trade recovers the whole sequence plus a small gain. Forex versions dress this up with indicators and filters. Underneath, the arithmetic stays the same.

So this guide does two things. It describes the martingale strategy accurately, then shows exactly what ends it, using the geometry of the position sizes and a real one-way market.

The panel above lays six steps side by side. Each bar doubles the one before it, and the total exposure line climbs far faster than most people picture when they read the rule.

The Martingale Strategy in One Paragraph

Start with the rule as written. Open a position. If it loses, open the next one at twice the size in the same direction, and repeat until a trade wins.

Where the Idea Comes From

Casinos supplied the original. A gambler betting on a coin doubles after each loss, so the first success returns every previous stake plus one unit.

Mathematically the sequence works, provided two conditions hold. The player needs unlimited money and the table needs no maximum bet.

Neither Condition Exists

Casinos cap the table precisely because of this system. Brokers cap position size, and your balance runs out long before either cap matters.

Those limits are not incidental details. They are the reason the sequence fails, and no amount of tuning removes them.

The Forex Version

Currency versions replace the coin with a signal. An oscillator, a moving average cross, or a simple time rule decides the direction of the first trade.

After that the doubling takes over. Signal quality then matters far less than most product pages imply, because sizing dominates the outcome.

How a Doubling Sequence Escalates

Six steps describe the whole method. Nothing hidden, nothing proprietary.

1. Open at base size. One trade, a fixed target and often no stop at all.
2. Take the small win. If it works, the sequence resets and starts again at base size.
3. Double after a loss. The next trade uses twice the volume, in the same direction.
4. Keep doubling. Each further loss doubles again, so the size climbs geometrically.
5. Recover on a win. One winner at the current size clears the accumulated losses and books a small gain.
6. Or hit a wall. Margin, a broker volume cap, or the balance itself ends the sequence first.

Steps one to five appear in every sales page. Step six appears in none of them, though it decides how the story ends.

The Arithmetic That Decides Everything

Doubling feels manageable for three or four steps. Then it stops feeling manageable very quickly.

Ten Steps in a Table

Read the third column carefully. It shows the total volume committed across the whole sequence, not just the current trade.

Step	Lot size	Total lots committed
1	0.01	0.01
2	0.02	0.03
3	0.04	0.07
4	0.08	0.15
5	0.16	0.31
6	0.32	0.63
7	0.64	1.27
8	1.28	2.55
9	2.56	5.11
10	5.12	10.23

What the Table Shows

Ten losing trades in a row take the position from a hundredth of a lot to over five lots. That is a five hundred fold increase in size on the tenth attempt.

Notice the last column too. The sequence has committed more than ten lots of total volume while trying to recover a starting risk of almost nothing.

Recovery Needs a Bigger Move Each Time

The average entry of the sequence drifts as each new trade lands. It moves toward the latest price, because the newest position dwarfs the older ones.

That helps a little and costs a lot. The distance back to break even shrinks slowly, while the money at stake grows twice as fast at every step.

Losing Runs Are Ordinary

Ten consecutive losses sound extreme, yet they arrive far more often than intuition suggests. A method that wins slightly more often than it loses still produces long adverse runs across a few thousand trades.

Run the numbers for your own settings with our risk of ruin calculator. The probability of a long streak rises sharply with the number of trades taken.

Cost Rises With Size

Spread and commission scale with volume. The tenth trade pays roughly five hundred times the transaction cost of the first.

Recovery therefore needs more than a simple retracement. It needs the move plus every accumulated cost along the way.

Why the Record Looks Wonderful for Months

Statements from martingale accounts flatter the method for a long time. Two features create that impression.

Almost Every Sequence Closes Green

A sequence ends when one trade wins, so the closed result reads positive nearly every time. Traders then quote a strike figure above ninety percent and feel justified.

That figure describes sequences rather than risk. It says nothing about how much money sat exposed while each sequence resolved.

The Losing Sequence Waits

Most months contain no sequence long enough to break the account. Six months of smooth growth therefore proves only that the long run has not arrived yet.

Time works against the trader here. Every extra month raises the chance of meeting the streak that ends it.

The Comparison That Helps

Think of a method that gains a little most weeks and loses everything rarely. Then think of one that loses a little most weeks and gains a lot rarely.

Both can show the same average. Only one leaves you with an account after the tail event, so read the shape rather than the headline.

Why the Idea Feels Right

The method persists because it fits how people think. Three habits of mind do the work.

A Run of Losses Feels Due for a Win

After five losses in a row, a sixth feels unlikely. Markets carry no such memory, so the next trade knows nothing about the last five.

That belief has a name. Anyone tempted by doubling should read up on the gambler’s fallacy first.

Booking a Loss Hurts More Than It Should

Closing a loser makes the loss real. Doubling instead keeps the hope alive for one more trade.

So the method appeals to a feeling rather than to a plan. It offers a way to avoid the moment of admission.

The Curve Rewards the Belief Early

Early sequences recover, and each recovery teaches the wrong lesson. Success arrives often enough to look like proof.

Rare events cannot be learned from experience. By the time the lesson lands, the account has usually gone.

A Trending Market Removes the Retracement

Doubling depends on one assumption. Price must come back far enough for the enlarged position to clear everything behind it.

Price ran up 7.0 ATR over 18 bars and the deepest pullback against that run was only 34 percent of the distance travelled — a position sold into it was never given a recovery.

That four-hour chart shows a normal trending week rather than a crash. Somebody selling into it doubled at each level and met a pullback that stopped a third of the way back.

A Third of the Way Back Is Not Enough

The enlarged position needs a specific distance, set by the average entry of the whole sequence. A partial retracement leaves the account underwater and the next double even larger.

So the failure is not dramatic. Price simply behaves ordinarily and declines to return.

Currencies Trend Regularly

Rate expectations shift, a commodity moves, or a central bank changes its tone. Any of those can produce weeks of one-way travel.

Nothing in the market owes a doubling sequence a retracement. Treat any claim that price must come back as an assumption under test, never as a property of markets.

What Actually Ends a Martingale Account

Three walls arrive, and usually in this order. None of them involves an unusual event.

Margin Binds First

Each new position consumes margin while the floating loss reduces equity. Free margin therefore falls from both directions at once.

The broker closes positions at the stop out level, and it does so at market. Our page on margin calls and stop outs describes that sequence in detail.

Volume Caps Interrupt the Sequence

Brokers set a maximum lot size per order and per symbol. Once the doubling reaches that ceiling, the next step cannot be placed at the required size.

The sequence then breaks mathematically. A partial double cannot recover the ground behind it, so the logic collapses even if money remains.

A Gap Skips the Recovery

Weekend gaps and news gaps move price without trading through the levels between. Any stop or recovery target sitting inside the gap simply gets passed over.

That single event can convert a manageable sequence into a terminal one. Nothing in the code prevents it.

Common Defences and Why They Fall Short

Defenders of the method offer four fixes. Each one helps a little and none removes the core problem.

The comparison above sets what a doubling sequence assumes against what markets actually do. Read it line by line before accepting any of the fixes below.

Use a Smaller Multiplier

Multiplying by one and a half instead of two slows the escalation. It also slows the recovery, so the sequence needs a bigger move to clear itself.

You have traded one risk for another. Slower growth simply pushes the wall further out.

Cap the Number of Steps

A hard cap at, say, six steps genuinely limits the damage. But hitting the cap means accepting a loss far larger than the intended base risk.

That is a sensible design. Notice, though, that it converts the method into an ordinary strategy with a very wide stop.

Add a Better Entry Signal

A stronger signal reduces how often long streaks appear. Streaks still appear, and the sizing decides the outcome when they do.

Our overview of expert advisor risk settings covers the controls that matter more than the entry.

Only Trade With the Trend

Trend filters help, and they fail exactly when trends turn. A sequence opened just before a reversal doubles straight into the new direction.

Filters also cut the number of trades. Fewer trades mean fewer sequences, so the smooth record takes longer to build and the tail risk stays the same size.

The Shape of the Equity Curve

Every martingale account produces the same picture eventually. Recognising the shape helps you spot the method behind a curve somebody shows you.

The panel above traces a long shallow climb made of many small gains, then one vertical drop straight through the starting line. Nothing gradual happens at the end.

Why the Drop Is Vertical

The largest positions in the sequence carry almost all the exposure. When they close, the loss arrives in a single moment rather than across a week.

A conventional strategy bleeds while it struggles. This one gives no warning at all.

Reading Somebody Else’s Curve

Smooth rising lines with hardly a down day should raise questions rather than confidence. Ask for the floating drawdown and the largest position ever opened.

Those two numbers expose the method immediately. A curve alone cannot, which is exactly why curves get published.

A Typical Account, Month by Month

Stories stick better than tables. Here is the arc these accounts tend to follow.

The First Month

The program runs. Small wins land most days. The balance ticks up in a neat line.

Nothing looks odd yet. So the trader relaxes and stops checking the open positions.

Weeks Six to Twelve

Growth carries on. One short streak hits in week eight, and the program doubles twice before a win lands.

The curve barely dips. Faith grows, because the rule appears to have handled a bad patch.

The Fourth Month

Confidence turns into action. The trader adds funds and raises the base lot size.

Both moves feel earned. Both also move the wall much closer, since every step now costs more margin.

The Fifth Month

A trend starts. The program sells into it and doubles at step three, then four, then five.

Free margin drops fast. The trader watches, hopes for a bounce, and waits.

The Last Hour

The broker closes the stack at market. Months of gains vanish in a single session.

Nothing broke. The rule simply ran out of room, which was always the way it would end.

Five Questions to Put to a Seller

Short questions work best here. Vague answers tell you as much as clear ones.

Does Size Ever Rise After a Loss?

Ask it plainly. A yes means you hold a doubling sequence, whatever the product calls itself.

What Was the Largest Lot Ever Opened?

Compare that number with your own margin. If your account cannot hold it, the test never applied to you.

How Deep Did Floating Loss Go?

Closed profit hides the risk on this method. Floating loss shows it, and every tester reports the figure.

Is There a Hard Step Cap?

A cap turns the method into an ordinary wide-stop strategy. No cap leaves the ending to your broker.

Check that the cap fires in a test rather than sitting unused in the settings. Plenty of programs list a limit that the code never actually reads.

Which Market Broke It?

Every rule has a bad market. A seller who cannot name theirs has not looked, or would rather not say.

Push for a date and a symbol. Then run the same settings over that stretch yourself and read the floating loss.

Softer Variants You Will Meet

Products rarely use the word martingale. Three near relatives show up under friendlier names.

Averaging Down

Adding at better prices without doubling grows exposure more slowly. Risk still rises as price moves against the position, and the same wall waits at the end.

Grid With Progressive Sizing

A ladder of orders that increases volume at deeper levels behaves like a doubling sequence with extra steps. Our comparison of grid trading against martingale separates the two clearly.

Hedged Recovery Baskets

Some products open an opposite trade instead of doubling straight away. Net exposure then looks small on the screen.

Both sides still sit open, so both pay spread and swap. The decision to unwind them has merely been delayed, and it usually arrives with less margin available than before.

Recovery Mode

Some programs advertise a recovery setting that reopens a larger trade after a loss. The label differs, and the arithmetic does not. Our guide to grid trading covers the closely related ladder version.

How to Test a Doubling Program Safely

Curiosity is reasonable, and the mechanism teaches a genuine lesson about exposure. Four precautions keep the education cheap.

Test on the Worst Periods Deliberately

Choose stretches where a currency ran hard in one direction, then run the same settings there. A quiet year tells you nothing useful.

Read Floating Drawdown, Never Closed Profit

Closed profit on this method looks excellent by design. The floating figure carries the information, so find it before anything else.

Count the Largest Position

Find the biggest single trade in the test and compare it against your own margin. Many programs pass only because the simulated account held more room than yours does.

Size the Base Risk for the Whole Sequence

Work out the total exposure at your step cap, then treat that figure as the real risk of the first trade. Our position size calculator turns that into a lot size you can live with.

Better Uses for the Same Effort

The appeal of doubling comes from wanting recovery without a plan. Three ordinary habits deliver more.

Fix the Risk Per Trade

A constant fraction of equity removes the temptation to chase. Losses then shrink the position automatically, which is the opposite of doubling.

Our guide to position sizing covers the standard approaches and where each one fits.

Understand Your Real Ruin Probability

Ruin depends on three things: the share of trades that work, the average size of wins against losses, and the fraction risked each time. Our note on risk of ruin shows how they interact.

Build a Process You Can Inspect

A rule you can read beats a recovery mode you cannot. Browse the forex strategy library for approaches with a defined loss on every trade.

FAQ

Does the martingale strategy ever work?

Sequences finish green very often, and that is precisely what makes the method dangerous. The rare sequence that fails removes far more than every successful sequence contributed, so the record and the risk point in opposite directions. Nobody can tell you when the long streak arrives, only that longer trading makes it more likely.

How many losing trades in a row can an account survive?

Fewer than most people expect. Divide your balance by the margin needed at each doubled step and count how far the ladder reaches, remembering that floating losses reduce equity at the same time. Most retail accounts run out somewhere between the seventh and tenth step, and broker volume caps can intervene earlier.

Is a smaller multiplier safer?

It slows the escalation, which buys extra steps. Recovery slows as well, since a smaller position needs a larger price move to clear the accumulated loss behind it. You are moving the wall, not removing it.

Why do martingale programs sell so well?

Because the equity curve looks superb until the day it does not. Months of small consistent gains persuade buyers, and the failure mode stays invisible until a trend arrives. Sellers rarely publish floating drawdown or the largest position opened, and those two numbers would end most conversations.

Can a stop loss make doubling safe?

A hard cap on the number of steps, combined with a stop, does limit the damage. At that point the method has become a wide-stop strategy with escalating size, which behaves nothing like the original idea. Evaluate it on those terms rather than on the doubling logic.

Is averaging down the same thing?

Not identical, though it shares the shape. Averaging down adds to a losing position without necessarily doubling, so exposure grows more slowly while still growing as price moves against you. Both approaches trade a comfortable record for a rare large loss. Results are not guaranteed; past performance is not indicative of future results.