A 99% win rate is the easiest number in the world to fall in love with. You look at the backtest, almost every trade is green, and the curve climbs like a staircase. If you’re looking at a martingale EA on MT5, or a grid, that staircase is real. So is the one night at the bottom of it.
I build EAs, and none of mine use a grid or a martingale. That’s my bias on the table. This post isn’t about any single product: it’s about the maths, so you can spot it yourself before you pay for it. Not financial advice.
Grid and martingale, in plain words
Both do the same thing when price moves against them: they don’t close the trade, they add another one.
- A grid adds a new trade every so many dollars, usually with the same or a growing lot. When price comes back a little, the whole basket closes in profit.
- A martingale doubles the lot after each loss or each step against it, so one small bounce recovers everything.
Most of the time, price does come back. That’s why the win rate looks so good. The question is what happens the time it doesn’t.
Why a 90% win rate can still lose money
Win rate tells you how often you win. It says nothing about how much. What decides whether a strategy makes money is the expectancy: win rate times the average win, minus loss rate times the average loss.
Take a hypothetical gold strategy that wins 90% of its trades. If the average win is $10 and the average loss is $120:
0.90 × $10 − 0.10 × $120 = $9 − $12 = −$3 per trade.
Nine winners out of ten, and it loses money. That’s the shape behind most “gold trading strategy with 90% win rate” claims: lots of small wins paid for by rare, large losses. Grids and martingales push that shape to the extreme, because the rare loss isn’t large. It’s the account.
The maths of one bad move
Here’s a hypothetical example on gold: one trade every $5 against you, starting at 0.01 lots (one ounce at most brokers), measured when price is that many steps below the first entry.
| Steps against you | Grid, 0.01 each: open lots / floating loss | Martingale, lot doubles: open lots / floating loss |
|---|---|---|
| 5 ($25 move) | 0.05 lots / $75 | 0.31 lots / $285 |
| 10 ($50 move) | 0.10 lots / $275 | 10.23 lots / $10,180 |
| 20 ($100 move) | 0.20 lots / $1,050 | Impossible: over 10,000 lots |
A $100 move in gold is about 2.4% at $4,100 an ounce. Not a crash. And while all that is open, every winning basket was paying you a few dollars. The grid needs hundreds of small wins to pay for one move like that. The martingale doesn’t even get there: the broker closes it for lack of margin long before.
That’s not theory. In my Quantum Queen X review, a grid EA won 1,589 baskets in a row on gold, then a drop of about one percent took a new $1,000 account to $57 in one night, on its default settings.
How to spot a grid or martingale EA before you buy
- No stop loss on the orders. Open the backtest report and look at the deals. If trades have no stop loss, something else decides when to close, usually a basket target.
- Several trades in the same direction at once. Three, five, nineteen buys open on the same symbol is a grid.
- Inputs like “step”, “distance”, “lot multiplier”, “max orders” or “recovery”. A multiplier above 1.0 is a martingale.
- A win rate above 90% with an average loss many times the average win. Check both numbers in the report, not just the first.
- Equity far below balance during open trades. If the report shows a balance drawdown much smaller than the equity drawdown, the losses were sitting open and hidden.
- A drawdown protection that comes switched off. If the safety is an input, check its default.
Then test the worst day yourself, with real ticks and your own account size: here’s how to backtest an EA on MT5 without fooling yourself. A single backtest is one path; a Monte Carlo simulation shows you the others.
If you still want to run one
Some people do, and some make money with them. If that’s you:
- Switch on its drawdown cut, or set a hard equity stop, and accept that it will sometimes take a big loss to save the account.
- Take your profits out as you make them. The money left in the account is what the next bad move takes.
- Never on a prop firm challenge: any daily loss limit is hit long before the broker’s stop out.
- Size it as if the worst night will happen, because over enough years it does.
What I run instead
My answer to the same problem is a portfolio, not a grid. DoIt MultiStrategy Pro runs seven strategies on gold, USDJPY and GBPJPY, and every trade has a stop loss from the moment it opens. No grid, no martingale: each strategy takes at most two trades a day, sizes every trade by the money at risk, and closes at the end of the day and before the weekend.
It has losing streaks, like any strategy that’s honest about its losses. The difference is that a bad day costs a known amount, not the account. Its live track started in September 2026, and you can follow it on the product page.
If you’re thinking about how much of your money one strategy should carry, I wrote about that in EA portfolio risk: count the bets behind your charts.
See a stress test done properly: in The Gold Reaper Review: I Tried to Break It, a breakout EA with a stop on every trade goes through 1,000 alternative histories. It’s the opposite case: it held up.
If you test on gold, the data matters. I use Axi‘s XAUUSD history for my tests. Affiliate link; I may earn a commission if you open an account through it. And if you’d rather grow with allocated capital than your own, read the rules first: Axi Select only allows a third-party EA for generating signals, with trades placed manually.
I break EAs so you don’t have to find out with your own money. Join the DoItTrading newsletter to get the next test.
The grid and martingale figures are a hypothetical example, not a backtest. Trading involves risk, and this is not financial advice.
Frequently Asked Questions
Is a martingale EA profitable?
It can be profitable for long stretches, because it wins most of its trades. But it adds or doubles positions when price moves against it, so one long move can lose more than all the previous wins together. Its expectancy depends on how big that rare loss is.
What is the difference between a grid EA and a martingale EA?
A grid adds a new trade every so many dollars against you, usually with the same or a growing lot. A martingale doubles the lot after each loss or step. Both close the whole basket when price comes back a little, and both can carry a large open loss while they wait.
How can I tell if an EA uses a grid or martingale?
Look for orders without a stop loss, several trades open in the same direction, inputs like step, lot multiplier or max orders, and a win rate above 90% with an average loss many times the average win. A big gap between equity and balance drawdown is another sign.
Can a gold strategy with a 90% win rate lose money?
Yes. With a 90% win rate, an average win of $10 and an average loss of $120, the expectancy is minus $3 per trade. What matters is the win rate and the size of wins and losses together.
Can I use a martingale EA on a prop firm account?
I wouldn’t. Prop firms have daily and total loss limits, and a grid or martingale carrying an open loss usually breaks them long before the broker would close the trades. Read each firm’s rules on EAs before you try.
Does DoIt MultiStrategy Pro use a grid or martingale?
No. Its seven strategies each use a stop loss on every trade, take at most two trades a day, size by the money at risk, and close at the end of the day and before the weekend.
