How to Spot a Forex EA Scam: 9 Checks

Most EAs that disappoint are not frauds. They are ordinary strategies sold with the losing half left out. These nine checks separate the two, and every one of them is something you can verify yourself in a few minutes, without taking anyone's word for it.

Last updated 18 September 2026

The distinction worth holding on to

There are two different problems, and they need different responses.

An overstated EA is a real strategy sold optimistically: the backtest is cherry-picked, the risk is underplayed, the screenshots are the good months. You can often still use it, at a smaller size than the seller suggests, once you have worked out the real risk yourself.

A dishonest one is selling something that does not exist: an invented track record, a demo account presented as live, or a strategy that hides a martingale behind the word "AI". Nothing you do with the settings makes that safe.

The checks below sort one from the other.

1. Is there a drawdown figure next to the return?

"340% in 8 months" is not a claim, it is half of one. The other half is what the account was worth at its worst point, and any seller who has run the strategy knows that number.

If the drawdown is absent from the sales page entirely, treat every other number on it as decoration.

2. Can you verify the account, and is it the account you would run?

A screenshot proves nothing; anyone can edit one. A verified Myfxbook page or an MQL5 signal proves the trades happened.

Then read the account details, which is where most of the useful information sits:

3. Are the default settings published?

You should be able to see the lot size, the multiplier if there is one, the maximum number of positions, the stop loss and the take profit before you pay.

With those numbers you can work out the worst case yourself, which is exactly why a dishonest seller does not publish them. If the inputs list only appears after purchase, assume the reason is arithmetic.

4. Does it multiply position size after a loss?

Ask directly: does it use a grid, a martingale, averaging down, or "recovery" mode? All four mean the same thing in practice.

This is not automatically disqualifying. A great many working EAs are grids, ours included. What matters is whether the seller says so plainly and shows the ladder, or calls it "smart recovery" and changes the subject. The arithmetic is in grid vs martingale.

5. Does the language promise what no strategy can deliver?

Certain phrases are not exaggeration, they are a category error. "No loss", "guaranteed profit", "100% win rate", "risk free". Markets do not offer those, so a seller using them is either lying or does not understand their own product.

"AI powered" deserves its own line. Sometimes it is true. Very often it describes a moving average crossover. Ask what the model is, what it was trained on, and what it does when it is wrong. A real answer exists; a marketing answer does not.

6. Can you test it before paying?

On the MQL5 Market, every product can be downloaded as a free demo and backtested in the Strategy Tester with no time limit. A seller on their own website should offer either a trial, a refund window, or a verifiable record.

Nothing to test and nothing to verify, on a product that costs hundreds, is the strongest signal on this page.

7. Is there a refund policy with conditions you can read?

A vague "money back guarantee" is worth less than a specific one. "14 days, provided the licence has not been activated" is a promise that can be kept or broken in a way you could point to afterwards.

Check where the payment goes, too. A card payment through a recognised processor gives you a dispute route; a direct crypto transfer to a personal wallet gives you none.

8. Is there a person behind it?

Look for a name, a country, a support address that replies before you pay, and a presence that is older than the product. Ask a specific technical question and read the answer: whether they know their own EA is obvious within two replies.

9. What do the negative reviews say?

Ignore the five star reviews; read the one and two star ones. You are not looking for whether people lost money, which happens with every strategy. You are looking for a pattern: did the seller stop answering, did the product change after purchase, did the licence stop working.

On MQL5, reviews can only be left by people who actually bought the product, which makes that section considerably more useful than reviews on the seller's own site.

Running the checks on us

It would be dishonest to publish a checklist and dodge it, so:

Where we do not pass: there is no free trial on the website, and the live track record is not yet long enough to include a violent market. Both are true, and you should weigh them.

If you are working through a specific product now, reading a backtest report covers what to check inside the numbers they send you.

Common questions

Are all forex EAs scams?

No, but the marketing around them is unreliable often enough that you should assume nothing. Automated strategies are ordinary software; the dishonesty, when it appears, is in the claims, the hidden risk model and the fake track records rather than in the idea of automation.

What is the single biggest warning sign?

A claimed return with no drawdown beside it. Any real strategy can tell you what its worst period cost, and any seller who cannot, or will not, is either hiding it or has never looked.

Is a verified Myfxbook or MQL5 signal proof that an EA works?

It is evidence that the trades happened, which is a great deal more than a screenshot. It does not prove the account is the one you would run, that the settings are the ones you are sold, or that the record is long enough to include a bad period. Check the account type, the age, and whether it is a cent account, which makes the currency figures look larger than they are.

What does a martingale EA look like in a track record?

A long, almost straight rising line with a very high win rate, and occasionally one enormous drop. If the equity curve is smooth and the win rate is above 90 per cent, assume position sizes increase after losses until you have been shown otherwise.

How much should a genuine EA cost?

There is no honest answer in the price alone: real products exist from about $50 to several hundred, and so do bad ones. What the price should buy you is a stated refund window, a licence that covers more than one account, and a person who answers questions before you pay.