Checklist for Spotting Dangerous EAs and Fraudulent Auto-Trading Systems | 10 Points to Avoid Being Fooled by EAs That Only Tout Win Rate and Monthly Returns

Many people have seen EA (automated trading system) advertisements filled with phrases like “95% win rate,” “30% monthly return,” and “just leave it running.” Yet stories of accounts being wiped out after actually using one never seem to stop. Why is that? The answer is simple: the numbers emphasized in the marketing and the numbers that actually matter once you’re running the system are completely different things.

This page is the flagship article we most want to invest in as a “verification media” outlet. Rather than naming and criticizing specific EAs, it brings together, in language and numerical examples accessible even to beginners, plus a checklist, how to read “warning signs” — what kind of presentation should make you stop and think twice. Reading this will help you notice the “numbers not being shown” behind flashy marketing copy. If you haven’t yet read What Is an EA (Automated Trading System)? or How to Read EA Performance Metrics, going through those first will make the content of this article much clearer.

What Is a “Dangerous EA”? (Why Do People Keep Getting Fooled?)

First, let’s clarify one thing: a “dangerous EA” doesn’t necessarily mean only a “malicious, fraudulent EA.” Even without malicious intent, if the risk isn’t adequately disclosed, or if the design happens to work only in certain market conditions, it can still be plenty dangerous for the person using it. Here, we use “dangerous EA” to refer broadly to any EA that makes it easy for users to blow their account.

So why do people keep getting fooled? There are three main reasons. First, people are drawn in by “good numbers.” When told “95% win rate,” it creates the impression of winning 95 times out of 100, but the actual design might be one where a single loss wipes out the profit from all 95 wins. Second, marketing works fine without showing the inconvenient numbers. Since there’s no obligation to disclose maximum drawdown (DD) or floating losses, most sellers keep them hidden. Third, beginners are the least likely to think in terms of “verifying” a claim. If you don’t know what to check before buying, all you can do is trust the advertisement.

The purpose of this article is to fill that third gap. People with experience learning the market through discretionary trading tend to notice something “off” about dangerous EAs more easily — and once that vague sense of “something’s fishy” is put into words, beginners can develop the same eye for it. Let’s go through it step by step.

10 Warning Signs of a Dangerous EA

Now for the main subject. Dangerous EAs share common patterns. Matching one of the ten signs below doesn’t automatically disqualify an EA, but the more of them that overlap, the sharper the risk rises. In particular, signs related to “not showing inconvenient numbers” are strong warning signals even on their own.

1. It Emphasizes Only Win Rate and Monthly Return

This is the most common warning sign. Numbers like “90% win rate” or “20% monthly return” guarantee absolutely nothing about safety on their own. The reason is explained in detail in How to Read EA Performance Metrics, but even with a high win rate, if the loss per losing trade is larger than the profit per winning trade, the strategy loses money overall. For example, even at a 90% win rate, if the average profit is 1,000 yen and the average loss is 15,000 yen, then over 100 trades you’d get “90 wins x 1,000 yen = 90,000 yen, 10 losses x 15,000 yen = 150,000 yen,” for a net loss of 60,000 yen. When an advertisement has only a single number — win rate — doing all the dancing, assume there’s a good chance it’s hiding other, less favorable numbers.

2. It Doesn’t Disclose Maximum DD or Floating Losses

Maximum drawdown (DD) is a number showing how far equity has fallen from its peak. Floating loss is the temporary, unrealized loss carried by positions that haven’t been closed yet. Together, these two figures are the lifeline of live operation, showing “how much of a drop you need to be prepared for along the way if you use this EA.” And yet, most advertisements don’t include them. Showing the profit (the result) while hiding the risk of the process that got you there — from a verification-media standpoint, this is the presentation to watch most closely. Conversely, the reason our lab discloses SMC Gold Sniper’s 8.2% maximum DD and MAC v2.0’s maximum floating loss is precisely to differentiate ourselves on this single point.

3. Backtest Only, No Forward-Test Track Record / Short Period or Few Trades

A backtest is a “mock exam taken against past market data,” while a forward test is verification by actually running the system on real, upcoming market conditions. An EA that’s simply been fitted tightly to past data (a practice called curve fitting, or over-optimization) can look brilliant in a backtest yet collapse easily once it goes live. That’s exactly why an EA with no forward-test track record should be considered “unproven in live conditions.” Caution is also needed if the verification period is only a few months or the trade count is only a few dozen. With so few trades, the results could just be a lucky streak, and statistical confidence is low. As a rough benchmark, our lab treats verification spanning multiple years and several hundred trades or more as one baseline standard.

4. Averaging-Down or Martingale With No Risk Explanation

Averaging down is a technique of adding to a position as price moves against you in order to lower your average entry price, while martingale takes it further by doubling the lot size each time. Both carry a strong bias: in exchange for “looking like” a high win rate, the floating loss balloons all at once when the trade eventually goes wrong. If the adverse move continues and the additions stretch to the maximum number of steps, the floating loss builds up stage by stage, and if there isn’t enough capital, it can end in a forced margin call (a single account-ending event). Averaging-down/martingale EAs aren’t inherently bad — what’s dangerous is not explaining that risk in numbers. The reason our lab’s MAC v2.0 (GOLD-only, 1.2x multiplier x up to 15 steps, 30-pip spacing, no hard SL) deliberately discloses the floating loss under the worst-case scenario is to fulfill this “duty of explanation.” The mechanics are explained with figures in Stop-Losses and Money Management.

5. Recommended Margin Is Vague / Lot Size Changes Partway Through

Recommended margin is a guideline for “how much capital you should run this with,” and properly speaking, it can be stated with a rationale, such as “required margin + max DD x 2” (see EA Money Management for details). If this is only stated vaguely, like “starting from 100,000 yen is fine,” there’s a good chance the maximum DD hasn’t even been calculated — meaning the risk isn’t understood. Even more dangerous is the case where, to make the results look better, the lot size (trade volume) is quietly increased partway through the verification. Raising the lot size only in the latter half makes the profit figure jump, but that’s not skill — it’s just “piling on risk.” Whether verification was carried out consistently with the same lot size the whole way through is a point you should always confirm.

6. It Hides Losing Months or Suspension History

Every EA, without exception, will have losing months where it doesn’t mesh with market conditions. If none of those are shown and only winning months are lined up, that’s a sign of “hiding the inconvenient months.” Likewise, there are cases where a history of large past losses that led to suspended operation (and the quiet resumption that followed) is kept hidden. An EA that publishes its losing months is, if anything, more trustworthy — this may sound counterintuitive, but it’s a consistent position our lab holds as a verification media outlet. That’s exactly why our lab publishes losing months as well on the performance dashboard.

7. It Hypes “Guaranteed,” “Explosive Profit,” or “Set and Forget”

“Guaranteed to win,” “explosive profits while you do nothing,” “anyone can easily make money” — if language like this is lined up, that alone is reason enough to keep your distance. There’s no such thing as certainty in the markets, and an EA isn’t magic — it’s nothing more than “written rules” (How an EA Works). Treat absolute, gambling-impulse-stoking language as something designed to rob you of calm judgment. In the financial world, as a rule, the more certainty an expression promises, the further it tends to be from reality.

8. No Data Source or Third-Party Verification

Performance screenshots are, in fact, easy to fabricate. That’s precisely why it matters whether the track record is linked to an account through a third-party verification service like Myfxbook, and whether the testing environment (broker, spread, period, modeling quality) is clearly stated. Consider a performance sheet with nothing but “just trust me” and no stated verification conditions weak as decision-making material. Our lab, too, states the conditions for SMC Gold Sniper’s verification as “GOLD / M30 / 2018-2026” precisely so that anyone can independently replicate the test afterward.

9. It Rushes You Toward Purchase or Sign-Up Instead of Explaining Risk

“Only now,” “just X spots left,” “sign up from this link right now” — a path that uses this kind of manufactured urgency to rush you into registering or buying without giving you time to verify is a warning sign. Properly, an EA or copy-trading service is something only chosen by people who understand what the numbers mean, have their money management in place, and are convinced. If there’s a need to rush you, it’s natural to assume there’s something that wouldn’t hold up under careful scrutiny.

10. It Sells a Design With No Stop-Loss (SL) at All as a “Strength”

Be especially wary of the pitch “it never loses because it never cuts losses.” Not realizing a floating loss just means the loss never gets recorded — in reality, it keeps growing as an unrealized loss. Combined with averaging down, this raises the risk of a sudden blowup if the adverse move continues. A design with no hard SL (a forced stop-loss) isn’t automatically disqualifying on its own, but if it’s being sold as “proof there’s no risk,” that should be seen as hiding the true nature of the risk. Our lab’s MAC v2.0 is also designed without a hard SL — which is exactly why we disclose the maximum floating loss in numbers and explicitly position it as “a high-risk verification allocation for surplus funds.”

Dangerous EA Checklist (Quick Reference)

The 10 signs above have been organized into a ready-to-use checklist. As you look at a marketing page or performance sheet, try marking off any items that apply. The more items that apply, the higher the risk. Weigh items related to “numbers not being shown” especially heavily.

Warning Sign	What to Check	Risk Level
Emphasizes only win rate/monthly return	Are other metrics (DD, RR, trade count) also disclosed?	High
Max DD/floating loss undisclosed	Does it state how far equity can drop along the way?	Highest
No forward-test track record / short period	Is live operation (real account or demo) demonstrated?	Highest
Few trades	Is verification based on hundreds of trades or more? (Tens of trades warrant caution)	High
Uses averaging-down/martingale without explanation	Is the worst-case floating loss, number of additions, and presence/absence of SL disclosed?	Highest
Recommended margin is vague	Is there a rationale such as “required margin + max DD x 2”?	Medium
Changes lot size mid-test	Was verification conducted consistently with the same lot size?	High
Hides losing months/suspension history	Are losing months and any suspension history also disclosed?	High
Hypes words like “guaranteed,” “explosive profit,” “set and forget”	Are there absolute claims or language designed to stoke gambling impulses?	High
No third-party verification or data source	Are verification conditions (broker, period, quality) clearly stated?	Medium
Rushes you to sign up	Does it give you time to verify before deciding?	Medium

Why This Kind of Discernment Matters — Buying Without Verifying Is Buying an “Invisible Bomb”

Let’s return to basics for a moment. Why is it necessary to check warning signs in this much detail? Because an EA is something that, “once switched on, keeps trading your capital automatically.” With discretionary trading, you can stop the moment something feels dangerous. But an EA keeps trading according to its programmed rules while you’re asleep, while you’re traveling — regardless. Deploying one without being able to spot the warning signs is like planting a bomb of unknown contents in your account.

Our lab’s position as a verification media outlet is clear: not “buy it because it looks good,” but “confirm even the inconvenient numbers first, then choose after understanding the risk.” Only people who can maintain this order can master an EA as a tool. Learning to discern these things is a skill less about making money and more, first and foremost, about not getting knocked out of the game.

Applying This to Discretionary Trading — Putting the “Something’s Off” Feeling Experienced Traders Notice Into Words

Anyone with even a bit of discretionary trading experience may get a vague sense of “something’s fishy” about a dangerous EA. Once you can put words to what that feeling actually is, it becomes an “eye for spotting it” that can be shared with beginners too.

For instance, watching the market through discretionary trading, you come to feel in your bones things like “there’s no way you keep winning in one direction forever” and “there’s no way the same approach wins equally well in both ranging and trending conditions.” Because of that, wariness kicks in naturally toward any EA that claims a high win rate across every market condition. Also, if you have the bitter experience of hesitating on a stop-loss and getting stuck holding a floating loss, you’ll intuitively grasp just how detached from reality the pitch “it never loses because it never cuts losses” really is. The floating loss hasn’t disappeared — it’s simply not yet realized.

This “sense of something being off, born from discretionary experience,” sharpens the more you learn market fundamentals through Stop-Losses and Money Management and Dow Theory. Discretionary trading and EAs might seem like completely different worlds, but they actually rest on the same premise: “you never know what the market will do.” The market sense built through discretionary trading becomes, as is, a weapon for spotting the dangers in an EA.

Useful for Evaluating EAs and Auto-Trading Systems — Reading a Performance Sheet in “Order of Risk”

When you’re actually handed an EA’s performance sheet, where should you start reading? From the standpoint of spotting a dangerous EA, the golden rule is to read starting from the “hidden numbers,” not the “good numbers.” Here’s the order.

First, check whether the maximum DD and floating loss are listed. If not, that alone makes it insufficient as decision-making material. Next, look at the forward-test track record along with the verification period and trade count. If it’s backtest-only, or the period is short, mark it as “unproven in live conditions.” Then check whether averaging-down or martingale is used, and whether the risk is explained. Finally, evaluate win rate and monthly return “as a set with the other numbers.” Reading in this order lets you catch the warning signs before your eyes get pulled in by flashy figures.

The foundation for this way of reading is How to Read EA Performance Metrics. If you know what indicators like profit factor, maximum DD, recovery factor, and expected value mean, you can immediately spot which parts of a performance sheet are left blank. The table below, which maps discretionary trading terms and SMC terms to how they’re handled in EA operation, will also help you translate an EA’s performance into “your own words.”

Common Discretionary Trading Term	SMC Term	How It’s Handled in EAs (Auto-Trading)
Stop-loss (giving up here)	Liquidity swept / outside structure	SL setting. Without a hard SL, losses accumulate as floating loss (Warning Sign 10)
Adding to a position (averaging down)	Additional entry at a discount	Averaging-down conditions. The number of additions, spacing, and multiplier determine the max floating loss (Warning Sign 4)
Losing streaks / big losses	Phases where the strategy fails to adapt to a shift in market conditions	Max DD, max losing streak. Hiding this is the biggest warning sign
Winning streaks / losing months	Compatibility with trending vs. ranging periods	Monthly track record. Non-disclosure of losing months = Warning Sign 6
“This feels like it’ll work”	Edge based on reading market context	Must be distinguished from optimization to past data (curve fitting)

Applying AI Analysis — Having AI Read a Suspicious Performance Sheet to Detect What’s Off

Checking warning signs one by one by hand can be a lot of work until you get used to it. So what our lab actually practices is having an AI read a performance sheet or an EA’s description and sort out “what’s disclosed and what’s missing.” Unmoved by emotion, the AI flatly picks out gaps such as “there’s a win rate but no DD,” “the trade count is low,” or “averaging down is mentioned but there’s no risk explanation.”

For example, if you have an AI read a description of an averaging-down EA and ask it to reconsider “up to how many steps does it add to the position,” “what is the total lot size and expected floating loss at that point,” and “is there a hard SL,” assumptions that aren’t written in the marketing copy start to surface. In the case of our lab’s MAC v2.0, having the AI translate the design “1.2x multiplier x up to 15 steps, 30-pip spacing, no hard SL” produces the conclusion: “if it stretches to 15 steps, the total lot size becomes more than ten times the initial size, and the floating loss balloons proportionally — which is why it should be limited to a verification allocation of surplus funds.” This matches what’s quantified in Stop-Losses and Money Management. AI isn’t a magic judgment machine, but as a “verification partner that reduces oversights,” it’s extremely effective. “Using AI to translate difficult metrics” refers to exactly this kind of use.

Conversely, a Framework for Choosing an EA Safely (→ EA Evaluation Template)

We’ve been looking at “warning signs” up to this point, but flip them around and they also serve as a reverse checklist for choosing a relatively safe EA. In other words, an EA that discloses not just win rate but also DD and trade count; that has both a backtest and a forward test; that, if it uses averaging down, explains the worst-case scenario in numbers; that has a rationale for its recommended margin; and that also publishes losing months — an EA meeting these conditions is, at minimum, one where “the decision-making material is all there.” That still doesn’t guarantee profit, but it puts you in a position to choose with an understanding of the risk.

This reverse checklist has been systematized into the FX AI Lab EA Evaluation Template. Across nine axes — backtest soundness, forward-test consistency, risk level, averaging-down risk, capital efficiency, stability, market resilience, operational difficulty, and beginner-friendliness — it puts into words not “strong vs. weak” but “what kind of risk tolerance this suits.” And the safeguard against blowing up after you’ve made your choice is EA Money Management. Only when all three are in place — an eye for spotting danger (this article), a framework for multi-angle evaluation (the evaluation template), and the skill of protecting yourself with capital (money management) — can you truly treat an EA as a tool. The EAs our lab actually operates disclose all of these numbers on the EA Library and the performance dashboard.

Summary

Boiled down, how to spot a dangerous EA comes down to a single principle: “don’t search for good numbers — search for hidden numbers.” If only win rate and monthly return are doing the dancing, while max DD, floating loss, forward-test track record, an explanation of averaging-down risk, and losing months are all missing, that may be a sign of something that “can’t be shown.” Keep the 10 warning signs and the checklist close at hand, and build the habit of reading marketing pages and performance sheets in “order of risk.”

Experience learning the market through discretionary trading becomes a weapon for spotting what’s off about an EA. And once you’ve spotted it, the framework for choosing safely is the EA Evaluation Template, and the safeguard against blowing up is EA Money Management. The reason our lab publishes even our own EAs’ maximum DD, floating losses, and losing months is that we believe the “transparency” raised throughout this article is precisely the foundation of trust for a verification media outlet. If you’d like to understand the meaning of these numbers more deeply, head to How to Read EA Performance Metrics; actual verification data can be checked on the performance dashboard.

Frequently Asked Questions

• Q. Is it okay to buy an EA with a win rate of 90% or higher?
  A. Win rate alone isn’t enough to judge. Even with a high win rate, if the loss on a single trade is bigger than the gain, you lose overall. This is a trap common among averaging-down/martingale types. Always check it together with maximum DD, average profit/loss, and trade count (How to Read EA Performance Metrics).
• Q. Are all averaging-down EAs dangerous?
  A. No. Averaging down as a technique isn’t inherently bad — what’s dangerous is not explaining the worst-case scenario (maximum number of steps, total lot size, expected floating loss, presence of an SL) in numbers. Our lab’s MAC v2.0 deliberately discloses this and explicitly states it as a high-risk verification allocation for surplus funds (Stop-Losses and Money Management).
• Q. Is it okay to trust an EA if there’s a performance screenshot?
  A. Because screenshots are easy to fabricate, they aren’t sufficient on their own. Prioritize ones where the verification conditions (broker, period, spread, trade count) are clearly stated and, ideally, can be confirmed through a third-party verification service. Consider a performance sheet with no stated conditions weak as decision-making material.