Buying an Expert Advisor is different from buying most software. You are not just paying for a tool — you are trusting a set of rules to place real trades with real money while you are asleep, at work, or simply not watching the screen. A slick landing page, a curve-fit equity chart, and a countdown timer telling you the "discount" ends tonight are not due diligence. Asking the right questions before you buy is the single highest-leverage thing you can do to separate a legitimate, rules-based system from a marketing gimmick. This guide walks through exactly what to ask, why each question matters, what a good answer sounds like, and what a red-flag answer sounds like — with worked numbers so you can see the difference in dollars, not just theory.

Why Vetting an EA Developer Matters More Than Vetting the Software Itself

Most buyers spend their time comparing win rates and equity curves. That is backwards. An equity curve is a snapshot of the past; the developer's transparency, honesty, and support practices are what determine whether you can trust the next twelve months. The CFTC's advisory on trading system fraud makes this explicit: many trading-system scams are built around backtested or hypothetical results that were never traded live, sold by people who disappear once the refund window closes. A developer who is confident in their product will answer hard questions directly. One who is not will change the subject, cite "proprietary" secrecy for basic operational facts, or point you to a testimonial page instead of a data source.

Before you hand over a license fee, treat the conversation with the developer the same way you would treat a job interview for someone who is about to manage part of your capital. You would not hire a portfolio manager based on a five-star review; you would ask about their process, their track record, and their risk controls. The same standard should apply to a piece of software that is about to place trades on your MetaTrader 4 or MetaTrader 5 terminal.

1. What Exactly Does the Strategy Do, and Can You Describe the Logic in Plain English?

You do not need — and should not expect — the developer to hand you their exact source code or indicator formulas. That is legitimate intellectual property, the same way a fund manager will not disclose their exact model. But a legitimate developer should be able to describe, in plain language, the category of strategy you are buying: Is it trend-following? Mean-reverting? Does it scalp small moves or hold trades for hours? Does it trade one instrument or many? What timeframe does it operate on?

A vague answer like "it uses advanced AI algorithms to predict the market" with no further detail is a warning sign. A better answer sounds like: "This is a rules-based trend and momentum confirmation system that trades XAUUSD only, evaluates setups on the H4 chart, and takes at most roughly one trade per day because it waits for a specific confluence of conditions before entering." That is specific enough to evaluate without giving away the proprietary logic.

Also ask directly: does the system use martingale, grid, or averaging-down to recover losing trades? These recovery techniques can make an equity curve look smooth for months and then produce a catastrophic drawdown in a single volatile session. A developer who says "no martingale, no grid, no averaging" and can explain how position sizing works instead (see Question 5) is giving you a real answer you can verify against their trade history.

2. Is the Performance Track Record Independently Verified?

This is arguably the most important question, and the easiest for a scam vendor to fake. Backtest screenshots and MetaTrader strategy tester reports are trivially easy to curve-fit — you can optimize a strategy after the fact to look flawless on historical data and still watch it fail in live conditions. Ask specifically: "Is there a live, verified trading account I can inspect, and who verifies it?"

Independent verification services like Myfxbook connect directly to a live broker account via read-only investor credentials and log every trade automatically, so the numbers cannot be edited after the fact. Myfxbook's own verification process confirms that the account is real and the statement has not been tampered with. A developer publishing a verified Myfxbook link (or a signal on the MQL5 Signals service, which uses a similar independent logging model) is giving you something you can check yourself, in real time, rather than asking you to trust a PDF.

When you review the account, look past the headline return and check: how many months of history are there, what is the maximum drawdown, how many trades have closed, and does the equity curve show gaps that suggest the account was reset or swapped. A three-week track record proves almost nothing. Six to twelve months of continuous, unedited live data is a far more meaningful sample.

Worked Example: Backtest Claims vs. Verified Live Results

Suppose a seller advertises a backtest showing 340% annual return with a 12% maximum drawdown over five years of historical data. Compare that to a verified live account showing 28% annual return with an 18% maximum drawdown over eight months. The backtest number looks far more impressive on paper. But the backtest was generated with perfect hindsight on data the strategy was tuned against, while the live account reflects real spreads, real slippage, and real execution happening in front of anyone who wants to check it. In practice, the smaller, verified number is the one you can actually plan around.

3. What Does It Actually Cost, Including Everything the Sales Page Doesn't Show Up Front?

Pricing structures for EAs vary widely, and the sticker price rarely tells the whole story. Ask directly: Is this a one-time purchase or a recurring subscription? Does the license cover both MT4 and MT5, or do you need to buy twice? Are updates included, or is there a separate maintenance fee? Is there a free trial or refund policy, and if not, why not?

Be skeptical of anyone who cannot answer these questions in one clear sentence. A trustworthy developer will state the terms plainly: for example, a one-time $199 lifetime license that covers both MT4 and MT5 with no subscription and no recurring fee, versus a $30/month copy-signal option for traders who would rather mirror trades than run the EA themselves. Both models are legitimate as long as they are disclosed clearly before you pay — what is not legitimate is a vendor who advertises a low headline price and then reveals mandatory add-ons, renewal fees, or "activation" charges only after checkout.

Worked Example: One-Time License vs. Subscription Over Three Years

Consider two hypothetical vendors. Vendor A sells a lifetime license for a flat $199 with no recurring charge. Vendor B sells a similar product for $39/month. Over 36 months, Vendor A's total cost stays at $199. Vendor B's total cost is $39 x 36 = $1,404 — more than seven times as much for a comparable product. Subscriptions are not automatically bad (they can fund ongoing development and support), but you should do this multiplication before you buy, not after your card has been charged for the eighteenth time.

4. Which Markets, Timeframes, and Platforms Does It Actually Trade?

Not every EA is built for every market or account type. Ask exactly which instrument or instruments the EA trades, what timeframe it evaluates, and whether it runs on MetaTrader 4, MetaTrader 5, or both. A system built to trade a single, well-understood instrument on a higher timeframe — for instance, an XAUUSD-only strategy operating on the H4 chart — is generally easier to reason about and stress-test than a "universal" EA that claims to trade dozens of pairs and every timeframe from M1 to daily with the same rule set.

Gold specifically behaves differently from currency pairs, reacting to real interest rates, central-bank buying, and shifting safe-haven demand in ways that a generic multi-pair EA is rarely tuned to handle well. A developer who built and tuned their system specifically around gold's volatility profile should be able to explain why that focus matters, rather than claiming the same logic works equally well on EURUSD, US30, and Bitcoin. If you're still deciding whether gold-focused automation fits your goals at all, it's worth reading through the fundamentals of automated gold trading profitability before you evaluate any specific vendor.

5. How Does the System Size Positions and Manage Risk?

This question separates serious developers from marketers. Ask: How is lot size calculated — fixed, or based on account equity and a defined risk percentage? Is there a stop-loss or safety-stop on every trade? Does the system lock in profit as a trade moves favorably, or does it let winners ride with no protection? What is the realistic range of maximum drawdown you should expect, based on the verified track record rather than the backtest?

A well-built system will size positions based on a percentage of account equity rather than a flat lot size, so that a $2,000 account and a $20,000 account are each risking a proportional, not identical, dollar amount per trade. Many well-designed EAs also offer multiple risk settings — commonly labeled something like Conservative, Normal, and Aggressive — so you can choose a drawdown profile that matches your own tolerance rather than being locked into one setting. For general background on how professional risk sizing works, Investopedia's risk management overview is a useful primer, and understanding what drawdown actually measures will help you interpret whatever numbers the developer shows you. Our own breakdown of how drawdown is calculated and why it matters goes deeper into reading these numbers correctly.

Worked Example: Risk-Based Lot Sizing Across Two Account Sizes

Say a system risks 1% of account equity per trade with a stop distance equivalent to $10 per 0.01 lot. On a $2,000 account, 1% is $20, sized to roughly 0.02 lots. On a $20,000 account, 1% is $200, sized to roughly 0.20 lots. Both traders are taking the same proportional risk even though their position sizes differ by a factor of ten. If a developer instead tells you the EA always trades a fixed 0.10 lot regardless of account size, a $2,000 account would be risking 5% or more per trade — a materially riskier setup that most experienced traders would flag immediately.

6. What Happens Around High-Impact News and During Periods of Low Liquidity?

Gold is especially sensitive to macroeconomic releases — Federal Reserve rate decisions, US inflation data, and geopolitical shocks can all move price sharply within seconds. Ask the developer how the system behaves during these windows: does it avoid opening new trades around major releases, does it widen its stop distance to account for a likely spread spike, or does it trade through news exactly the same way it trades any other hour? None of these answers is automatically wrong, but you need to know which one you are getting, because slippage and temporary spread widening around news events are simply a feature of how markets work, not a flaw unique to any one broker. If you want to understand how economic releases move gold specifically, our guide on how economic news affects gold prices covers this in more depth.

Be cautious of any developer who claims their EA "automatically filters all bad news" or "eliminates slippage" — no automated system can fully eliminate execution risk, and some EAs marketed with these claims simply do not behave the way the sales copy suggests. If a specific feature like this matters to you, ask for evidence of how it performed during a known volatile session rather than accepting the claim at face value.

7. What Support, Updates, and Documentation Come With the Purchase?

Software breaks, brokers change execution conditions, and terminal updates occasionally require configuration changes. Ask: What support channel is available, and what is the realistic response time? Are updates included for the life of the license, or only for a limited window? Is there written documentation covering installation, input settings, and troubleshooting, or will you be left guessing?

A developer who offers a real support channel — whether that's email, Telegram, or WhatsApp — and points you to documentation before you even ask is signaling that they expect to be around after the sale. Our guide to understanding EA input settings is a useful reference for the kind of documentation depth you should expect a legitimate vendor to provide on their own product. You should also confirm the EA is a genuine listing on a reputable channel such as the MQL5 Market, which enforces a baseline technical review before a product can be listed, rather than an undocumented custom platform with no outside oversight.

Red Flags That Signal a Scam or an Unreliable Developer

Some warning signs are consistent across nearly every trading-system scam the CFTC has documented and the patterns the FTC lists under investment scams. Watch for:

• Guarantees of "risk-free" profit, "guaranteed returns," or claims that the system "never loses" — no legitimate trading strategy can honestly promise this, and any developer who does is either lying or does not understand basic market risk.
• Countdown timers, artificially limited "spots," or pressure to buy immediately before you can verify anything.
• Refusal to show any independently verified track record, only marketing screenshots.
• Testimonials with no verifiable source, or reviews that all sound identical in phrasing.
• Vague or evasive answers to specific, reasonable questions about strategy category, cost structure, or risk controls.
• Pressure to pay through untraceable methods, or requests for account passwords rather than read-only investor credentials for verification.

None of these signs alone proves fraud, but two or three together should make you walk away. Conversely, a developer who openly discusses drawdown, declines to promise guaranteed profit, and directs you to a verifiable track record is behaving the way a legitimate seller should.

A Practical Due-Diligence Checklist Before You Buy

Use the table below as a script for your first conversation with any EA developer. Bring it to a live chat, an email exchange, or a sales call, and note how directly each question gets answered.

How to Test the Answers Before You Commit Real Capital

Once a developer's answers pass the questions above, verify them yourself rather than taking the conversation at face value. Run the EA on a demo account for a meaningful stretch and, where the vendor provides source access or a strategy-tester file, run your own backtest to confirm the logic behaves the way it was described — the step-by-step process is the same one covered in our MT4 and MT5 backtesting walkthroughs linked above. If the developer publishes a verified account, learn how to connect your own MT4 account to Myfxbook so you understand exactly how that verification works before you rely on someone else's numbers. Finally, treat any EA purchase as one piece of a broader plan rather than your entire trading capital — the principles in our guide to capital preservation apply just as much to automated systems as to manual trading. You can review Golden Viper EA's own published terms and verified track record on the product page as one example of how these disclosures should look in practice.

A short, honest note before you go further: trading gold or any other instrument with an automated system carries real risk. Past performance, whether from a backtest or a verified live account, does not guarantee future results, and losses are possible with any strategy. Only allocate capital you can genuinely afford to lose, and treat every claim — including the ones in this article — as something to verify independently before you act on it.

Frequently Asked Questions

Do I need to see the EA's exact source code before buying?

No. A developer keeping their exact indicator logic or entry rules proprietary is normal and is not, by itself, a red flag. What matters is that they can describe the strategy category, risk approach, and instrument/timeframe in plain language and back performance claims with independently verified data rather than asking you to trust the code blindly.

What's the difference between a backtest and a verified live track record?

A backtest is generated by running the strategy against historical price data, often after the strategy has already been tuned to that same data, which makes it easy to overstate future performance. A verified live track record, such as one logged through Myfxbook, reflects real trades placed on a real broker account with real spreads and slippage, and cannot be edited after the fact.

Is a one-time license always better than a subscription?

Not automatically, but over any multi-year holding period a flat one-time fee is usually far cheaper than a recurring monthly charge for a comparable product, as shown in the cost table above. Subscriptions can make sense if they fund active ongoing development, but you should compare the three-year total cost before deciding, not just the first month's price.

Should I be suspicious if a developer won't offer a free trial or refund?

Not necessarily on its own. Some legitimate vendors sell a one-time license with no trial or refund but compensate with a fully transparent, independently verified track record you can review before paying — which is arguably more useful than a trial period, since you can see months of real trading history rather than a few days of your own testing. What matters more is whether the developer is transparent about this policy upfront and backs it with verifiable proof.

How much verified trading history should I look for before trusting an EA?

There's no universal minimum, but a few weeks of data tells you very little. Six months or more of continuous, unedited verified results, covering a range of market conditions, gives you a meaningfully better sense of how the system behaves than a short or cherry-picked window.

What does it mean if an EA uses martingale or grid trading?

Martingale and grid approaches increase position size after losing trades in an attempt to recover losses on the next move, which can produce a smooth-looking equity curve for a long stretch and then a severe, sometimes account-ending drawdown during an unusually strong trending move. Asking directly whether a system uses these techniques, and getting a clear answer, is one of the most important questions on this entire list.

Can an EA guarantee it will avoid losing trades?

No. Any developer who claims their system is risk-free, guarantees profit, or "never loses" is making a claim that cannot be honestly true. This type of promise is a hallmark of trading-system fraud, and every legitimate automated strategy carries the possibility of losing trades and periods of drawdown.

Does it matter whether the EA works on both MT4 and MT5?

It can matter depending on your broker and existing setup. Ask whether one license covers both platforms or whether you'd need to purchase separately for each, and confirm which platform version your broker actually supports before committing to either one.

What should I ask about how the EA sizes trades on a small account?

Ask specifically whether lot sizing is fixed or calculated as a percentage of your account equity, and what the minimum recommended account balance is. A system with proportional, equity-based sizing and a published minimum balance recommendation is generally more appropriate for smaller accounts than one that trades a fixed lot size regardless of your capital.

Is it reasonable to ask a developer for their own risk disclosure?

Yes, and you should. A developer who volunteers, without being pressed, that trading involves risk of loss and that past results do not guarantee future performance is demonstrating exactly the kind of honesty this entire vetting process is designed to find.