Vendor Interview Guide
12 questions · ideal vs red-flag answers · why each matters
Most traders buy EAs without asking a single question. This guide walks you through the 12 questions that separate informed buyers from those who rely on hope. Click each question to see what a good answer looks like — and what a bad one tells you.
What Questions Should You Ask
Before Buying a Gold Trading EA?
The 12 questions to put to any EA vendor before handing over money — what a good answer sounds like, what a red flag sounds like, and why each question matters.
Published 15 June 2026 · Updated as vendor practices evolve
The five non-negotiable questions: (1) What is the named strategy? (2) Can you share the 3-year backtest .htm report? (3) Is there a live forward test link? (4) Does every trade have a hard broker-side stop loss? (5) What is the maximum historical drawdown? If any of these are evaded, stop. The remaining seven questions deepen your vendor assessment once the basics are confirmed.
Why Most Traders Don't Ask — And Why That's a Problem
The default EA buying experience looks like this: trader finds a product, reads the description, sees some impressive screenshots, checks the price, and pays. No questions asked. This is the experience EA vendors optimise their sales pages for — and it is the experience that most often ends in disappointment.
There are three reasons traders don't ask questions. First, they feel awkward — as if asking is an accusation of dishonesty. It is not. It is standard due diligence. Second, they don't know what to ask — which is why this guide exists. Third, they assume that a high price or a professional website signals quality. It signals budget, not substance.
The psychology of EA salespeople is worth understanding. Most are not malicious — they genuinely believe in their product. But belief in a product and evidence of its quality are different things. Your job as a buyer is to convert belief into verifiable evidence. These 12 questions are the mechanism for that conversion.
Most
Traders who ask no questions before buying
3–5
Questions that would catch most EA problems
5 min
Time to send a pre-sales email
The 12 Questions — Click to Expand Each
Each question shows the ideal vendor answer, the red-flag answer, and why the question matters.
Ideal answer
A named, specific strategy: e.g., "M15 range breakout with H4 trend filter." The vendor explains entry conditions, indicators used, and how exits are managed — in plain language, without code disclosure.
Red flag answer
"Proprietary AI algorithm." Any description that avoids naming the strategy type or explaining what actually triggers a trade is hiding something — either a simple strategy oversold as complex, or one with poor fundamentals.
Why this question matters
You cannot evaluate risk without knowing what the strategy does. An unnamed "algorithm" cannot be fact-checked, stress-tested conceptually, or compared to market conditions it may not suit.
Ideal answer
Yes — and they provide the full MT5 Strategy Tester .htm report file, not a screenshot. The test covers at minimum 2022–2025, includes the COVID period if possible, and shows modelling quality above 90%.
Red flag answer
"Results speak for themselves." Or sharing only a single-year backtest on a bull-run period. Screenshots instead of the .htm file are suspect — they can be edited in any image tool.
Why this question matters
Three years of data that includes different market regimes (trending, ranging, volatile) is the minimum to distinguish a robust strategy from a curve-fitted one.
Ideal answer
Yes — with a verifiable link to MyFXBook, FXBlue, or similar. At least 6 months of live or demo-with-real-conditions data. Account is not "unverified" on the platform.
Red flag answer
"We don't share that." Or: providing only a demo account test without real-broker data. Vendors with no forward test are asking you to be the first real-world experiment.
Why this question matters
Backtests use historical data the EA's developer may have optimised against. Forward tests are on data the EA has never seen — the only true measure of out-of-sample performance.
Ideal answer
Yes — a hard stop loss placed at the broker level on every trade. The SL size is stated (e.g., 15–20 pips fixed, or ATR-based). This is confirmed in documentation, not just verbally.
Red flag answer
"Sometimes we let winners run" — this conflates SL with trailing stops. Or: "virtual stop loss managed internally." A virtual SL means the stop exists only in the EA's memory; if the EA crashes or connection drops, the trade is unprotected at the broker level.
Why this question matters
No hard SL is the single biggest risk factor in EA trading. A martingale or grid EA with no SL can lose entire accounts in one volatile session. This question has only one acceptable answer.
Ideal answer
A specific percentage, with context: "Maximum drawdown was 18% during the March 2020 gold volatility spike, with the account recovering within 6 weeks." The figure matches what the backtest report shows.
Red flag answer
"Very low, don't worry." Or a number without context — "only 5%" on a 6-month backtest with low volatility. Suspiciously low drawdown on a short backtest is a red flag, not a green one.
Why this question matters
Drawdown is the key risk metric. Understanding the worst historical period tells you what capital you need to survive it and whether the risk-to-reward profile matches your tolerance.
Ideal answer
No. Or, if it uses position-sizing progression: "Yes, but with a hard account drawdown cap at X% that closes all positions." The vendor is transparent about any position recovery logic.
Red flag answer
"It has a recovery system" without specifying what that system is. Recovery systems are often martingale by another name — doubling position size after a loss. Without a hard account-level stop, these strategies carry unlimited drawdown risk.
Why this question matters
Martingale and grid strategies produce flattering results in range-bound markets and catastrophic losses in trending markets. Gold (XAUUSD) is known for strong, sustained trends — exactly the market condition these strategies fail in.
Ideal answer
Specific broker types: ECN, MT5 platform, spreads on XAUUSD typically under 25 pips. Named broker partners if they exist. Honest caveats about brokers it performs poorly with.
Red flag answer
"All brokers." No EA works identically across all brokers. Spread differences, execution speed, and account type all significantly affect EA performance. Claiming universal compatibility suggests the vendor has not tested across broker conditions.
Why this question matters
Execution quality, spread, and slippage vary enormously between brokers. A backtest at a tight-spread ECN broker will not replicate on a market-maker broker with 40-pip spreads on XAUUSD.
Ideal answer
A specific channel (email, Telegram group, ticketing system) and a stated response time such as "within 24 business hours." The vendor can demonstrate responsiveness with their reply to your pre-sales question.
Red flag answer
"Read the manual." Or: community forum support only, with no direct vendor channel. Last response visible in any public channel was months ago. Support is the first thing to vanish when a vendor moves on from a product.
Why this question matters
MT5 updates, broker changes, and evolving market conditions all require EA updates and configuration guidance. Support after purchase is part of the product, not a bonus.
Ideal answer
A changelog with dates showing updates within the last 90 days, with notes explaining what changed and why (e.g., "adjusted spread filter for current broker environment"). Ideally available publicly on the MQL5 product page.
Red flag answer
"We constantly improve it" with no documentation. Or: last visible update was 18+ months ago. An EA left without updates in a changing market is an EA that will silently underperform without anyone noticing.
Why this question matters
Gold market conditions evolve. Broker execution characteristics change. MT5 itself receives updates. An EA vendor who cannot point to a recent changelog is either not actively developing or not communicating.
Ideal answer
Yes — the vendor actively encourages demo testing and ideally provides a recommended demo testing period (e.g., "run on demo for at least 4 weeks before going live with real capital").
Red flag answer
"Demo isn't realistic, just go live." While demo spreads can differ, this answer is a pressure tactic. Any EA that cannot be tested on demo should trigger caution — demo testing reveals setup issues, parameter mismatches, and strategy behaviour before real capital is at risk.
Why this question matters
Demo testing is your lowest-cost opportunity to verify that the EA you received matches the one described in documentation. Vendors who discourage it have a reason to discourage it.
Ideal answer
A specific policy: "14-day refund if the EA does not function as documented on a supported MT5 broker." Or: a clear explanation of why refunds are not offered — such as digital delivery preventing re-sale once delivered.
Red flag answer
"All sales final" with no reason. Any vendor offering zero refund path on a digital product costing $100+ without explanation has no incentive to ensure the product works as advertised after purchase.
Why this question matters
Refund policy reveals vendor confidence. A vendor who stands behind their product offers a path to resolution. A flat "no refunds" policy with no reasoning shifts all risk to the buyer.
Ideal answer
A specific number: "3 live accounts + 1 demo account." Or: a MetaQuotes ID-based licence that travels with your login regardless of broker. Clear process for changing the installed broker if needed.
Red flag answer
A vague answer: "it depends" or "contact us when you need to change." Licence ambiguity almost always resolves against the buyer. If you need to switch brokers mid-trade and cannot install on the new account, you are stuck.
Why this question matters
You may need to switch brokers — conditions change, brokers close, execution quality deteriorates. A restrictive licence can strand you in a poor situation. Know the terms before you pay.
Reading Vendor Behaviour — Beyond the Answers
The answers to these questions matter. But how a vendor responds — their tone, speed, and attitude — is equally revealing.
Confident vendor signals
- Responds within 24 hours with detailed, specific answers
- Shares backtest files proactively — often before you ask
- Acknowledges limitations: "This strategy underperforms in ranging markets"
- References documentation and changelog in their replies
- Encourages demo testing and provides a recommended duration
- Does not create urgency or pressure around your timeline to buy
Defensive vendor signals
- Replies take days, or answers are one-line deflections
- Testimonials and screenshots instead of data when asked for data
- Becomes defensive or vague when stop loss questions are raised
- Creates artificial urgency: "This offer expires soon"
- Suggests demo testing is not representative — without explanation
- Cannot point to a recent update or changelog when asked
As a practical reference point: Goldie Razor V2.8.4 documents its strategy (M15 range breakout with H4 200 EMA trend filter), confirms hard stop loss placement on every trade, and provides complete setup documentation. These are the answers you should be looking for from any vendor you contact.
Further Reading
Complete your pre-purchase research with these related guides.
What Documentation Should You Demand Before Buying a Gold EA?
The specific documents to request from vendors — a companion to this question guide.
What Should You Check Before Buying a Gold Trading EA?
What to verify independently — the self-directed side of due diligence.
Best Expert Advisor for XAUUSD in 2026
EA evaluation scorecard once you have completed your vendor interrogation.
How Much Should You Spend on a Gold EA?
Understanding vendor cost structures helps calibrate expectations about documentation and support.
Who Makes the Best Gold EA?
Evaluating vendors as a whole, not just the specific product.
Frequently Asked Questions
Treat evasion as information. If a vendor cannot or will not answer a direct question about their strategy type, stop loss mechanism, or backtest source data, that non-answer tells you something important: either the answer is poor, or they are not accustomed to being held accountable. A confident vendor with a quality product answers these questions readily — often before they are even asked. Politely restate the question once. If it is dodged again, that is your answer.
Yes — and professional EA vendors expect it from serious buyers. Pre-sales due diligence is standard in any high-value purchase. The vendor's attitude toward your questions is itself a data point: vendors who are irritated by due diligence questions prefer uninformed buyers. Vendors who welcome them have confidence in what they are selling. The quality and speed of pre-sales responses often predicts the quality of post-sales support.
Source code protection is legitimate — you are not entitled to see the .mq5 source code. However, you are entitled to know: (1) what strategy type the EA uses (breakout, mean reversion, trend-following), (2) whether it places hard stop losses on every trade, and (3) what the backtest statistics show. None of these require code disclosure. If a vendor refuses to share these three things under the guise of "IP protection," they are using IP protection as cover for something else.
No. Backtest data requires no code disclosure — it is output data from running a compiled .ex5 file on historical prices. A vendor who refuses to share it either has poor backtest results they do not want you to see, or has not run one properly. Either reason disqualifies the product. The only partial exception is a vendor who has so much verified live-account data that backtest results become secondary, but even then, refusing to share them is unusual for a professional operation.
A confident vendor shares documentation proactively, often before you ask. Their support response is detailed, references the specific question, and cites documentation or data. They acknowledge limitations honestly — "our backtest shows 2022 was a difficult year for this strategy" — rather than overselling. A defensive vendor deflects specific questions with generalities ("the EA has performed excellently"), cites testimonials instead of data, and creates urgency to buy before you ask more questions.
Goldie Razor V2.8.4
M15 breakout + H4 EMA filter — built for XAUUSD on MT5