The Gold Reaper EA Review (2026): 102 Weeks of Live Signal Data, Scored on 8 Criteria

59.3 / 100

weighted verdict score on eight published criteria

102 weeks

live MQL5 signal record read on 17 Oct 2026, 895 trades

7% to 96%

stressed drawdown from $8,000 down to the $600 minimum, same EA, per an independent test

The verdict at a glance: a real record, a fragile minimum deposit

A verdict is a set of trade-offs, so here is the shape of this one before any detail. The Gold Reaper holds up where the last verdict did not: it has a long, platform-hosted, public record. It loses points where most breakout systems do: a modest payoff per trade, lumpy returns, and a drawdown that explodes on a small account. The table and radar show how the 59.3 was built.

What The Gold Reaper actually is

The Gold Reaper MT5 is an Expert Advisor for MetaTrader 5 sold by Profalgo Limited of Malta, run by developer Wim Schrynemakers. The seller profile shows a 4.1 rating from 1,010 votes, 27 products and 29 signals. The Market listing describes a breakout system for gold: it places pending orders beyond support and resistance levels it builds, across several timeframes at once, and the setup guide recommends the H1 chart. It was published on 1 February 2024 and showed version 5.4 on 14 October 2026.

It is not one strategy but several. An independent reviewer counted eight (daily, four-hour and hourly levels, by his own reading), while the NYCServers review says nine. Neither figure comes from an official table, which is a small but telling point about how much of the engine is visible from outside. The vendor's own pitch is that several internal strategies spread risk, and that randomisation of entries, exits and trailing values stops two accounts at one broker from trading identically.

How we judged it, and who has something to sell

Evidence on EAs is never neutral, so here is every party and its stake, ours included.

• The vendor sells the EA and runs the signal. His record is public and checkable, which is the point, but he also has every reason to run the signal well.
• ForexCopilotEA sells a listing called Gold Reaper. We gain if you buy it, which is why this verdict opens with that fact and why it applies the same rules we applied to Quantum Queen X.
• Diego Arribas Lopez wrote the independent MQL5 stress test. He says he has no tie to the developer, but he builds and sells competing EAs and the article carries Axi broker affiliate links.
• NYCServers reviewed it and sells the VPS hosting its pages promote. Its page does not carry a date we could read, so we treat its figures as undated.
• Buyers and signal subscribers wrote the reviews. Testimony, not data.

Grades C and D are reported but never scored as proof. A grade A record lifts the cap on the evidence criterion, and it has done so here, which is the biggest difference between this verdict and the last one.

102 weeks on a live MQL5 signal: what the record shows

The vendor's live signal, named Gold Reaper New V2 2, has been trading XAUUSD since 22 October 2024 on a CapitalPointTrading MT5 account at 1:500 leverage. At 35 dollars a month it has 68 subscribers. Because MQL5 generates the statistics from the broker's trade history, they are the closest thing this market has to an audited record. Here are the figures as we read them on 17 October 2026.

Read that table slowly, because two numbers in it pull in opposite directions. A profit factor of 2.33 and a recovery factor above 11 describe a system that has paid well for the risk it has run on this account. An average loss bigger than the average win describes one with little room for a bad patch. Both are true at once, and the next sections explain why.

The warning MQL5 itself prints

The signal page carries an automated notice that 80% of the growth came within 31 days, 4.36% of the signal's 711-day life. Earlier snapshots in its log show the same pattern: 29 days on 15 June, 26 days on 23 April, with notices saying the share of days for 80% of growth is 'too low'. In plain terms, most of the return arrived in a few bursts, and long stretches in between were flat or negative. The independent reviewer recorded +309% on 27 September; we read +353.71% on 17 October. Roughly 45 percentage points in three weeks is exactly what a burst looks like.

The signal also has gaps worth stating plainly. Its account tab labels it 'Running Latest Version of the EA' but shows no version number, and the independent reviewer, on 27 September, found an older version listed and noted its numbers are not comparable to a Strategy Tester run of the Market product. We therefore treat 'latest version' as a vendor statement we cannot check. The account type, real or demo, is not shown, and there is no month-by-month table or open-position view without a login. And the page's own warning log records 'no trading activity' notices from April to October 2026, including six idle days on 15 October. None of that is a defect in a breakout EA, which can sit still for days. It is a reason to judge the EA over months, not weeks.

How a breakout system makes money, and how it loses it

To judge The Gold Reaper you need to understand the family it belongs to, because its risks are the family's risks. A breakout system waits for price to leave a range and joins the move. The Gold Reaper does it with pending orders placed beyond levels it calculates, with a stop loss and a take profit on each trade. The same idea is explained from first principles in our VWAP breakout research; the picture below shows the single most important thing about it.

Where a grid wins small and loses big, a breakout system tends to lose small and win in bursts. That is a healthier shape, because each loss is capped by a stop placed before the trade, and the signal's own numbers match it: a worst consecutive loss of about $460 across five trades against a balance of several thousand. It is also why the equity curve looks like a staircase with long landings, not a ramp.

The cost of the shape is clustering. Fakeouts do not arrive one at a time; they arrive in choppy weeks. A reviewer on the Market page puts it bluntly, calling it a breakout system with frequent fakeouts that cause short-term losses. Another reports early losses on positions filled across a weekend gap, to which the developer replied that the Friday Stop Hour setting prevents it. Both complaints are structural, not scandalous: a breakout system is exposed to the exact moments breakouts fail, and the settings that manage that exposure are the buyer's to choose.

The overlap finding that matters most

The independent stress test found something a headline win rate hides: the sub-strategies are not independent. One strategy lost on 186 days, and another lost on 134 of the same days; a different pair overlapped on 19 of 38. About a third of the losses in the test's worst stress window came from a single strategy. Remove that strategy and the historical drawdown fell from 8.46% to 5.6%, at a cost of about $1,500 in profit. The practical lesson: eight strategies are not eight lottery tickets. They share days, so when one has a bad day several often do.

A 72% win rate with a payoff below one: the maths of living on your win rate

Look again at the signal's averages: $14.49 per winning trade, $16.10 per losing trade. Divide the average loss by the sum of both and you get the win rate a system needs just to break even: 16.10 / (14.49 + 16.10), which is 52.6%. The signal's actual win rate is 72.17%. That gap of about 20 percentage points is the whole cushion between a profitable record and a flat one.

Expectancy, the average profit per trade, follows: 0.7217 times 14.49, minus 0.2783 times 16.10, equals about $5.98, which is exactly the expected payoff MQL5 prints. Try it with your own figures below, and drop the win rate by ten points to see how quickly a payoff below one stops being a comfortable edge.

Why this matters beyond arithmetic: the NYCServers review lists a Myfxbook account for The Gold Reaper with a win rate of 72%, an average win of $20.94 and an average loss of $51.05. Breakeven there is 70.9%, so a 72% win rate left roughly one point of cushion, and the page reports a profit factor of 1.08 and a maximum drawdown of 41.66%. Same win rate as the signal, a very different payoff, a very different outcome. That page gives no date, no account link and no number of weeks, so we cannot say which version or risk setting produced it, and we grade it C. What we can say is that the same EA can show a profit factor of 2.33 or 1.08 depending on settings, broker and period, and that the win rate alone tells you almost nothing.

Drawdown depends on how much money you put in

This is the section that decides whether The Gold Reaper suits you. The EA sizes its lots from the account balance and a maximum-allowed-drawdown input. The independent reviewer's key finding is that this is a lot-sizing input, not a cap on losses: it tells the EA how big to trade, not when to stop. And at the minimum lot size the broker allows, a small account cannot trade small enough.

Read that chart as the most important sentence in this verdict. The vendor's listed minimum deposit is $600; the independent stress test puts the 95th-percentile drawdown at the minimum lot size on a $600 account at 95.6%. The test measures drawdown between daily closes as a percentage of starting capital, uses 1-minute OHLC data and fixed 0.01 lots per strategy, so it is a model and not a promise. But it is the only independent attempt we found to answer 'what happens to a small account', and the answer is unambiguous: the minimum is a technical floor, not a safety level.

The signal's own drawdown, 16.89% relative by balance, sits between the $4,000 and $2,000 rows, which fits a signal account of a few thousand dollars. Our guides on drawdown and lot size and risk explain why the account you run it on matters more than the EA.

A drawdown has to be earned back, and the maths is not symmetrical. Lose 29% (the $2,000 row) and you need a 40% gain to recover; lose 57% and you need 133%; lose 96% and you need more than 2,000%. The comfortable rows are comfortable partly because they keep you out of that region.

The independent test, and its limits

The MQL5 blog test by Diego Arribas Lopez, published 14 October 2026, is the most thorough outside look we found, and its headline numbers are worth knowing. On Axi data from February 2020 to September 2026, using 1-minute OHLC modelling with a $4,000 reference deposit and fixed 0.01 lots, all eight strategies together earned $9,578.50. Fifty-three of 56 strategy-years were positive (2020 and 2026 are partial years). The worst historical drawdown was 8.46%, about $339. On a real-tick cross-check from November 2025 to April 2026, the win rate fell from 78.26% to 71.38% and money drawdown rose from $219.57 to $243.74: worse, without collapsing.

Its Monte Carlo stress test ran 1,000 scenarios built from 10-day blocks, with all strategies kept in sync, 10% of trades skipped at random and results scaled between 0.85 and 1.15 times. In 664 of the 1,000 scenarios the drawdown was worse than the historical 8.46%. The 95th percentile was 14.34% and the 99th was 17.89%. The author's conclusion is to plan on about 14%, not 8.46%, and he scores the EA 8 out of 10 with one big asterisk for small accounts. He would not run it on one.

Now the limits, because they are real. The test is a model run on history, not a live record. Spread, leverage and trade count are not stated, commission is zero in the reports, and Auto trade-frequency mode was not tested. The author sells competing EAs and uses broker affiliate links, which he discloses. The strategy timeframes are his own inference. For a general framework to check any such test yourself, see our guides to reading a backtest report and to whether backtests can be manipulated. Nothing here contradicts the live signal; it complements it. But two models and one live account are still a small sample of the market.

What buyers and signal subscribers say

The Market listing shows 4.49 from 117 reviews and 334 comments. Unlike the last verdict, the split here is closer to a normal distribution of experience: mostly satisfied buyers, a few unhappy ones, and a developer who replies with specifics. The dated reviews we could read tell a consistent story.

• Positive: a buyer reports passing a funded account and six months of payouts with low drawdown (14 September); another calls it efficient and easy for a beginner to set up (15 October); another reports a gain over about two and a half months, noting multi-day gaps then bursts of profit.
• Negative: a one-star review from 26 September reports losses after a month. The developer replies that the buyer saw about 5% drawdown on default settings while backtests allow up to about 30%, and recommends visual backtests.
• Structural: early losses from weekend-gap fills (answered with the Friday Stop Hour setting) and a reminder that breakouts bring frequent fakeouts.
• Setup friction: the AutoGMT news filter needs worldtimeserver.com on the terminal's allowed URL list, and at least one buyer asked about the prompt.

The signal page has its own 16 reviews and a 4.1 average, and they are a more useful warning than the product reviews. Several are not about the strategy at all; they are about copy trading. Subscribers report that trades were not copied, that broker symbol suffixes broke the mapping, that a $1,000 follower account against a leader balance of about $5,000 produced no copies, and that slippage turned the shown results into lower profits. One long-term reviewer compared it with five other gold signals and noted it was the only one with several consecutive losing days. A signal's numbers describe the signal account. A copier's results depend on their broker, deposit and latency.

Where the sources disagree

Three things stand out. First, the price: NYCServers shows $599 and a $300 minimum balance, the listing now shows $949 and $600. The listing's own copy talks of a final price of $990, so this is a rising-price mechanism like the last verdict's, only faster and without a published staircase. Second, the strategy count differs by one. Third, and most useful, the drawdown estimates range from 8.46% (backtest) to 16.89% (signal) to 41.66% (an undated Myfxbook account at moderate risk). They are not contradictions. They are the same EA run at different sizes, brokers and periods, which is exactly the point of the drawdown section.

The real price of The Gold Reaper

The sticker price on the Market is $949 to own, or $399, $499 and $699 to rent for three months, six months and a year. Renting is the sensible way to run a verification period: three months on your own broker costs well under half the purchase price, long enough to see a fakeout week, and it ends by itself. If the demo and the rental both pass your checks, buying is a separate decision.

Add the costs the sticker leaves out. A VPS is recommended for a 24/5 breakout system with pending orders. Slippage and the broker matter more than usual: the signal's own page lists slippage by broker, with most brokers near zero and a few not. And capital: on the evidence above, the account that makes the EA behave is $2,000 to $4,000 or more, which dwarfs the licence fee. The signal subscription at $35 a month looks like the cheap alternative, but as the buyers section shows, copy trading brings its own drag. Our store's listing is a separate sale; read its product page for its own version, terms and price.

Can your account actually run it?

• Capital. Plan around $2,000 as a floor for a 29% stressed drawdown and $4,000 for around 14%. The $600 minimum is technical.
• Broker. Tight spreads and clean pending-order execution matter. See our guide to spread, slippage and swap and to running EAs on gold.
• Hosting. A VPS keeps the EA running through the pending-order window. See the VPS guide.
• Prop firms. The listing calls it prop-firm ready and ships a set file, and the daily drawdown limit can close everything before a firm's limit. Rules vary and change; read how to use EAs with prop firms and check the firm's current terms before you pay a challenge fee.
• Weekends. Set the Friday Stop Hour. A weekend gap through a pending order is one of the few ways a stop-protected breakout system still takes a bigger loss than planned.

Will it last? Gold, breakouts and a moving target

Breakout systems need movement. Gold in a long, directional phase is their best friend; gold in a tight, choppy range is their worst. The 53-of-56 positive strategy-years across 2020 to 2026 suggest the edge has crossed several regimes, which is rare among gold EAs and counts in its favour. So does the product's age (published February 2024), the vendor's long catalogue of products and signals, and a release cadence that put version 5.4 on the listing on 14 October 2026.

Against it: the strategies overlap, so diversification is weaker than the count of eight implies; the record's best returns are concentrated in short bursts; and a constantly updated product means the version you buy may differ from the one on the signal. Our durability score reflects that balance: better than the last verdict, below a product with a multi-year, multi-regime verified record.

Criterion by criterion: how the 59.3 was built

Each score below is out of 10, multiplied by the criterion's weight, and sits inside the tier described by the written scoring anchors. The reasoning is short on purpose; the sections above carry the evidence.

Who it suits, and who should walk away

It may suit a trader with $2,000 to $4,000 or more earmarked for one gold strategy, who accepts breakout behaviour (small losses, bursts of profit, quiet weeks), who will rent or demo first, and who runs it on their own broker with a VPS and a Friday stop. The independent reviewer's own advice is that it works best as one gold component in a diversified portfolio, not as a whole-account solution.

It does not suit a trader trying to grow a $500 to $1,000 account, anyone who wants smooth monthly returns, anyone planning to copy the signal and expect the signal's numbers, and anyone who judges the EA by one month. If that is you, read our guide to spotting EA scams and forward testing on demo first.

If you want to test it: a 12-week plan

A breakout system needs more patience to test than a grid, because its edge shows up across choppy and trending stretches. Twelve weeks is a minimum, and the rental option is built for exactly this.

1. 1

   Step 1: Backtest the demo, with default settings and with a changed maximum drawdown

   Use XAUUSD with real ticks and your broker's spread. Note the maximum equity drawdown at your intended account size, then change the maximum-allowed-drawdown input and watch how drawdown moves.

2. 2

   Step 2: Run it forward on a demo account for 12 weeks

   Same broker, same account type, on a VPS. Log slippage, fills across the Friday close and the gap of any weekend.

3. 3

   Step 3: Compare your demo with the signal

   Open the signal and compare trade counts, average win and loss and drawdown. A large gap points to your broker, settings or version, and is worth knowing before real money.

4. 4

   Step 4: Size from the drawdown, not the deposit

   Choose the stressed drawdown you could accept and set the lot-sizing and daily drawdown limit to match, using the account-size chart above as a guide.

5. 5

   Step 5: Set a total-loss rule outside the EA

   A fixed percentage of the account, enforced by an account-level stop or a manual rule.

6. 6

   Step 6: Rent before you buy

   If the demo passes, a three-month rental on a small live account extends the test at limited cost.

What would move this score

1. A signal or Myfxbook account running the current Market version on a regulated broker for 12 months or more, with growth spread more evenly across months. This is the largest single gain available.
2. A published, official strategy table with each sub-strategy's timeframe and role. This raises trust and clarifies the overlap question.
3. Official guidance on minimum capital that matches the stress-test results, rather than a $600 floor.
4. A second independent stress test by someone with no competing product.
5. Downward risks: a deeper drawdown on the live signal, a signal gap or removal, or a pattern of unresolved reports from buyers.

The vendor, or any reader holding verified records for the current build, can send them through the contact page. We will recalculate and log the change with the date at the top of this page.

How it compares with the last verdict

Set it beside our Quantum Queen X verdict, which scored 36.5 with a High risk label. The difference is not the strategy type alone. It is that The Gold Reaper has a public, platform-hosted record, a stop on every trade and a developer who shows his inputs, so the evidence criterion moves from 3 to 6 and the risk label from High to Medium. A grid and a breakout system can both be good, but the breakout system's losses are bounded before the trade, and the grid's are not. For a framework on choosing between them, see our strategy comparison and our gold category.