The Gold Reaper review: 80% of the profit came from 31 days

2026-08-13 · based on our full audit of v4.6. Read the complete audit free: it is one of our three open samples.

The Gold Reaper produced the strongest tester run in our catalog: +$271,579 net over 6.6 years from a $10,000 start, profit factor 1.92 in the report head, 4,258 trades, and a cost share that stays at a healthy 16.5% of gross profit. If we ranked EAs, this one would sit at the top of our list. We do not rank, we dissect. And the dissection finds one number that changes how you should read all the others: 80% of the total profit was earned in 31 trading days. That is 3.5% of the 882 days on which the EA traded.

What we ran

Standard catalog protocol: 2020-01 to 2026-08 on XAUUSD, M1 chart, every tick based on real ticks, 10 ms execution delay, $10,000 at 1:100 (the vendor publishes no deposit recommendation, so the protocol default applies), the tester journal paired deal by deal. The run crossed midnight and produced two journal files; our engine merges them and reconstructs the full sequence. The complete deliverable is open as a free sample.

First, credit where the evidence supports it

Two things distinguish this run from the gold EAs we usually audit. The stacking goes with the market: of 3,392 stacked entries, 2,164 added in the direction of the move and 494 against it, with a median size-escalation factor of 0.92. That is pyramiding, not averaging a losing basket. And the cost profile is sane: swap and commission together take 16.5% of gross, compared with the 114% we measured on Quantum Queen X. Whoever built this made structural choices that survive our forensics. The audit still carries a stacking caution, because up to 16 simultaneous positions is one large bet however you stack it, but the geometry is the healthier kind.

The calendar problem

Now the number that reframes everything. Sort the trading days by contribution and you find 80% of the profit sitting in 31 days out of 882. The average day tells you almost nothing about this system. Its economics live in a few short bursts, presumably the big gold trend legs, and the rest of the time it grinds sideways or gives some back.

Three practical consequences follow. First, whoever starts today should expect the everyday version, unspectacular stretches with occasional losing phases, not the average curve from the sales page. The compounding you see in a 6.6-year backtest includes the bursts; your first months probably will not. Second, miss the bursts and the edge is gone. A vacation, a VPS outage, a nervous pause after a losing week: on this profile, being offline in the wrong 31 days costs most of the result. Third, consistency rules at prop firms punish exactly this shape. Our replay flags consistency breaches in two separate years for the IQ Capital profile. A strategy that earns in bursts collides with rules that demand evenness.

Sharpe 4.1 in the header, 1.86 at end of day

The report head prints a Sharpe ratio of 4.14. Our end-of-day recomputation, which is the resolution a human account holder actually experiences across days, lands at 1.86 with an annualized volatility of 28% and a maximum end-of-day drawdown of 22%. Still respectable, and honestly among the better values we have measured. But the header number is what the marketing quotes, and it is more than twice the lived one. The gap is not fraud, it is aggregation. It is also why we recompute instead of copying.

What the prop replay says

Run this EA at full tested size against prop-firm rule sets and it dies: 767 trades exceed the per-position loss limit of the IQ Capital profile at full size, and 410 still do at one-third size. The first death-free sizing is one quarter of the tested size, which leaves roughly 13% a year of withdrawal. That is a perfectly real number, and it is nowhere near what a +$271k backtest suggests at first glance. The full sizing matrix, the challenge pass curve and the withdrawal replay are in the audit.

Run these checks yourself

The deal-level version of these checks, with the exact sizing matrix and the pass curve, is in the free sample audit. Your own tester report goes through the same engine: the browser check is free, the full audit is $19.

The honest limits. One tester run, protocol-default settings of v4.6, 39% of the window on real ticks, in-sample by definition. A strong backtest is not proof of a live edge; concentration is not proof of failure. What the numbers support is narrower and more useful: this system's result depends on a small number of days, and any plan that cannot sit through the quiet stretches, or that must obey consistency rules, is running a different strategy than the one in this report.

Related: the audit page for this EA · all audits · more articles