This is what a $300 private audit contains, filled in with a real subject from the free batch: file 066, sold as Hurricane Scalper ECN 2.0 PLUS. Nobody paid for this page and nobody was asked. The only difference from a paid report is that you are reading it.
Not a scam, and not a strategy. A martingale grid that earned its entire 32-month profit in shallow water. The deep water — the region that ends accounts — was never sampled in 32 months, and the shipped defaults do everything wrong in exactly the order that pays: 1.5× lot escalation per adverse pip, basket exit at 2.5 pip, and no stop of any kind. Its profit distribution is “small money most days, occasionally zero”, and the zero never happened on this sample. That is history’s gift, not a property of the system.
| Item | Value |
|---|---|
| Symbol / timeframe | EURUSD H1 |
| Window | 2024-02-01 → 2026-09-30 (32 months) |
| Spread | 17 points fixed (1.70 pip) |
| Start / sizing | $10,000 · 0.01 lots |
| Model | Open prices only · MetaTrader 4 build 1492 |
| Net & drawdown source | Printed by the EA itself through monitor code injected into OnTester (MT4-computed balance, equity peak, max drawdown). Not reconstructed from logs. |
| Measure | Value |
|---|---|
| Positions opened | 5,099 |
| Positions closed | 583 |
| Closed / opened ratio | 11.4% |
| Win rate (closed trades only) | 86.1% |
| Lot range | 0.01 → 0.83 |
| Stop-out | none |
An 11.4% close ratio means positions are closed in baskets by a basket-average rule, not one by one. The 86.1% win rate is computed over the 583 trades the basket rule chose to close — the ones that reached the profit side. The other 4,516 positions were still open when the window closed, and their floating P/L of +$2,340 is the whole difference between the two net figures on this page.
Delete the best 5% of closed trades and the ledger goes from +$1,418 to +$803. Read both ways:
By year the closed ledger reads 2024: +$253 · 2025: +$576 · 2026: +$588. Positive every year — which shows the favourable regime persisted, not that the system is robust.
The package ships full .mq4 source (decompiled — variable names are mangled — but the Chinese parameter names survive, so the logic reconstructs exactly). Key inputs, as shipped:
| Input (translated) | Default | What it does |
|---|---|---|
| Lot multiplier | 1.5 | Each adverse step sizes the next order at 1.5× the previous |
| Max orders per side | 25 | The escalation ceiling |
| Adverse add interval | 5 pip −0.1/step | Adds every 5 pip of adverse movement, interval shrinking per order |
| Basket take-profit | 2.5 pip −0.1/step | Closes the whole basket past its average; target shrinks as the basket grows |
| Total-loss close | 0 | Disabled — there is no stop |
| Percent equity stop switch | FALSE | The circuit breaker exists and is off |
| Four signal switches (Bands/RSI/Stoch/CCI) | all FALSE | The entry does not use any indicator by default |
Three lines decide everything. Translated from the source:
Two things the source does to stay unread: the first 10 tickets are virtual — recorded internally, never sent, so the platform’s visible order count understates true exposure by 10 — and the code calls HideTestIndicators(true) so the tester shows nothing it depends on. A strategy sold as transparent has no legitimate use for either.
| Open orders | Next order | Cumul. adverse | Total lots |
|---|---|---|---|
| 1 | 0.010 | 0.0 pip | 0.01 |
| 6 | 0.076 | 23.5 pip | 0.21 |
| 12 | 0.600 (cap) | 48.4 pip | 2.31 |
| 25 | 0.600 | 90.0 pip | 10.11 |
At 90 pip of adverse movement the account holds 10.11 lots against a $10,000 balance. From there, another 100 pip one-way — about two days of a real trend — is a $10,110 floating loss. The account is gone, and no line of the shipped configuration objects. In 32 months the deepest basket resolved before this point, every time. The Kelly framing: the whole expectation hangs on the annual probability of that one event, which 32 months cannot estimate. At a 4% annual chance it is +9.4%/yr; at 15% it is −3.1%/yr. No backtest distinguishes these.
To separate the multiplier’s money from the strategy’s money, the same code was rebuilt with only the extern defaults changed (plus the monitor block) and re-run on the identical window:
| Build | Multiplier | Max orders | Signal | Net | Max DD | Net/DD |
|---|---|---|---|---|---|---|
| As sold | 1.5 | 25 | — | +$3,758 | 18.04% | 1.51 |
| Linear, deep grid | 1.0 | 25 | — | +$557 | 18.48% | 0.28 |
| Linear, shallow grid | 1.0 | 6 | — | −$388 | 7.07% | neg. |
| Linear, shallow + RSI | 1.0 | 6 | RSI 30/70 | +$247 | 1.42% | 1.70 |
Turn the multiplier off and 85% of the profit disappears while the drawdown barely moves. The grid edge alone is worth about 2% a year at an 18% drawdown. The one clean configuration — linear sizing, shallow grid, RSI filter — makes less, but at 1.42% drawdown the ratio is the best of the four. Its caveat: 61 closed trades in 32 months is not a sample, and about $167 of the +$247 is end-of-window floating profit realized at close. It is a candidate, not a verdict.
If the vendor shipped this, the report would read differently:
| Change | Why |
|---|---|
| Lot multiplier 1.5 → 1.0 | The only mandatory change; everything else is secondary |
| Max orders per side 25 → 6 | Depth is the exposure: 6 orders cap adverse excursion at 23.5 pip |
| RSI switch TRUE (30/70) | The only switch that measurably improved entry quality |
| Percent equity stop TRUE at 25% | Turns “occasionally zero” into “occasionally −25%” |
| 90 days demo before any size | Checks that live trade count, slippage and patterns match the tester’s 61 trades / 32 months |
Reconciliation note. The database page shows +$1,418; this page shows +$3,758. Both figures are correct and the difference is itself a finding: +$1,418 is the closed-trade ledger, +$3,758 is the MT4-measured net including the floating P/L of the 4,516 positions still open at window close. A net figure that depends on which positions you count is one more thing a screenshot will never tell you.
Isolated MT4 instance (separate data directory from any live terminal), build 1492 tester, EURUSD H1, 2024-02-01 to 2026-09-30, fixed 17-point spread, $10,000 start, open-prices-only model. Net, equity peak and drawdown printed by the EA through monitor code injected into OnTester and read from the tester output — not rebuilt from logs. Source files are the original package’s .mq4 (UTF-16LE with BOM); modifications were limited to extern default values and the monitor block; nothing in the compiled logic was touched. Subject is free-batch file 066; file 069 is byte-identical output of the same EA and is not double-counted anywhere on this site. No vendor was paid, contacted, or given preview.
The format above — conditions, structure, tail, source, exposure arithmetic, variant runs, recommendation — is the private audit. If you want one for your own build before a buyer runs it for you: the testing service →