Tester · MetaTrader 5 EURUSD 1… AI Review | QuantTester
Tester · MetaTrader 5 EURUSD: return -3.0%, DD 8.22%, PF 0.51, 111 trades. Read the review.
Strategy Safety Grade: E (5 / 100) · High Risk Alert
Trading risk analysis result
Strategy Type: Forced Liquidation (Stop Out). Classified as Forced Liquidation (Stop Out), not a viable strategy: the EA accrued a symmetric stream of small wins over 448 days (86 wins averaging $3.68) before a single synchronized 9-position closure at 2026.11.05 23:59 produced a $455.80 loss and drove the account to a -3.04% net result with a 0.51 profit factor; the precomputed veto places it at Grade E.
Darwinex 集中度压力测试 (Top 5% Windfall Dependence Test)
剔除排名前 6 笔（5%）极端盈利单后：毛利占比 70.21%，剩余净利润 $-526.57，剩余 PF 0.15。
Data transparency & Environment
- MT4/MT5: MetaTrader 5
- Symbol: EURUSD
- Timeframe: 1 Hour (H1) 2025.08.
- Net profit: $-304.20
- Return: -3.0%
- Max drawdown: 8.22%
- Total Trades: 111
- Profit Factor (PF): 0.51
Critical Algorithmic Vulnerabilities
- Detected 9 close-at-stop forced liquidations: clear evidence of a blown account / Stop Out
- Return Consistency Defect: The top 5% of closed rows (6 trades) account for 70.21% of gross profit; removing them leaves residual net profit at -$526.57 with a profit factor of 0.15, demonstrating that expectancy depends on extraction of statistical outliers rather than a systematic edge.
- Payoff Asymmetry Fragility: A 77.5% win rate with a 0.15 realized payoff ratio means the average winner ($3.68) is roughly one-seventh of the average loser ($24.84); a single adverse sequence (2026.11.05, 9 simultaneous closures, -$455.80) erased the entire cumulative gain accumulated over 448 days.
- Forced-Liquidation Concentration: All 9 terminal positions closed at 23:59 on the final data date at an identical price of 1.12188 with total exposure of 0.46 lots; this is consistent with a basket being forcibly unwound rather than individually managed stops, which is the main live transmission channel for capital impairment.
- Position-Size Escalation: Lot sizes rose from a 0.02 baseline to 0.12 on the terminal sequence, with an identical pattern on 2026.02.17 (0.08 lots); given the unpaired close-only rows, a pyramid/martingale link is an unverified hypothesis, but the dollar-risk expansion on the final date is observed and material.
Score Breakdown Details
- [Base Score] All valid reports start at 50 points (neutral base)
- -45 pts [Stop-Out Meltdown] Broker close-at-stop/margin call detected, veto capped at 5 pts (Grade E)
- Final Strategy Safety Score: 5 / 100 (Grade E).
Recommended Next Actions
- Request the full MT5 statement with paired entry tickets, directions and open timestamps to verify whether the terminal 9-position cluster was a pyramid or an adverse scale-in.
- Re-run the backtest with the broker's real tick history and explicit commission/swap settings to confirm how much of the 0.51 profit factor is cost-eroded.
- Inspect the EA's iGrid, iIncreaseLotsK (1.5) and iMaxLossPercentage (45) parameters against the observed 0.12-lot escalation to confirm whether the sizing logic is intentional or unintended risk compounding.
- Establish a hard account-level drawdown cap (iGMaxDD currently 0) and a per-position maximum-loss ceiling to prevent recurrence of the $455.80 synchronized unwind.
- If forward validation is pursued, run a Demo forward account for at least 90 days and compare realized profit factor and maximum simultaneous exposure against this backtest.
Q&A About This Report (Quant Due Diligence)
Why did this EA receive Grade E despite a 77.5% win rate and only an 8.22% maximum drawdown?
The grade is driven by the terminal liquidation event rather than the headline win rate. On 2026.11.05 at 23:59, 9 positions closed simultaneously at 1.12188 with a combined loss of $455.80, wiping out the cumulative gain built over 448 days. The realized payoff ratio of 0.15 (average win $3.68 versus average loss $24.84) means the win rate cannot protect capital once losses concentrate.
Was the strategy running a martingale or a grid, and how is that evidence treated?
The parsed rows are close-only and do not supply entry timestamps, directions or ticket pairing, so a confirmed martingale or grid sequence cannot be established from this evidence. What is observable is that lot size stepped from a 0.02 baseline to 0.12 lots in the final sequence and to 0.08 lots on 2026.02.17, and that 9 positions were closed together at 2026.11.05 23:59. The scale-in structure remains an unverified hypothesis pending paired-ticket data.
The maximum drawdown is reported at 8.22%. Does that mean downside risk was contained?
No. Drawdown in the report is measured against the running equity peak and understates the final realized outcome: net return is -3.04% and the profit factor is 0.51. The stop-loss evidence shows an unbounded tail (largest loss $131.87 versus median loss $1.52), meaning realized losses do not terminate at a consistent level. The relevant capital-safety question is not the single peak-to-trough statistic but whether losses are bounded, and here they are not.
How concentrated is profitability, and what does that imply for live allocation?
The top 5% of closed rows (6 trades) account for 70.21% of gross profit. Removing them leaves residual net profit of -$526.57 and a profit factor of 0.15, while the top 1% alone (2 trades) carries 39.35% of gross profit and, if removed, leaves a profit factor of 0.31. The result is a small set of outsized winners funding a long sequence of thin gains, which is exactly the profile that fails when a single adverse sequence (as on 2026.11.05) removes the accumulated buffer.
How is QuantTester different from asking a general-purpose AI assistant myself?
QuantTester parses MT4/MT5 reports into structured metrics and trade records before AI review. Supported calculations can be checked against the original export, guaranteeing consistent audit facts.
Which broker is recommended for trading and running MT4/MT5 strategies?
For algorithmic trading, EA execution and reliable backtesting, we strongly recommend IC Markets (Raw Spread Account) with 0.0 pip spreads.