Vigorous EA · MT4 EURUSD H1 Backtest AI 测评｜QuantTester
Vigorous EA · MT4 EURUSD H1 Backtes 报告：收益 -3.0%、回撤 4.98%、PF 0.51、111 笔交易。AI 风险解读。
策略安全信用评级: E (5 / 100) · 高危预警 (High Risk)
交易风险分析结果
Strategy Type: High-Risk Grid / Blown Account. Classified as High-Risk Grid / Blown Account: over 2025-08-14 to 2026-11-05 (448 days) on EURUSD H1, Vigorous EA produced a net loss of -$304.20 (PF 0.51, expectancy -$2.74/trade, payoff 0.15), with the final bar recording nine simultaneous close-at-stop liquidations spanning 0.02-0.12 lots that erased a 1,151.6 peak gain down to a 9,695.8 balance, demonstrating an unviable, structurally fragile averaging-down architecture whose entire P/L profile depends on statistical outliers.
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。
数据透明度与环境说明
- MT4/MT5: MT4
- 品种: EURUSD
- 周期: H1
- 总净利润: $-304.21
- 总收益率: -3.0%
- 最大回撤: 4.98%
- 总交易笔数: 111
- 盈利因子 (PF): 0.51
AI 识别的核心高危漏洞清单
- Detected 9 close-at-stop forced liquidations: clear evidence of a blown account / Stop Out
- [Risk Stability Breakdown / Rs] Loss Escalation & Forced Liquidation: On 2026-11-05, 9 close-at-stop rows included -$130.16, -$131.87, -$89.24 and -$90.98 on lots up to 0.12, driving the balance from a 10,151.6 peak to 9,695.8; the inability to cut risk before stop-out exposes live capital to catastrophic loss and violates replication stability.
- [Momentum Decay / D-Score Quote Shape] Protracted Stagnation & Alpha Decay: The equity curve gained only ~1.5% cumulatively over 448 days with a 77.48% win rate but PF 0.51; the terminal stop-out erased nearly all accumulated gains, producing a broken quote shape with no durable risk-adjusted trajectory.
- [Outlier Dependency / R+] Statistical Outlier Reliance: The top 5% of trades (6 rows) account for 70.21% of gross profit; excluding them leaves residual net profit at -$526.57 with PF 0.15, demonstrating expectancy is driven by a handful of favourable closures rather than systematic edge.
- [Payoff Asymmetry Fragility / La] Win-Rate & Payoff Asymmetry: 77.48% win rate paired with average win $3.68 versus average loss -$24.84 (payoff 0.15); a single 0.15-payoff loss outruns dozens of small winners, so any interruption in the win streak cascades into arithmetic insolvency.
- [Capacity & Friction / Execution Feasibility] Low-Timeframe Friction Sensitivity: Backtest uses Control-point modelling with fixed slippage 10 and no hard SL configured; lacking live broker execution proof, the thin per-trade margins (~$3.68 average win) cannot be assumed to survive real spread widening and slippage.
客观数学打分明细
- 【Baseline】Starts at 50 pts neutral baseline
- -45 pts / Forced Liquidation (Stop Out): Detected 9 close-at-stop/margin call liquidations, triggering hard safety floor Grade E
- Final Strategy Safety Score: 5 pts (Grade E).
下一步行动与优化建议
- Re-run the backtest on the exact intended chart history after resolving the 26 mismatched charts errors, using the 'Every tick' model, and report the resulting modelling-quality figure.
- Re-export with paired entry/exit ticket data, direction and entry price to confirm tiered-basket structure and true stop-loss behaviour.
- Introduce explicit pre-set hard stop-losses on every position and cap aggregate basket exposure to prevent the 2026-11-05 forced-liquidation scenario.
- Conduct stress-testing of the tiered basket under adverse 1% and 5% drawdown scenarios to quantify worst-case capital loss.
- Validate on out-of-sample data and a Demo/forward record before any capital allocation.
关于本报告的关键质询与答疑 (Q&A About This Report)
Why does the strategy receive Grade E despite a 77.48% win rate?
Because the score is governed by downside risk, not win frequency. The realized payoff ratio is 0.15 (average win $3.68 vs average loss -$24.84), expectancy is -$2.74 per trade, and the terminal bar records 9 close-at-stop liquidations that wiped the balance from a 10,151.6 peak to 9,695.8. The forced-liquidation event caps the grade at E.
Is the profit concentrated rather than systematically generated?
Yes. The top 5% of trades (6 rows) account for 70.21% of gross profit; excluding them leaves residual net profit at -$526.57 with PF 0.15. This demonstrates the account relies on a few outlier closures rather than a repeatable edge.
What does the 2026-11-05 terminal bar reveal?
Nine simultaneous close-at-stop rows at the same timestamp and price (1.12188) with lots from 0.02 to 0.12, including losses of -$130.16, -$131.87, -$89.24 and -$90.98. This is the profile of a tiered basket that was not protected by a pre-set hard stop and was ultimately liquidated.
How reliable is this backtest evidence?
Limited. The tester explicitly warns that results must not be considered; modelling is 'Control points (a very crude method)', modelling quality is n/a, and 26 mismatched charts errors are logged. These test-methodology limitations reduce evidence confidence but are separate from the strategy's internal risk metrics.
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.