GoldX9_Base · MetaTrader 5 XAUUSD H1… AI 测评｜QuantTester
GoldX9_Base · MetaTrader 5 XAUUSD H 报告：收益 132.8%、回撤 19.55%、PF 1.99、444 笔交易。AI 风险解读。
策略安全信用评级: A (85 / 100) · 稳健受控 (Robust)
交易风险分析结果
Classified as Hybrid (Tiered Basket + Trend Following): a short-biased XAUUSD basket EA that layers multiple same-direction entries within minutes and exits them collectively, achieving a 1.99 profit factor and 93.2% win rate over 388 days while retaining a 1.43 profit factor after removing its best 5% of trades, but whose 0.14 payoff ratio and 30 heavy losses (up to -1,118.5 on a single position) mean a small number of adverse sequences dictate survival, making controlled 19.55% drawdown the binding real-market boundary rather than return potential.
Darwinex 集中度压力测试 (Top 5% Windfall Dependence Test)
剔除排名前 23 笔（5%）极端盈利单后：毛利占比 28.3%，剩余净利润 $5,717.90，剩余 PF 1.43。
数据透明度与环境说明
- MT4/MT5: MetaTrader 5
- 品种: XAUUSD
- 周期: H1 (2025.01.03 - 202
- 总净利润: $13,279.50
- 总收益率: 132.8%
- 最大回撤: 19.55%
- 总交易笔数: 444
- 盈利因子 (PF): 1.99
AI 识别的核心高危漏洞清单
- Payoff Asymmetry Fragility (La): The realized payoff ratio is 0.14 - average win 64.53 versus average loss -447.93 - so the 93.2% win rate is doing all of the work. The largest single loss of -1,118.5 requires roughly 17.3 average winners to offset; a modest cluster of adverse short-side sequences would flip the arithmetic expectancy negative despite the headline PF of 1.99.
- Risk Stability Breakdown (Rs): Loss severity escalated through the sample. Early losses were -529.6 (2025-01-16); mid-sample clustered at -626.5 to -688.7 (2025-04-04 to 2025-06-17) and -841.5 (2025-11-14); the tail includes -879.9 (2026-01-02), -909.1 (2025-12-29), -927.4 (2026-01-16) and the single maximum of -1,118.5 (2026-01-29). Widening absolute dollar exposure undermines replication stability for an allocator sizing on historical loss distribution.
- Concurrent Exposure Concentration (Capacity/Execution): The audit records up to 9 simultaneously open orders with 0.1 lots each, and the same-minute exits frequently close 5-9 tickets together. Combined basket exposure on a single instrument is roughly 9x the nominal per-trade size; a sustained short-side sweep during a gold rally amplifies correlated mark-to-market loss beyond the per-trade risk model.
- Directional Bias (Return Consistency R+): 98.6% of closed trades are sells, against a 1.4% long share. Profit is structurally contingent on mean-reverting or declining XAUUSD behavior; a persistent one-way upward impulse would be materially more damaging than the 19.55% historical drawdown suggests, and the record contains no regime that severely tested this asymmetry.
- Outlier-Adjacent Dependence (R+): The residual PF of 1.43 after excluding the top 5% is acceptable, but the top 1% of trades (5 rows) already carry 8.11% of gross profit. Combined with the 0.14 payoff ratio, tolerance for missing a small number of favorable exits is narrow.
客观数学打分明细
- [Base Score] All valid statements start at 50 points neutral benchmark (Grade B-)
- +5 pts [Order Basket]: Controlled multi-entry execution in same minute; layered basket architecture
- +5 pts [Expectancy]: Overall Profit Factor reached PF 1.99 with solid positive mathematical expectancy
- +10 pts [Organic Profit]: Removing top 5% windfall deals leaves PF 1.43 >= 1.2; robust and non-luck-driven
- +5 pts [Controlled Drawdown]: Max historical drawdown is 19.6% (10%-20%), within stable controlled zone
- +10 pts [Sample Horizon]: Sample duration spans 388 days (>= 180 days), covering diverse regimes
- Final Strategy Safety Score: 85 / 100 (Grade A).
下一步行动与优化建议
- Obtain a broker-level confirmation of the effective stop mechanism (EA-side close, broker SL, or trailing) by sampling realized stop levels from a full MT4/MT5 statement with populated SL fields.
- Re-run the backtest with realistic spread, commission, and swap for a multi-day-holding XAUUSD basket to quantify net-of-cost expectancy and confirm the 1.99 PF survives friction.
- Stress-test the short-biased basket against a synthetic one-way gold-uptrend series to estimate the true tail drawdown beyond the observed 19.55%.
- Measure maximum simultaneous exposure and margin utilization from the full unfiltered order data, and compare it to the account's stated 0.5% risk-per-trade input.
- Forward-observe on a demo account for at least 60-90 days to verify that basket fill quality and slippage match the backtest assumptions.
关于本报告的关键质询与答疑 (Q&A About This Report)
The strategy achieved a Grade A Safety Grade; what are the core strengths and rationale supporting this rating?
The Grade A rating is backed by robust risk controls in the trade data: strict execution discipline (e.g. pre-set stop losses, synchronized basket entries without toxic martingale averaging), and historical drawdowns kept within institutional thresholds.
The strategy achieved a Grade A Safety Grade; what are the core strengths and rationale supporting this rating?
The Grade A rating is backed by robust risk controls in the trade data: strict execution discipline (e.g. pre-set stop losses, synchronized basket entries without toxic martingale averaging), and historical drawdowns kept within institutional thresholds.
The report shows zero values in the SL and TP columns - does the strategy trade without protection?
The supplied rows carry no populated SL/TP, and realized losses terminate at variable levels (clustered share 6.7%, max-over-median 1.9), so executed stop discipline cannot be confirmed from the loss pattern. The parameter set, however, configures InpStopLossPct 2.0, take-profit 1.0%, trailing, breakeven, and a salvage-exit rule, and there is no stop-out or forced liquidation in the record. The honest position is that protection exists in configuration and no runaway loss is observed, but hard stop execution remains unverified.
Given the 98.6% short bias, what single condition would most seriously challenge the 19.55% drawdown figure?
A sustained upward XAUUSD impulse that forces correlated mark-to-market losses across a multi-order short basket. The observed worst episodes already show the pattern: 2025-10-09 saw concurrent -811.0 losses, 2026-01-16 a -927.4 short, and 2026-01-29 a -1,118.5 short, yet the sample as a whole was net profitable for shorts. Because only 6 of 444 trades are long, the strategy has no demonstrated restoration mechanism if the gold uptrend persists rather than reverts.
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.