TitanForge v1.3_fix · MT5 XAUUSD M30… AI 测评｜QuantTester
TitanForge v1.3_fix · MT5 XAUUSD M3 报告：收益 1,730.6%、回撤 25.52%、PF 3.07、362 笔交易。AI 风险解读。
策略安全信用评级: B (60 / 100) · 常态中性 (Moderate)
交易风险分析结果
Strategy Type: Strict Lot Discipline. Classified as Strict Lot Discipline: A high-win-rate XAUUSD strategy with dynamic lot sizing proportional to equity demonstrated a robust PF of 3.07 over 672 days, but its viability is compromised by a realized payoff ratio of 0.08, where just nine large losses (average -$9,306) nearly offset the cumulative effect of 353 small profits.
Darwinex 集中度压力测试 (Top 5% Windfall Dependence Test)
剔除排名前 19 笔（5%）极端盈利单后：毛利占比 28.13%，剩余净利润 $100,820.63，剩余 PF 2.20。
数据透明度与环境说明
- MT4/MT5: MT5
- 品种: XAUUSD
- 周期: M30
- 总净利润: $173,056.19
- 总收益率: 1,730.6%
- 最大回撤: 25.52%
- 总交易笔数: 362
- 盈利因子 (PF): 3.07
AI 识别的核心高危漏洞清单
- Payoff Asymmetry Fragility: With a 97.51% win rate but a payoff ratio of 0.08, average wins ($727.5) are dwarfed by average losses ($9,305.78). The nine recorded losses, ranging from -$5,377 to -$14,877, collectively erased $83,752, equivalent to 65% of gross profit. This structure is mathematically fragile: a modest increase in the frequency of large losses would rapidly deplete capital.
- Loss Scale Escalation: The loss distribution shows no consistent stop level, with 'variable' stop-loss evidence and max_over_median ratio of 1.83. Mid-series losses clustered near -$6,000 to -$8,100, but later losses escalated to -$11,818, -$12,692, and -$14,877, indicating widening stop distances or increased dollar exposure over time.
- Outlier Dependency: The top five trades account for 28.13% of gross profit; removing them reduces PF to 2.20, but the remaining edge is still dependent on the consistency of small wins. However, the size of the outliers needed to offset the few large losses makes the strategy vulnerable to any missed trend expansion.
客观数学打分明细
- [Base Score] All valid statements start at 50 points neutral benchmark (Grade B-)
- +15 pts [Lot Discipline]: Fixed/proportional lot execution, no adverse counter-trend doubling
- -20 pts [Poor Payoff]: Win/loss payoff ratio only 0.08 (< 0.6), losses heavily erode gains
- +5 pts [Expectancy]: Overall Profit Factor reached PF 3.07 with solid positive mathematical expectancy
- +10 pts [Organic Profit]: Removing top 5% windfall deals leaves PF 2.20 >= 1.2; robust and non-luck-driven
- -10 pts [Elevated Drawdown]: Max historical drawdown reached 25.5% (20%-35%), downside pressure elevated
- +10 pts [Sample Horizon]: Sample duration spans 672 days (>= 180 days), covering diverse regimes
- Final Strategy Safety Score: 60 / 100 (Grade B).
下一步行动与优化建议
- Obtain paired-ticket entry data to confirm that no concurrent multi-position exposure occurs and to accurately compute per-trade holding periods.
- Request verification of stop-loss placement method (e.g., fixed SL, trailing stop) from the EA source code to explain why losses vary from -$5,377 to -$14,877.
- Re-run the backtest with realistic broker commission and variable spread settings to validate the net expectancy net of costs.
- Forward-test on a demo account for at least 3-6 months to sample a different market regime and assess live slippage.
- Conduct a sensitivity analysis on the dynamic volume risk profile (ap_risk_profile=3) to understand the impact of compounding on drawdown depth.
关于本报告的关键质询与答疑 (Q&A About This Report)
Why is the strategy safety grade only B despite a 3.07 profit factor and a 97.51% win rate?
The B grade reflects the strategy's asymmetric payoff structure. Although it wins 97.51% of the time, its average loss of $9,305.78 is 12.8 times larger than its average win of $727.5, a realized payoff ratio of 0.08. This requires an extremely fragile balance: just a few more large losses would erase a substantial portion of the accumulated small gains, as demonstrated by the nine losses totaling $83,752 (65% of gross profit). The 25.5% maximum drawdown also shows that the strategy has already experienced material equity erosion.
What do the stop-loss exits reveal about the strategy's risk management?
The stop-loss evidence shows a 'variable' pattern with a clustered share of only 11.1% and a maximum-to-median loss ratio of 1.83. This indicates that losses are not terminating at a consistent level; instead, they range widely from -$5,377 to -$14,877. While the presence of a trailing stop is configured in the parameters (ap_trailing_enabled=true), the actual executed stop levels appear to drift, suggesting possible manual intervention, widening stop distances, or varying position sizes that amplify dollar risk.
Is the profit concentrated in a few outlier trades?
Yes, to a moderate degree. The top 5% of trades (19 out of 362) accounted for 28.13% of gross profit. However, removing these top trades still leaves a profit factor of 2.20, which indicates the strategy retains a viable edge without the windfalls. This is consistent with a trend-following/high-win-rate archetype where occasional large wins are expected, but the overall system does not rely solely on these outliers for profitability.
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.