The Gold Prop Guardian_XAUU · MT4/MT5… AI 测评｜QuantTester
The Gold Prop Guardian_XAUU · MT4/M 报告：收益 -23.2%、回撤 28.88%、PF 0.62、223 笔交易。AI 风险解读。
策略安全信用评级: D+ (18 / 100) · 高危预警 (High Risk)
交易风险分析结果
A 93.7-day real-account XAUUSD statement shows a losing, cost-negative profile (-23.22% return, PF 0.62, 28.88% max drawdown) driven by a small number of cluster-closed rows, so the strategy's account-level risk is confirmed weak on this evidence while attribution of the cluster structure remains unverified.
数据透明度与环境说明
- MT4/MT5: MT4/MT5
- 品种: XAUUSD
- 周期:
- 总净利润: $-990.21
- 总收益率: -23.2%
- 最大回撤: 28.88%
- 总交易笔数: 223
- 盈利因子 (PF): 0.62
AI 识别的核心高危漏洞清单
- Account lost 23.22% with a 28.88% maximum drawdown; the loss profile is confirmed by the statement totals, not by sampling.
- Profit factor 0.62 with a 64.57% win rate: the average loss must materially exceed the average win for this combination, i.e. negative payoff asymmetry on the same account.
- Concentration: without the 3 best rows net profit becomes -1,211.31 (PF 0.53); without the 12 best it becomes -1,642.64 (PF 0.36).
- Repeated same-timestamp closing clusters (up to 6 rows at one timestamp, 2026-07-16 16:02:38) sit on the largest loss days, indicating that the worst outcomes arrive as grouped exposures rather than isolated trades.
- The equity base was reduced by 3,273.57 of withdrawals during a losing run, which mechanically shrinks the cushion available for recovery.
客观数学打分明细
- Verified negative: 223 closed rows over 2026-06-30 to 2026-10-02 (93.7 days) produced -23.22% return and -990.21 net profit on a 4,263.78 capital base, with profit factor 0.62.
- Verified negative: max drawdown 28.88% against a 64.57% win rate shows losses are much larger than wins (negative payoff asymmetry).
- Verified negative: top 1% of rows (3 rows) contribute 13.94% of gross profit; excluding them the remainder is -1,211.31 net at PF 0.53; top 5% (12 rows) contribute 41.15%, leaving -1,642.64 net at PF 0.36. Profit depends on a very small set of rows.
- Verified negative: repeated large clusters closed at identical timestamps (e.g. 2026-07-16 16:02:38 with 6 rows from -111.60 to +20.61, 2026-10-02 04:09:29 with 3 rows of -94.77/-111.84/-85.86, 2026-09-28 03:31:09 with 5 rows) — these coincide with the account's worst loss days and are a documented loss-concentration pattern.
- Verified positive: 64.57% of closed rows are winners, and 4,263.78 deposits were partly returned via 3,273.57 withdrawals during the period.
- Pending: gain/loss attribution by entry ticket is not possible because the parsed rows are close-event rows; the full report fields are not the limit of the source data.
- Pending: no independent Demo or Real forward comparison period exists to confirm the profile outside this statement window.
下一步行动与优化建议
- Obtain the raw MT4/MT5 report with entry tickets and opening times to pair entries and exits, then reclassify the same-timestamp closing clusters accurately.
- Compute per-row cost (commission, swap, spread) or rerun the strategy with broker cost fields included to separate execution cost from directional loss.
- Re-derive the drawdown as an episode series from the full equity column to see how many distinct drawdown periods exist and their depths.
- Ask the operator for the documented lot-sizing and risk rules, then verify them against the observed 0.01–0.04 lot sequence.
- Collect a longer real-account period (beyond 93.7 days) and, if available, a comparable backtest of the same EA settings to check whether the loss profile is regime-specific.
关于本报告的关键质询与答疑 (Q&A About This Report)
Why is the return negative when 64.57% of trades are winners?
The statement shows profit factor 0.62 with a 64.57% win rate, so the strategy wins often but not enough in size: the average losing row is much larger than the average winning row, and the totals are -990.21 on a 4,263.78 base (-23.22%).
How dependent is the result on a few large trades?
Very concentrated. computed_trade_stats show the best 3 rows (top 1%) carry 13.94% of gross profit, and removing them turns the rest into -1,211.31 with PF 0.53; the best 12 rows (top 5%) carry 41.15%, and removing them leaves -1,642.64 with PF 0.36.
What do the repeated same-timestamp closing rows mean?
Rows such as the six closed at 2026-07-16 16:02:38 and the three closed at 2026-10-02 04:09:29 are grouped exits around the largest loss days. Because the parsed rows do not carry entry-ticket pairing, the data only shows repeated same-timestamp closed rows and cannot confirm whether they were a grid, scale-in, or another cluster structure.
How long is the record and does it include forward testing?
The parsed statement covers 2026-06-30 to 2026-10-02, about 93.7 days, with 223 closed rows on a real account; there is no separate Demo or backtest comparison period in this evidence, so the loss profile is documented for this window only.
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.