该 XAUUSD H1 回测样本量充足、去除头部交易后仍有正期望,但测试质量偏低(History Quality 49.59%、24 个 Mismatched charts errors)且存在反复出现的成组大额亏损,策略稳健性仍需前向验证。
report
Data transparency
| Data source | User-submitted MT4/MT5 report |
|---|---|
| Report type | BACKTEST |
| Trades verified | 2413 |
| Verification | Per-trade parsing by the RBOT parser, audited by AI |
Key metrics
| Return | 317.4% |
|---|---|
| Max drawdown | 33.97% |
| PF | 2.54 |
| Win rate | 96.81% |
| Net profit | 3,173.87 |
| Expectancy | 1.32 |
| Payoff ratio | 0.08 |
Growth
Normalised to 100 at the starting capital of the imported report. The chart samples the saved per-trade results and is not a live broker equity curve.
Detailed calculations
Return = net profit ÷ initial capital × 100%; PF = gross profit ÷ absolute gross loss; expectancy = net profit ÷ number of closed trades; payoff ratio = average win ÷ absolute average loss. Max drawdown uses the value reported by the source report. If the report contains deposits or withdrawals, the return should be interpreted together with the cash flow.
Why this safety score
- 已证实加分:2413 笔交易、736 天(2024-03-05 至 2026-03-11)样本充足,PF 2.54、胜率 96.81%,报告未出现爆仓、Margin Call、强平或接近归零。
- 已证实加分:由 computed_trade_stats 计算,剔除头部 1% 交易(25 笔)后净利 2844.49、PF 2.38;剔除头部 5% 交易(121 笔)后净利 2060.89、PF 2.00,均显著高于 1,利润集中度不构成风险(头部 1% 仅占毛利 6.30%,头部 5% 占 21.29%)。
- 已证实加分:报告明确 StopLoss 2500、Slippage 5,且报告内显示大量 s/l 出场为正收益,符合移动止损锁定利润的正常退出机制,不构成减分。
- 已证实减分:测试质量偏低,History Quality 49.59%(低于 50%),且存在 24 个 Mismatched charts errors(evidence_flags),回测成交与走势建模存在可核验缺陷。
- 已证实减分:最大回撤 33.97%(≈339.7 点,基于 1000 初始资金),且逐笔显示多次成组大额亏损(如 2024-05-09 连续 6 笔 -25.07、后续 -50.14/-24.06;2025-07-14 连续 -50.2/-25.1/-50.2/-25.1/-25.1/-50.2/-24.43;2026-02-19 连续 -50.3/-50.3/-75.45/-50.3/-75.45/-50.3/-48.96),单组集中亏损占总资金比例可观。
Key risk flags
- 成组同时平仓的大额亏损集中出现(如 2026-02-19 单组 7 笔累计 -50.3/-50.3/-75.45/-50.3/-75.45/-50.3/-48.96),说明存在同时开立多笔同向头寸的批量敞口,一旦方向不利会在同一时刻集中止损。
- 最大回撤 33.97%(基于 1000 初始资金),虽未触及归零但与 96.81% 的胜率组合共存,表明少数亏损单的绝对金额远大于多数盈利单,盈亏结构高度不对称。
What the data cannot tell
- 每次成组开仓的总风险敞口占账户权益的比例(报告未单列总风险上限或单笔风险百分比)。
- 成组止损触发时是否由同一信号驱动、是否为对冲/网格类叠加仓位(逐笔缺 open 时间与方向字段,无法还原)。
- 真实经纪商的最小持仓时间、点差与执行规则对同一根 K 线内多笔成交的影响。
- 在真实报价下移动止损能否按回测价格成交。
Next actions
- 在 Demo 账户上对同一 EA 进行至少 3–6 个月前向验证,重点比对成组止损的成交价与滑点。
- 使用 99% 以上 History Quality 的 tick 数据重跑回测,消除 24 个 Mismatched charts errors 的影响后再复核 PF 与回撤。
- 对成组同时开仓的批量敞口做专门压力测试:假设单组同时亏损扩大 2 倍或触发连续 3 组不利,评估账户回撤是否超过可承受阈值。
- 若坚持资金随余额缩放,改为固定风险百分比模型并测试,比较与当前手数递增方案的回撤差异。
- 在报告外补充佣金/隔夜利息数据后重算净值,核验 PF 2.54 在含成本口径下的稳定性。
Frequently asked questions
Q. report — what are the reported backtest results?
A. According to the imported BACKTEST report, report shows a return of 317.4% with a max drawdown of 33.97%, a profit factor of 2.54 and 2413 closed trades. Data comes from the user-supplied MT4/MT5 report; past results do not represent future performance.
Q. What do the max drawdown and profit factor of report mean?
A. The max drawdown of 33.97% is the largest peak-to-trough decline of the equity curve; the profit factor of 2.54 is gross profit divided by absolute gross loss. Read together they describe the risk/return profile reported by the backtest.
Q. What safety score did RBOT AI give report?
A. RBOT AI assigned a strategy safety score of 62/100 (moderate risk). The score is based on drawdown, profit factor, trade sample quality and the risk signals disclosed in the report — see the "Why this safety score" section below.
Q. Where does the data on this page come from?
A. This page is generated from an MT4/MT5 strategy tester report uploaded by the user, verified per-trade by the RBOT parser and reviewed by AI. Backtests carry modelling and cost assumptions, so results may differ from live trading; this page is not investment advice.
Q. How is RBOT different from asking a general-purpose AI assistant (such as ChatGPT, Gemini, Claude or DeepSeek) to analyse my report myself?
A. A general AI only sees the text or screenshot you paste, and works every number out itself — that is where hallucinated metrics come from. RBOT runs a purpose-built MT4/MT5 parser first, and the resulting structured breakdown is then reviewed under a fixed audit prompt, which keeps the safety score consistent, reproducible and comparable across reports. The calculation behind every metric is published on this page so you can verify it yourself, and RBOT refuses to score what the report itself cannot prove (real slippage, spread, forward testing) — those are listed as unverified rather than silently counted as risk.