My AI Gold V1.82 EA · MT5 XAUUSD… AI Review | QuantTester
My AI Gold V1.82 EA · MT5 XA: return 31.7%, DD 2.00%, PF 2.03, 469 trades. Read the review.
Strategy Safety Grade: A (85 / 100) · Robust & Controlled
Trading risk analysis result
Strategy Type: Strict Lot Discipline. Classified as Strict Lot Discipline (single-order trend/swing with fixed 0.01 lots): across 469 parsed closed rows (2025-01-03 to 2026-11-16, 681 days) the EA delivered a 2.03 profit factor, 80.38% win rate, 31.72% return on a 10,000 base and only 2.0% maximum drawdown, with robust non-outlier expectancy (PF 1.84 after removing the top 5% of trades) and losses terminating at a consistent -40 USD hard stop, though the realized 0.47 payoff ratio makes results heavily dependent on sustained win-rate discipline.
Darwinex 集中度压力测试 (Top 5% Windfall Dependence Test)
剔除排名前 24 笔(5%)极端盈利单后:毛利占比 9.62%,剩余净利润 $2,572.00,剩余 PF 1.84。
Data transparency & Environment
- MT4/MT5: MT5
- Symbol: XAUUSD
- Timeframe: H1
- Net profit: $3,172.00
- Return: 31.7%
- Max drawdown: 2.00%
- Total Trades: 469
- Profit Factor (PF): 2.03
Critical Algorithmic Vulnerabilities
- Payoff Asymmetry Fragility (La Loss Aversion): With an 80.38% win rate but a realized payoff ratio of 0.47 (average win 16 USD vs average loss -34 USD), expectancy of 6.76 USD per trade relies on sustaining the high win rate; a modest degradation in hit rate (roughly a 6-7 percentage point drop) mathematically flips arithmetic expectancy negative and cannot be absorbed by larger winners, which are capped at 25 USD (largest win).
- Return Consistency / Regime Sensitivity (R+): Profit accumulation rests on repeated modest wins (gross profit 6,253 USD across 371 winning rows) while every loser is full-size at -40 USD (gross loss 3,067 USD across 88 rows); in a choppy range that converts several consecutive wins into stops, the longest loss streak is only 3 in-sample, so the realized streak profile offers no evidence of tolerance to a longer adverse run.
- Risk Stability (Rs): Maximum simultaneous open exposure as inferable from the audit facts is 2 orders (max_sim_loss 80 USD), meaning each stop event consumes roughly 2.4 average wins; any widening of the executed -40 USD stop distance or an increase in consecutive stop clusters would scale dollar risk faster than the 16 USD baseline win.
- Execution Friction Sensitivity: The strategy enters at fixed clock hours and operates on XAUUSD H1 with 0.69 trades per day and an average hold of 513.6 minutes; absent a documented spread/slippage model in report_context and with zero commission/swap recorded, the 16 USD modal win cannot be assumed to survive realistic spread widening or execution latency in live conditions.
- Capacity / Holding Risk: 64.6% of trades are long and average holding time is ~8.6 hours, so overnight swap cost and gap exposure are structurally embedded; these are not quantified in the supplied report and represent unverified live transmission channels.
Score Breakdown Details
- [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 0.47 (< 0.6), losses heavily erode gains
- +5 pts [Expectancy]: Overall Profit Factor reached PF 2.03 with solid positive mathematical expectancy
- +10 pts [Organic Profit]: Removing top 5% windfall deals leaves PF 1.84 >= 1.2; robust and non-luck-driven
- +15 pts [Minimal Drawdown]: Max historical drawdown only 2.0% (< 10%), excellent downside preservation
- +10 pts [Sample Horizon]: Sample duration spans 681 days (>= 180 days), covering diverse regimes
- Final Strategy Safety Score: 85 / 100 (Grade A).
Recommended Next Actions
- Re-export the MT5 statement with full paired ticket/entry data to confirm order structure and verify that no concurrent position sets were correlated.
- Run a dedicated spread/slippage sensitivity test on XAUUSD H1 to quantify how much modal win erosion (16 USD baseline) the 0.47 payoff can absorb before expectancy drops below zero.
- Conduct a forward demo observation to confirm the win rate holds near 80% under live spreads rather than the in-sample backtest conditions.
- Verify the actual hard stop implementation (trade-side SL vs internal equity cut) against the observed -40 USD clustered loss level.
- Stress-test the loss-streak profile beyond the in-sample longest streak of 3 to size capital for a plausible adverse run.
Q&A About This Report (Quant Due Diligence)
The payoff ratio is only 0.47 (average win 16 USD vs average loss -34 USD), yet the grade is A - how do those two facts reconcile?
The 0.47 payoff is a structural feature of this high-win-rate design: 80.38% of the 469 trades win at roughly 16 USD, and the 88 losing rows cluster at -40 USD (median and maximum both -40, 83% clustered). The profit factor of 2.03 and expectancy of 6.76 USD per trade show the win frequency more than offsets the loss size. The A grade reflects that overall expectancy, the 2.0% maximum drawdown and the 681-day sample horizon, while the payoff deficiency is explicitly accounted for as a deduction in the assessment.
Is the profit dependent on a few lucky outliers?
No. The computed statistics show the top 1% of trades (5 rows) account for 2% of gross profit, and the top 5% (24 rows) account for 9.62%. Removing the top 5% still leaves net profit of 2,572 USD and a profit factor of 1.84 versus the 2.03 baseline. The edge is distributed across many rows rather than concentrated in a handful of windfalls.
How does the EA terminate losing trades, and is stop-loss discipline actually demonstrated?
The stop_loss_evidence computation shows 88 losing trades with a median and maximum loss both at -40 USD and a max_over_median ratio of 1.0, with 83% of losses at that clustered level - a signature of a consistent executed stop rather than runaway holding. Note that the parameters do not enumerate a dedicated stop-loss input, so this is inferred from realised loss clustering rather than from documented input settings. A small number of shallower losses at -2 to -27 USD appear in the table, consistent with an internal exit or trailing mechanism alongside the fixed cut.
What would invalidate the current Grade A assessment?
Two observable thresholds matter most. First, the realized payoff of 0.47 means a roughly 6-7 percentage point decline in the 80.38% win rate would push expectancy to zero or below, because average winners (16 USD, largest 25 USD) cannot compensate for larger losses (-40 USD). Second, the sample's longest loss streak is only 3 trades; an in-sample run of longer consecutive stops would consume a larger share of capital than the 2.0% maximum drawdown shows. If either the win rate or the loss-streak profile degrades materially in forward data, the assessment would need revision.
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.