Piner_EA_fix · MetaTrader 5… AI Review | QuantTester
Piner_EA_fix · MetaTrader 5 : return 380,448.0%, DD 64.67%, PF 11.33, 192084 trades. Read the review.
Strategy Safety Grade: D (10 / 100) · High Risk Alert
Trading risk analysis result
Strategy Type: Martingale. Classified as Martingale (exponential adverse-lot grid): the strategy achieved a backtest PF of 11.33 and 83.4% win rate over 425.8 days, but a 10-level geometric lot ladder (0.01 to 2.56, plus 20 capped 2.56-lot levels), zero hard stop-losses (sl=0 on every row), an unbounded loss tail (max/median loss ratio 669), and a 64.67% maximum drawdown make the return stream structurally fragile and unfit for live capital allocation at current sizing.
Darwinex 集中度压力测试 (Top 5% Windfall Dependence Test)
剔除排名前 9605 笔（5%）极端盈利单后：毛利占比 68.59%，剩余净利润 $9,424,072.56，剩余 PF 3.56。
Data transparency & Environment
- MT4/MT5: MetaTrader 5
- Symbol: XAUUSD
- Timeframe: H1 (2025.01.03 - 202
- Net profit: $38,044,795.28
- Return: 380,448.0%
- Max drawdown: 64.67%
- Total Trades: 192084
- Profit Factor (PF): 11.33
Critical Algorithmic Vulnerabilities
- [La — Payoff Asymmetry Fragility / Loss-Scale Break] The realised loss tail is uncontained: median loss 13.21 versus largest single loss -23,119.36 (max/median ratio 669.17) and clustered_share_pct only 4.5, meaning losses do not terminate at a consistent stop level. In live conditions a single adverse trend extension transmitted through the 2.56-lot outer tiers converts dozens of +2 to +100 micro-wins into a loss that wipes the corresponding equity tier.
- [Rs — Risk Stability Breakdown / Sizing Ladder] The configured ladder compounds 0.01 → 0.02 → 0.04 → 0.08 → 0.16 → 0.32 → 0.64 → 1.28 → 2.56 (geometric), then holds twenty 2.56-lot tiers. audit_facts records max_sim_orders 64 and max_sim_loss 338,934.76 against a 10,000 capital base — a 33.9x notional-risk multiple of starting equity. Live spread widening on XAUUSD across a 1.0-pip grid step would activate fills materially faster than modelled.
- [R+ — Return Consistency / Drawdown Magnitude] Maximum equity drawdown reached 64.67% while the account compounded to 38,044,795.28 net profit. The return is real within the test, but the drawdown is a sizing artefact: at the peak tier exposure, a further adverse leg has no terminal cutoff (no hard SL on any parsed row), so the loss distribution is left-truncated only by margin, not by design.
- [D-Score Quote Shape / Recovery Latency] With 31,613 losing trades and a 19-trade longest loss streak at an 83.4% win rate, the equity path is a high-frequency step-up curve punctuated by concentrated tier losses (e.g. -23,119.36, -8,839.68, -2,515.20, -2,017.28 in the supplied rows). Such loss clusters produce abrupt quote-shape discontinuities that degrade D-Score momentum stability even when the cumulative curve rises.
- [Execution Feasibility] The engine operates at 451.09 trades/day on a single 1.0-pip grid step. The backtest used a configured SlippagePoints=50 test setting; there is no verified live broker fill record establishing that grid anchors and basket take-profits (BasketTakeProfit 20.0) survive real spread and latency at this cadence.
Score Breakdown Details
- [Base Score] All valid statements start at 50 points neutral benchmark (Grade B-)
- -35 pts [Martingale]: Exponential lot size doubling under floating loss detected; catastrophic tail risk
- -25 pts [No Hard SL]: No disciplined hard stop-loss; holding unhedged drawdown to severe depths
- +15 pts [Payoff Ratio]: Win/loss ratio reached 2.23 (> 1.8), strong asymmetric payoff
- +5 pts [Expectancy]: Overall Profit Factor reached PF 11.33 with solid positive mathematical expectancy
- +10 pts [Organic Profit]: Removing top 5% windfall deals leaves PF 3.56 >= 1.2; robust and non-luck-driven
- -20 pts [Extreme Drawdown]: Max historical drawdown reached 64.7% (> 35%), weak defensive execution
- +10 pts [Sample Horizon]: Sample duration spans 426 days (>= 180 days), covering diverse regimes
- Final Strategy Safety Score: 10 / 100 (Grade D).
Recommended Next Actions
- Re-run the backtest with realistic per-lot commission, swap, and a widened asymmetric slippage model on XAUUSD to test whether the outer 1.28/2.56 tiers still close profitably
- Stress-test the ladder under an adverse trend extension beyond the 40 configured levels (20 upper + 20 lower) to quantify the loss when the lot cap at 2.56 is exhausted
- Introduce and verify a hard equity stop or basket-level maximum-loss cutoff, then re-measure max drawdown and the loss tail (max/median ratio) to confirm the 669x outlier is eliminated
- Set the capital base so that the 338,934.76 simulated peak concurrent loss represents a tolerable fraction (e.g. under 20%) of equity, rather than 33.9x the 10,000 starting nominal
- Validate on a Demo forward account for at least 3–6 months to observe real fill quality on the 1.0-pip grid before any capital commitment
Q&A About This Report (Quant Due Diligence)
Why does a strategy with an 11.33 profit factor and 83.4% win rate only receive a Grade D?
The grade reflects internal downside mechanics, not headline profitability. The strategy doubles lot size geometrically down an adverse ladder (0.01 to 2.56, then twenty 2.56-lot tiers), holds no hard stop-loss on any parsed row, and has already realised a 64.67% maximum equity drawdown with a largest single loss of -23,119.36 against a median loss of 13.21. The positive expectancy is real, but it is purchased with a loss distribution that has no designed terminal cutoff.
The top 5% of trades account for 68.59% of gross profit — does that mean the edge is luck?
No. Removing those 9,605 rows still leaves residual net profit of 9,424,072.56 at a profit factor of 3.56, comfortably above break-even. The core micro-edge is repeatable within the test; the concentration sits in the large outer-tier exits, which is the expected shape of a lot-ladder system rather than an artefact.
Is the extended holding time on some trades evidence of missing stop-loss discipline?
Holding duration alone is not the issue — the decisive evidence is the realised loss distribution. The sl field is 0 on every parsed row, clustered_share_pct is only 4.5%, and the max-to-median loss ratio is 669.17, which matches an unbounded_tail pattern. Losses terminate when the basket reverses or margin forces it, not at a pre-set stop level. That is the structural weakness, independent of how long each position sits open.
At what position size would this ladder be institutionally tolerable?
The report shows a simulated peak concurrent floating loss of 338,934.76 with up to 64 simultaneous orders. Against the 10,000 nominal starting capital that is a 33.9x multiple of equity, which is why the ladder was flagged. Any deployment would require either drastically scaling the 2.56-lot cap down or funding the account so that peak concurrent exposure stays a small single-digit fraction of equity — but no such resizing is evidenced in this record.
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.