# Professional Eval Report: Focused Hurst MR Alpaca SIP 5m refreshed 2026-06-02

## Executive Decision

- 结论: FAIL / 先不要推进
- Bias: `BiasTest[Focused Hurst MR Alpaca SIP 5m refreshed 2026-06-02]: FAIL | shift=✗ autocorr=✓ truncate=—`
- Gross Sharpe: 0.667
- Net Sharpe: 0.605
- Gross MaxDD: -1.91%
- Net MaxDD: -2.02%
- TC Drag: 0.0574% NAV/yr
- Alpha t-stat: 2.915

## P0 Mandatory Gate

| Dimension | Status | Result |
|---|---:|---|
| Core return/risk | computed | CAGR net +0.56%, vol 0.93%, Calmar 0.276, Sortino 0.925 |
| Drawdown/recovery | computed | avg DD -0.26%, max recovery 673 days |
| Turnover | computed | annual turnover 1.436, holding 175.5 days, turnover-adj Sharpe 0.388 |
| Capacity/scalability | computed | max participation 901.32% |
| Statistical significance | partial | t-stat 2.915; DSR=needs_input; PBO=needs_input |
| Operational reliability | computed | lookahead pass=False |

## P1 Promotion Review

| Dimension | Status | Result |
|---|---:|---|
| Factor exposure | computed | alpha annualized +0.50%, R2 0.072 |
| Out-of-sample | computed | OOS Sharpe 0.110, IS Sharpe 0.704, decay -0.594 |
| Rolling Sharpe | computed | 252d latest -1.045, min -1.750, negative-window ratio 15.5% |
| Fee sensitivity | computed | see `summary.json` rows for 0x/0.5x/1x/2x/3x TC |
| Paper trading | needs_input | Requires paper/live return stream |
| Portfolio interaction | needs_input | max abs corr: needs_input; provide `--existing-returns` |

## P2 Live Readiness

| Dimension | Status | Result |
|---|---:|---|
| Short rolling Sharpe | computed | 126d latest -0.939, min -3.405 |
| Panda P&L correlation | needs_input | target < 0.8 when existing strategy P&L is supplied |
| Fill assumptions | needs_input | queue position, partial fills, latency, bid/ask crossing still required |

## 下一步 review 重点

- P0 fail 时不进入 promotion review；优先查 lookahead、时间戳对齐、label 泄露、全样本标准化。
- P1 缺输入时不能假定通过；补 factor、OOS、paper trading、existing strategy P&L 后再评估。
- P2 是 live readiness；没有 fill/latency/queue 假设时不要讨论上线。