跳到主要內容

Feature Iteration Run #421: zz_sylvain

Status: completed | Started: 2026-07-16 17:27:45 | Ended: 2026-07-16 17:29:46

📊 Iterations

Iter Y Label Model Warm Proxy Parent FS Feature DSL DSL Valid Baseline PF Modified PF Scalar Score t-stat p-value Cohen's d Result Reason
1 — gradient_boosting — — — k_loop_iter_1_1424d9 k_high - k_low.shift(1) + k_volume.pct_… ✅ 1.037 0.929 -0.7988 0.000 1.000 0.000 ❌ REJECT p=1.000 ≥ 0.05, cohens_d=0.000 ≤ 0.3, constraints: ['sharpe_ratio >= 1.0: got -2.2654']
2 — gradient_boosting — — — k_loop_iter_2_681381 k_close.pct_change(12) * vol_ratio + K_… ✅ 1.198 0.000 -1.9420 0.000 1.000 0.000 ❌ REJECT p=1.000 ≥ 0.05, cohens_d=0.000 ≤ 0.3, constraints: ['sharpe_ratio >= 1.0: got 0.0000', 'win_rate >= 0.4: got 0.0000']

📄 Report Markdown

# AUTOLOOP: zz_sylvain (2 iterations)

> **Phase**: Y4.7 Option B (DSL → Python code → FeatureSet candidate)
> **Target indicator**: zz_sylvain
> **Base feature_set_id**: 7
> **Claim**: (default) Propose a new k_* feature that improves risk-adjusted returns
> **Objective mode**: scalarize
> **Metrics**: ['profit_factor', 'sharpe_ratio', 'win_rate', 'max_drawdown', 'test_accuracy']
> **Total duration**: 121.42s
> **Total accepted**: 0
> **Total rejected**: 2

## Iterations 詳情

### Iteration 1
- DSL: `k_high - k_low.shift(1) + k_volume.pct_change(5)` (feature=k_loop_iter_1_1424d9)
- DSL valid: True
- FeatureSet candidate id: 438
- baseline PF: 1.037 → modified PF: 0.929
- t_stat: 0.000, p_value: 1.000, cohens_d: 0.000
- scalar_score: -0.7988
- accepted: False, reason: p=1.000 ≥ 0.05, cohens_d=0.000 ≤ 0.3, constraints: ['sharpe_ratio >= 1.0: got -2.2654']

**Metrics summary:**
  - profit_factor: baseline=1.037, modified=0.929, diff=-0.108
  - sharpe_ratio: baseline=0.995, modified=-2.265, diff=-3.260
  - win_rate: baseline=0.419, modified=0.428, diff=0.009
  - max_drawdown: baseline=0.120, modified=0.452, diff=0.332
  - test_accuracy: baseline=0.971, modified=0.941, diff=-0.030

### Iteration 2
- DSL: `k_close.pct_change(12) * vol_ratio + K_close_position_in_hl * k_loop_iter_2_0d3edb.pct_change(5)` (feature=k_loop_iter_2_681381)
- DSL valid: True
- FeatureSet candidate id: 439
- baseline PF: 1.198 → modified PF: 0.000
- t_stat: 0.000, p_value: 1.000, cohens_d: 0.000
- scalar_score: -1.9420
- accepted: False, reason: p=1.000 ≥ 0.05, cohens_d=0.000 ≤ 0.3, constraints: ['sharpe_ratio >= 1.0: got 0.0000', 'win_rate >= 0.4: got 0.0000']

**Metrics summary:**
  - profit_factor: baseline=1.198, modified=0.000, diff=-1.198
  - sharpe_ratio: baseline=5.537, modified=0.000, diff=-5.537
  - win_rate: baseline=0.443, modified=0.000, diff=-0.443
  - max_drawdown: baseline=0.082, modified=0.000, diff=-0.082
  - test_accuracy: baseline=0.973, modified=0.000, diff=-0.973

📊 Multi-Objective Metrics

Iteration 1: k_loop_iter_1_1424d9

Y Label: — | Parent FS: #—

Objective mode: scalarize

Scalar score: -0.7988

Failed constraints: sharpe_ratio >= 1.0: got -2.2654

MetricBaselineModifiedDiff
win_rate 0.419 0.428 0.009
max_drawdown 0.120 0.452 0.332
sharpe_ratio 0.995 -2.265 -3.260
profit_factor 1.037 0.929 -0.108
test_accuracy 0.971 0.941 -0.030

Iteration 2: k_loop_iter_2_681381

Y Label: — | Parent FS: #—

Objective mode: scalarize

Scalar score: -1.9420

Failed constraints: sharpe_ratio >= 1.0: got 0.0000, win_rate >= 0.4: got 0.0000

MetricBaselineModifiedDiff
win_rate 0.443 0.000 -0.443
max_drawdown 0.082 0.000 -0.082
sharpe_ratio 5.537 0.000 -5.537
profit_factor 1.198 0.000 -1.198
test_accuracy 0.973 0.000 -0.973

📁 Offline Report

Markdown backup: reports/autoloop/zz_sylvain/20260717_012946.md