Feature Iteration Run #575: zz_sylvain
Status: completed | Started: 2026-07-18 03:50:41 | Ended: 2026-07-18 03:52:10
⭐ Best Accepted Iteration
Iteration 1: (i_rsi(k_close, 14) - 50) * k_high.pct_change(12)
baseline PF: 1.286 → modified PF: 1.360 | Cohen's d: 0.000 | p-value: 1.000
📊 Iterations
| Iter | Y Label | Model | Warm | Proxy | Parent FS | Feature | Baseline PF | Modified PF | Scalar Score | Result |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | — |
gradient_boosting |
— | — | #7 | k_loop_iter_1_f3389e | 1.286 | 1.360 | 0.4154 | ✅ ACCEPT |
| 2 | — |
gradient_boosting |
— | — | #7 | k_iter_2_i_rsik_close, 14) - 50) * i_atrk_high, k_low, k_close, 10) / k_close | 0.000 | 0.000 | 0.0000 | ❌ REJECT |
| 3 | — |
gradient_boosting |
— | — | #7 | k_loop_iter_3_0228fb | 1.286 | 1.390 | 0.5766 | ✅ ACCEPT |
📄 Report Markdown
# AUTOLOOP: zz_sylvain (3 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**: 88.11s
> **Total accepted**: 2
> **Total rejected**: 1
## Iterations 詳情
### Iteration 1
- DSL: `(i_rsi(k_close, 14) - 50) * k_high.pct_change(12)` (feature=k_loop_iter_1_f3389e)
- DSL valid: True
- FeatureSet candidate id: 679
- baseline PF: 1.286 → modified PF: 1.360
- t_stat: 0.000, p_value: 1.000, cohens_d: 0.000
- scalar_score: 0.4154
- accepted: True, reason: accepted
**Metrics summary:**
- profit_factor: baseline=1.286, modified=1.360, diff=0.075
- sharpe_ratio: baseline=7.879, modified=9.490, diff=1.611
- win_rate: baseline=0.553, modified=0.519, diff=-0.034
- max_drawdown: baseline=0.075, modified=0.055, diff=-0.020
- test_accuracy: baseline=0.986, modified=0.986, diff=0.000
### Iteration 2
- DSL: `(i_rsi(k_close, 14) - 50) * i_atr(k_high, k_low, k_close, 10) / k_close` (feature=k_iter_2_i_rsik_close, 14) - 50) * i_atrk_high, k_low, k_close, 10) / k_close)
- DSL valid: False
- FeatureSet candidate id: None
- baseline PF: 0.000 → modified PF: 0.000
- t_stat: 0.000, p_value: 1.000, cohens_d: 0.000
- scalar_score: 0.0000
- accepted: False, reason: safe_eval DSL produces inf when ['k_close'] is zero; avoid unprotected division by these variables
### Iteration 3
- DSL: `(i_rsi(k_close, 14) - 50) * (k_high.pct_change(12) - k_low.pct_change(12))` (feature=k_loop_iter_3_0228fb)
- DSL valid: True
- FeatureSet candidate id: 680
- baseline PF: 1.286 → modified PF: 1.390
- t_stat: 0.000, p_value: 1.000, cohens_d: 0.000
- scalar_score: 0.5766
- accepted: True, reason: accepted
**Metrics summary:**
- profit_factor: baseline=1.286, modified=1.390, diff=0.104
- sharpe_ratio: baseline=7.879, modified=10.082, diff=2.203
- win_rate: baseline=0.553, modified=0.540, diff=-0.013
- max_drawdown: baseline=0.075, modified=0.057, diff=-0.018
- test_accuracy: baseline=0.986, modified=0.986, diff=0.000
📊 Multi-Objective Metrics
Iteration 1: k_loop_iter_1_f3389e
Y Label: — | Parent FS: #7
Objective mode: scalarize
Scalar score: 0.4154
| Metric | Baseline | Modified | Diff |
|---|---|---|---|
| win_rate | 0.553 | 0.519 | -0.034 |
| max_drawdown | 0.075 | 0.055 | -0.020 |
| sharpe_ratio | 7.879 | 9.490 | 1.611 |
| profit_factor | 1.286 | 1.360 | 0.075 |
| test_accuracy | 0.986 | 0.986 | 0.000 |
Iteration 2: k_iter_2_i_rsik_close, 14) - 50) * i_atrk_high, k_low, k_close, 10) / k_close
Y Label: — | Parent FS: #7
Objective mode: single
Scalar score: 0.0000
Iteration 3: k_loop_iter_3_0228fb
Y Label: — | Parent FS: #7
Objective mode: scalarize
Scalar score: 0.5766
| Metric | Baseline | Modified | Diff |
|---|---|---|---|
| win_rate | 0.553 | 0.540 | -0.013 |
| max_drawdown | 0.075 | 0.057 | -0.018 |
| sharpe_ratio | 7.879 | 10.082 | 2.203 |
| profit_factor | 1.286 | 1.390 | 0.104 |
| test_accuracy | 0.986 | 0.986 | 0.000 |
📁 Offline Report
Markdown backup: reports/autoloop/zz_sylvain/20260718_115210.md