跳到主要內容

Feature Iteration Run #539: zz_sylvain

Status: completed | Started: 2026-07-17 19:00:30 | Ended: 2026-07-17 19:01:57

⭐ Best Accepted Iteration

Iteration 1: (i_hma(k_close, 20) - k_close) / i_atr(k_high, k_low, k_clo…

baseline PF: 1.286 → modified PF: 1.775 | Cohen's d: 0.000 | p-value: 1.000

📊 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_0fd4e9 (i_hma(k_close, 20) - k_close) / i_atr(… ✅ 1.286 1.775 2.5773 0.000 1.000 0.000 ✅ ACCEPT accepted
2 — gradient_boosting — — — k_loop_iter_2_fcddd5 (k_close - i_ema(k_close, 50)) / k_clos… ✅ 1.286 1.497 1.1896 0.000 1.000 0.000 ✅ ACCEPT accepted

📄 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**: 87.25s
> **Total accepted**: 2
> **Total rejected**: 0

## Iterations 詳情

### Iteration 1
- DSL: `(i_hma(k_close, 20) - k_close) / i_atr(k_high, k_low, k_close, 14)` (feature=k_loop_iter_1_0fd4e9)
- DSL valid: True
- FeatureSet candidate id: 615
- baseline PF: 1.286 → modified PF: 1.775
- t_stat: 0.000, p_value: 1.000, cohens_d: 0.000
- scalar_score: 2.5773
- accepted: True, reason: accepted

**Metrics summary:**
  - profit_factor: baseline=1.286, modified=1.775, diff=0.489
  - sharpe_ratio: baseline=7.879, modified=17.592, diff=9.713
  - win_rate: baseline=0.553, modified=0.558, diff=0.005
  - max_drawdown: baseline=0.075, modified=0.084, diff=0.010
  - test_accuracy: baseline=0.986, modified=0.986, diff=0.000

### Iteration 2
- DSL: `(k_close - i_ema(k_close, 50)) / k_close.rolling(30).std()` (feature=k_loop_iter_2_fcddd5)
- DSL valid: True
- FeatureSet candidate id: 616
- baseline PF: 1.286 → modified PF: 1.497
- t_stat: 0.000, p_value: 1.000, cohens_d: 0.000
- scalar_score: 1.1896
- accepted: True, reason: accepted

**Metrics summary:**
  - profit_factor: baseline=1.286, modified=1.497, diff=0.211
  - sharpe_ratio: baseline=7.879, modified=12.393, diff=4.514
  - win_rate: baseline=0.553, modified=0.560, diff=0.007
  - max_drawdown: baseline=0.075, modified=0.051, diff=-0.023
  - test_accuracy: baseline=0.986, modified=0.987, diff=0.000

📊 Multi-Objective Metrics

Iteration 1: k_loop_iter_1_0fd4e9

Y Label: — | Parent FS: #—

Objective mode: scalarize

Scalar score: 2.5773

MetricBaselineModifiedDiff
win_rate 0.553 0.558 0.005
max_drawdown 0.075 0.084 0.010
sharpe_ratio 7.879 17.592 9.713
profit_factor 1.286 1.775 0.489
test_accuracy 0.986 0.986 0.000

Iteration 2: k_loop_iter_2_fcddd5

Y Label: — | Parent FS: #—

Objective mode: scalarize

Scalar score: 1.1896

MetricBaselineModifiedDiff
win_rate 0.553 0.560 0.007
max_drawdown 0.075 0.051 -0.023
sharpe_ratio 7.879 12.393 4.514
profit_factor 1.286 1.497 0.211
test_accuracy 0.986 0.987 0.000

📁 Offline Report

Markdown backup: reports/autoloop/zz_sylvain/20260718_030157.md