Feature Iteration Run #984: zz_sylvain

# 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**: 83.00s
> **Total accepted**: 0
> **Total rejected**: 3

## Iterations 詳情

### Iteration 1
- DSL: `(obv_ensemble_3 * (1 - hl_range)) / (1 + i_atr(k_high, k_low, k_close, 14)) * i_macd(k_close, 12, 26, 9)` (feature=k_loop_iter_1_5e68f1)
- DSL valid: True
- FeatureSet candidate id: 1257
- baseline PF: 1.947 → modified PF: 1.408
- t_stat: 0.000, p_value: 0.269, cohens_d: 0.292
- scalar_score: -0.2659
- accepted: False, reason: normalized_score -0.1418 <= 0.0, sharpe_delta -8.9106 < -0.1, hypothesis test not confirmed

**Metrics summary:**
  - profit_factor: baseline=1.947, modified=1.408, diff=-0.540
  - sharpe_ratio: baseline=19.162, modified=10.252, diff=-8.911
  - win_rate: baseline=0.600, modified=0.482, diff=-0.118
  - max_drawdown: baseline=0.056, modified=0.219, diff=0.163
  - test_accuracy: baseline=0.992, modified=0.981, diff=-0.012

### Iteration 2
- DSL: `k_high.pct_change(12) * (1 + hl_range) / (1 + atr_norm) * (1 - dist_to_high) * (1 - return_5)` (feature=k_iter2_k_high_pct_change_hl_range_atr_norm)
- 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: quality_gate: C8 leakage: ['dist_to_high']

### Iteration 3
- DSL: `k_momentum_pct_change * (1 - k_close_position_in_hl) * i_rsi(k_close, 14) / (1 + atr_norm * 10)` (feature=k_iter3_k_momentum_pct_chang_k_close_position_in__i_rsi_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: 變數未定義: k_close_position_in_hl