Feature Iteration Run #912: zz_sylvain

# AUTOLOOP: zz_sylvain (5 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**: 130.08s
> **Total accepted**: 0
> **Total rejected**: 5

## Iterations 詳情

### Iteration 1
- DSL: `obv_ensemble_3 * (1 - dist_to_high) / (1 + i_atr(k_high, k_low, k_close, 14))` (feature=k_iter1_dist_to_high_i_atr_k_high_k_low)
- 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 2
- DSL: `k_momentum_pct_change * obv_ensemble_3 / (1 + i_atr(k_high, k_low, k_close, 14) / k_close)` (feature=k_loop_iter_2_86396b)
- DSL valid: True
- FeatureSet candidate id: 1236
- baseline PF: 1.947 → modified PF: 1.138
- t_stat: 0.000, p_value: 0.512, cohens_d: 0.261
- scalar_score: -0.3923
- accepted: False, reason: normalized_score -0.2435 <= 0.0, sharpe_delta -14.6732 < -0.1, hypothesis test not confirmed

**Metrics summary:**
  - profit_factor: baseline=1.947, modified=1.138, diff=-0.810
  - sharpe_ratio: baseline=19.162, modified=4.489, diff=-14.673
  - win_rate: baseline=0.600, modified=0.466, diff=-0.134
  - max_drawdown: baseline=0.056, modified=0.166, diff=0.110
  - test_accuracy: baseline=0.992, modified=0.981, diff=-0.012

### Iteration 3
- DSL: `k_momentum_pct_change * obv_ensemble_3 / (1 + atr_norm * 10)` (feature=k_iter3_k_momentum_pct_chang_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: canonical duplicate: k_momentum_pct_change * obv_ensemble_3 / (1 + atr_norm * 10)

### Iteration 4
- DSL: `k_momentum_pct_change * i_rsi(k_close, 14) / (1 + atr_norm * 100)` (feature=k_iter4_k_momentum_pct_chang_i_rsi_k_close_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: NaN ratio 1.000 exceeds 0.3

### Iteration 5
- DSL: `k_momentum_pct_change * o_pct_change_12 / (1 + atr_norm * 10)` (feature=k_loop_iter_5_497257)
- DSL valid: True
- FeatureSet candidate id: 1237
- baseline PF: 1.947 → modified PF: 1.085
- t_stat: 0.000, p_value: 0.463, cohens_d: 0.243
- scalar_score: -0.4478
- accepted: False, reason: normalized_score -0.2581 <= 0.0, sharpe_delta -16.8215 < -0.1, hypothesis test not confirmed

**Metrics summary:**
  - profit_factor: baseline=1.947, modified=1.085, diff=-0.862
  - sharpe_ratio: baseline=19.162, modified=2.341, diff=-16.821
  - win_rate: baseline=0.600, modified=0.500, diff=-0.100
  - max_drawdown: baseline=0.056, modified=0.187, diff=0.131
  - test_accuracy: baseline=0.992, modified=0.992, diff=-0.000