Feature Iteration Run #1036

📊 Iterations

Iter	Y Label	Model	Warm	Proxy	Parent FS	Feature	DSL	DSL Valid	p-value	Result	Reason
1	`C8`	`lightgbm`	—	—	#7	k_iter1_i_atr_k_high_k_low_k_close	`i_atr(k_high, k_low, k_close, 14) / k_c…`	❌	1.000	❌ REJECT	safe_eval DSL: 變數未定義: abs
2	`C8`	`lightgbm`	—	—	#7	k_iter2_k_high_pct_change_hl_range_dist_to_high	`k_high.pct_change(12) * (1 + hl_range) …`	❌	1.000	❌ REJECT	quality_gate: C8 leakage: ['dist_to_high']
3	`C8`	`lightgbm`	—	—	#7	k_loop_iter_3_cca5e1	`obv_rebased.pct_change() * (1 + hl_rang…`	✅	1.000	❌ REJECT	1-trial proxy PF=0.000 < 0.95, no improvement expected
4	`C8`	`lightgbm`	—	—	#7	k_iter4_k_high_pct_change_hl_range_price_position	`k_high.pct_change(12) * (1 + hl_range) …`	❌	1.000	❌ REJECT	quality_gate: canonical duplicate: k_high.pct_change * (1 + hl_range) * (1 - price_position)
5	`C8`	`lightgbm`	—	—	#7	k_iter5_price_position_atr_norm	`o_div_amplitude_12 * (1 - price_positio…`	❌	1.000	❌ REJECT	quality_gate: canonical duplicate: o_div_amplitude_12 * (1 - price_position) / (1 + atr_norm)
6	`C8`	`lightgbm`	—	—	#7	k_loop_iter_6_58dc34	`i_macd(k_close, 12, 26, 9) / (1 + i_atr…`	✅	1.000	❌ REJECT	1-trial proxy PF=0.000 < 0.95, no improvement expected
7	`C8`	`lightgbm`	—	—	#7	k_loop_iter_7_35271b	`i_macd(k_close, 12, 26, 9) * (1 + hl_ra…`	✅	1.000	❌ REJECT	1-trial proxy PF=0.000 < 0.95, no improvement expected
8	`C8`	`lightgbm`	—	—	#7	k_loop_iter_8_7e64b2	`o_div_amplitude_12 / (1 + atr_norm) * (…`	✅	1.000	❌ REJECT	1-trial proxy PF=0.000 < 0.95, no improvement expected
9	`C8`	`lightgbm`	—	—	#7	k_loop_iter_9_ee3d29	`i_macd(k_close, 6, 13, 5) * vol_ratio /…`	✅	1.000	❌ REJECT	1-trial proxy PF=0.000 < 0.95, no improvement expected
10	`C8`	`lightgbm`	—	—	#7	k_loop_iter_10_388e74	`i_macd(k_close, 12, 26, 9) / (1 + i_atr…`	✅	1.000	❌ REJECT	1-trial proxy PF=0.000 < 0.95, no improvement expected
11	`C8`	`lightgbm`	—	—	#7	k_loop_iter_11_e1abc3	`o_pct_change_12 * (k_close - i_ema(k_cl…`	✅	1.000	❌ REJECT	1-trial proxy PF=0.000 < 0.95, no improvement expected
12	`C8`	`lightgbm`	—	—	#7	k_loop_iter_12_d8ccea	`k_volume * (k_close - i_ema(k_close, 50…`	✅	1.000	❌ REJECT	1-trial proxy PF=0.000 < 0.95, no improvement expected
13	`C8`	`lightgbm`	—	—	#7	k_iter13_i_rsi_k_close_price_position_k_volume	`i_rsi(k_close, 14) * (1 - price_positio…`	❌	1.000	❌ REJECT	quality_gate: NaN ratio 1.000 exceeds 0.3
14	`C8`	`lightgbm`	—	—	#7	k_loop_iter_14_f056e1	`(o_div_amplitude_12 - o_div_amplitude_4…`	✅	1.000	❌ REJECT	1-trial proxy PF=0.000 < 0.95, no improvement expected
15	`C8`	`lightgbm`	—	—	#7	k_loop_iter_15_5834af	`i_atr(k_high, k_low, k_close, 14) / (i_…`	✅	1.000	❌ REJECT	1-trial proxy PF=0.000 < 0.95, no improvement expected

📄 Report Markdown

# AUTOLOOP: zz_sylvain (15 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**: 228.65s
> **Total accepted**: 0
> **Total rejected**: 15

## Iterations 詳情

### Iteration 1
- DSL: `i_atr(k_high, k_low, k_close, 14) / k_close * (1 - i_rsi(k_close, 14) / 100) / (1 + abs(k_close.pct_change()))` (feature=k_iter1_i_atr_k_high_k_low_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: abs

### Iteration 2
- DSL: `k_high.pct_change(12) * (1 + hl_range) * (1 - dist_to_high) / (1 + i_atr(k_high, k_low, k_close, 14) / k_close)` (feature=k_iter2_k_high_pct_change_hl_range_dist_to_high)
- 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: `obv_rebased.pct_change() * (1 + hl_range) * (1 - price_position) / (1 + i_atr(k_high, k_low, k_close, 14) / k_close)` (feature=k_loop_iter_3_cca5e1)
- DSL valid: True
- FeatureSet candidate id: 1365
- 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: 1-trial proxy PF=0.000 < 0.95, no improvement expected

### Iteration 4
- DSL: `k_high.pct_change(12) * (1 + hl_range) * (1 - price_position)` (feature=k_iter4_k_high_pct_change_hl_range_price_position)
- 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_high.pct_change * (1 + hl_range) * (1 - price_position)

### Iteration 5
- DSL: `o_div_amplitude_12 * (1 - price_position) / (1 + atr_norm)` (feature=k_iter5_price_position_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: o_div_amplitude_12 * (1 - price_position) / (1 + atr_norm)

### Iteration 6
- DSL: `i_macd(k_close, 12, 26, 9) / (1 + i_atr(k_high, k_low, k_close, 14) / k_close) * (1 - price_position) * vol_ratio` (feature=k_loop_iter_6_58dc34)
- DSL valid: True
- FeatureSet candidate id: 1366
- 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: 1-trial proxy PF=0.000 < 0.95, no improvement expected

### Iteration 7
- DSL: `i_macd(k_close, 12, 26, 9) * (1 + hl_range) * o_div_amplitude_12 / (1 + atr_norm)` (feature=k_loop_iter_7_35271b)
- DSL valid: True
- FeatureSet candidate id: 1367
- 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: 1-trial proxy PF=0.000 < 0.95, no improvement expected

### Iteration 8
- DSL: `o_div_amplitude_12 / (1 + atr_norm) * (1 - price_position) * vol_ratio` (feature=k_loop_iter_8_7e64b2)
- DSL valid: True
- FeatureSet candidate id: 1368
- 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: 1-trial proxy PF=0.000 < 0.95, no improvement expected

### Iteration 9
- DSL: `i_macd(k_close, 6, 13, 5) * vol_ratio / (1 + atr_norm * 10) * (1 - price_position)` (feature=k_loop_iter_9_ee3d29)
- DSL valid: True
- FeatureSet candidate id: 1369
- 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: 1-trial proxy PF=0.000 < 0.95, no improvement expected

### Iteration 10
- DSL: `i_macd(k_close, 12, 26, 9) / (1 + i_atr(k_high, k_low, k_close, 14)) * vol_ratio` (feature=k_loop_iter_10_388e74)
- DSL valid: True
- FeatureSet candidate id: 1370
- 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: 1-trial proxy PF=0.000 < 0.95, no improvement expected

### Iteration 11
- DSL: `o_pct_change_12 * (k_close - i_ema(k_close, 50)) / (i_atr(k_high, k_low, k_close, 14) + 1e-10) * (1 - return_5) * (1 - price_position)` (feature=k_loop_iter_11_e1abc3)
- DSL valid: True
- FeatureSet candidate id: 1371
- 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: 1-trial proxy PF=0.000 < 0.95, no improvement expected

### Iteration 12
- DSL: `k_volume * (k_close - i_ema(k_close, 50)) / (k_volume.rolling(20).mean() * (1 + atr_norm)) * (1 - price_position)` (feature=k_loop_iter_12_d8ccea)
- DSL valid: True
- FeatureSet candidate id: 1372
- 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: 1-trial proxy PF=0.000 < 0.95, no improvement expected

### Iteration 13
- DSL: `i_rsi(k_close, 14) * (1 - price_position) * (k_volume / (k_volume.rolling(20).mean() + 1e-10)) / (1 + i_atr(k_high, k_low, k_close, 14) / k_close)` (feature=k_iter13_i_rsi_k_close_price_position_k_volume)
- 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 14
- DSL: `(o_div_amplitude_12 - o_div_amplitude_48) / (1 + atr_norm) * (1 - price_position)` (feature=k_loop_iter_14_f056e1)
- DSL valid: True
- FeatureSet candidate id: 1373
- 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: 1-trial proxy PF=0.000 < 0.95, no improvement expected

### Iteration 15
- DSL: `i_atr(k_high, k_low, k_close, 14) / (i_ema(k_close, 20) + 1e-10) * (1 - price_position) * vol_ratio` (feature=k_loop_iter_15_5834af)
- DSL valid: True
- FeatureSet candidate id: 1374
- 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: 1-trial proxy PF=0.000 < 0.95, no improvement expected

📊 Multi-Objective Metrics

Iteration 1: k_iter1_i_atr_k_high_k_low_k_close

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 2: k_iter2_k_high_pct_change_hl_range_dist_to_high

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 3: k_loop_iter_3_cca5e1

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 4: k_iter4_k_high_pct_change_hl_range_price_position

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 5: k_iter5_price_position_atr_norm

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 6: k_loop_iter_6_58dc34

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 7: k_loop_iter_7_35271b

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 8: k_loop_iter_8_7e64b2

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 9: k_loop_iter_9_ee3d29

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 10: k_loop_iter_10_388e74

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 11: k_loop_iter_11_e1abc3

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 12: k_loop_iter_12_d8ccea

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 13: k_iter13_i_rsi_k_close_price_position_k_volume

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 14: k_loop_iter_14_f056e1

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

Iteration 15: k_loop_iter_15_5834af

Y Label: C8 | Parent FS: #7

Objective mode: single

Scalar score: 0.0000

📁 Offline Report

Markdown backup: reports/autoloop/zz_sylvain/20260727_011349.md

Feature Iteration Run #1036: zz_sylvain

📊 Iterations

📄 Report Markdown

📊 Multi-Objective Metrics

Iteration 1: k_iter1_i_atr_k_high_k_low_k_close

Iteration 2: k_iter2_k_high_pct_change_hl_range_dist_to_high

Iteration 3: k_loop_iter_3_cca5e1

Iteration 4: k_iter4_k_high_pct_change_hl_range_price_position

Iteration 5: k_iter5_price_position_atr_norm

Iteration 6: k_loop_iter_6_58dc34

Iteration 7: k_loop_iter_7_35271b

Iteration 8: k_loop_iter_8_7e64b2

Iteration 9: k_loop_iter_9_ee3d29

Iteration 10: k_loop_iter_10_388e74

Iteration 11: k_loop_iter_11_e1abc3

Iteration 12: k_loop_iter_12_d8ccea

Iteration 13: k_iter13_i_rsi_k_close_price_position_k_volume

Iteration 14: k_loop_iter_14_f056e1

Iteration 15: k_loop_iter_15_5834af

📁 Offline Report