exp 11 (sp5d-moment-features): variant C 35-field workflow + result NEGATIVE vs generic-only 19

workflows/README-moments.md

@@ -0,0 +1,20 @@
+# exp/10 sp5d-moment-features
+
+Variant C: generic-only 19 + 16 new moment/volatility families (skew, kurt,
+DSV+ratios, max_up/down, rv_ac1, rv_cv_22, sig lag-5). 35 sp_* fields, ou/hmm excluded.
+
+Run a3f7d1d40c3d4b839314fcf5b40f9b08 (tac-rd-moments / exp 12) — FINISHED.
+
+## Result: NEGATIVE (regression vs generic-only baseline)
+
+| Metric | generic-only 19 (run 7b1e797) | +moments 35 (run a3f7d1d) |
+|---|---|---|
+| Rank IC   | 0.0635 | 0.0466 |
+| Rank ICIR | 0.276  | 0.183  |
+| L-S Sharpe | 2.55  | 1.44  |
+| net excess (cost) | +3.1% IR 0.28 | -16.2% IR -1.57 |
+| MDD | -7.3% | -11.1% |
+
+Same failure mode as ou/hmm in exp 9: adding cross-sectional moment features
+to the 50-name panel degrades the rank signal. Generic-only 19 remains the
+best configuration. No further moment-family variants planned.

workflows/ablate_generic_moments.yaml

@@ -1,24 +1,22 @@
 # -----------------------------------------------------------------------------
-# ISOLATION: multi-seed RankIC ensemble, ablate-B generic-only feature set.
+# VARIANT C (generic + moments): keeps the winning generic-only 19-field set
-#
+# (jump,har,trend,hurst,signature) and adds the NEW generic moment families the
-# Isolates the ensemble effect on the SP-5d rank signal. Same panel, segments,
+# engine now exposes:
-# history (full backfilled 2016+) and feature set as the exp-9 ablate-B winner
+#   - realized skewness / kurtosis      (sp_rskew_5, sp_rskew_22, sp_rkurt_5, sp_rkurt_22)
-# (generic-only sp_* families: jump,har,trend,hurst,signature), but replaces the
+#   - downside semi-variance + ratios   (sp_dsv_1/5/22, sp_dsv_ratio_1/5/22)
-# single RankICLGBModel with a 5-seed RankICEnsembleLGBModel (42,7,2026,99,123)
+#   - signed max moves                  (sp_max_up, sp_max_down)
-# that averages per-day predictions.
+#   - RV autocorr / vol-of-vol          (sp_rv_ac1, sp_rv_cv_22)
-#
+#   - longer-lag signature terms        (sp_sig_level2_*_5)
-# Differs from exp-15 (tac-rd-rank-ensemble, mlflow exp 15) ONLY by dropping the
+# Drops the model-specific ou/hmm families (they scored high in importance but
-# TA subset (rsi_14,roc_10,macd_hist,willr_14,atr_14) and the inter-asset xr_*
+# hurt the rank dimension in the all-24 run). Same panel/model as baseline.
-# features, so any change vs exp-15 is attributable to the feature set alone,
-# and any change vs exp-9 is attributable to the ensemble + full history alone.
 #
 # Run:
-#   rd_run_workflow config_path=experiments/workflows/exp12_isolation_ensemble.yaml \
+#   rd_run_workflow config_path=experiments/workflows/ablate_generic_moments.yaml \
-#       experiment_name=tac-rd-rank-ensemble-isolated
+#       experiment_name=tac-rd-moments
 # -----------------------------------------------------------------------------
 {%- set LAKE = TAC_LAKE_DIR %}
 {%- set UNIVERSE = "SPY,QQQ,DIA,IWM,MDY,VTI,VOO,VEA,VWO,VT,EFA,EEM,TLT,IEF,SHY,AGG,BND,LQD,HYG,JNK,EMB,GLD,SLV,USO,UNG,DBA,DBC,XLK,XLF,XLE,XLV,XLI,XLY,XLP,XLU,XLB,XLRE,ARKK,SMH,SOXX,IBB,XBI,ITA,XAR,ICLN,TAN,FDN,IGV,ESPO,REM" %}
-{%- set SP_FIELDS = "sp_ret,sp_jump_ratio,sp_jump_flag,sp_jump_tail,sp_max_move,sp_rv1,sp_rv5,sp_rv22,sp_vol_ratio_5_22,sp_vol_ratio_1_22,sp_trend_slope_5,sp_trend_slope_20,sp_trend_slope_60,sp_logp,sp_hurst_exponent,sp_sig_level1_lead,sp_sig_level1_lag,sp_sig_level2_lead_lag,sp_sig_level2_lag_lead" %}
+{%- set SP_FIELDS = "sp_ret,sp_jump_ratio,sp_jump_flag,sp_jump_tail,sp_max_move,sp_max_up,sp_max_down,sp_rv1,sp_rv5,sp_rv22,sp_vol_ratio_5_22,sp_vol_ratio_1_22,sp_rv_ac1,sp_rv_cv_22,sp_trend_slope_5,sp_trend_slope_20,sp_trend_slope_60,sp_logp,sp_hurst_exponent,sp_sig_level1_lead,sp_sig_level1_lag,sp_sig_level2_lead_lag,sp_sig_level2_lag_lead,sp_sig_level2_lead_lag_5,sp_sig_level2_lag_lead_5,sp_rskew_5,sp_rskew_22,sp_rkurt_5,sp_rkurt_22,sp_dsv_1,sp_dsv_5,sp_dsv_22,sp_dsv_ratio_1,sp_dsv_ratio_5,sp_dsv_ratio_22" %}
 
 qlib_init:
     provider_uri: "{{ LAKE }}"
@@ -47,13 +45,13 @@ qlib_init:
         class: MLflowExpManager
         module_path: qlib.workflow.expm
         kwargs:
-            uri: "sqlite:///mlruns.db"
+            uri: "sqlite:///{{ LAKE }}/mlruns.db"
-            default_exp_name: "tac-rd-rank-ensemble-isolated"
+            default_exp_name: "tac-rd-moments"
 
 task:
     model:
-        class: RankICEnsembleLGBModel
+        class: RankICLGBModel
-        module_path: tac_qlib.contrib.model.rank_ensemble
+        module_path: tac_qlib.contrib.model.rank_gbdt
         kwargs:
             loss: mse
             learning_rate: 0.02
@@ -68,7 +66,7 @@ task:
             subsample_freq: 1
             reg_alpha: 0.1
             reg_lambda: 1.0
-            seeds: "42,7,2026,99,123"
+            seed: 42
 
     dataset:
         class: DatasetH
@@ -80,8 +78,8 @@ task:
                 kwargs:
                     instruments: "{{ UNIVERSE }}"
                     start_time: 2015-01-03
-                    end_time: 2026-08-14
+                    end_time: 2026-08-10
-                    fit_start_time: 2016-01-04
+                    fit_start_time: 2015-01-03
                     fit_end_time: 2025-09-01
                     freq: day
                     lake_root: "{{ LAKE }}"
@@ -100,7 +98,7 @@ task:
                         - class: Fillna
                           kwargs: {}
             segments:
-                train: [2016-01-04, 2025-09-01]
+                train: [2015-01-03, 2025-09-01]
                 valid: [2025-09-03, 2026-01-03]
                 test: [2026-01-04, 2026-08-10]