{ "analysis": "pre-committed repeated scaffold-disjoint split robustness check for Delta R2", "gate": { "checks": { "delta_r2": { "abs_error": 0.0, "expected": 0.15844577805518212, "observed": 0.15844577805518212, "passes": true, "tolerance": 0.002 }, "rho_base": { "abs_error": 0.0, "expected": 0.6130893034049266, "observed": 0.6130893034049266, "passes": true, "tolerance": 0.002 }, "rho_struct": { "abs_error": 0.0, "expected": 0.6633801733693918, "observed": 0.6633801733693918, "passes": true, "tolerance": 0.002 } }, "seed": 20260710, "status": "pass" }, "model_specs": { "baseline": "baseline_physical_gbm", "fixed_hyperparameters": { "imputer": "SimpleImputer(strategy='median')", "l2_regularization": 0.1, "learning_rate": 0.04, "max_iter": 200, "model": "HistGradientBoostingRegressor", "random_state": 17 }, "source": "scripts/train_substantivity.py fit_predict(model_name='gbm')", "structure": "structure_full_gbm" }, "n_model_ready": 170, "pimt_version": "v8", "seed_list_precommitted": [ 20260710, 20260711, 20260712, 20260713, 20260714, 20260715, 20260716, 20260717, 20260718, 20260719, 20260720, 20260721, 20260722, 20260723, 20260724, 20260725, 20260726, 20260727, 20260728, 20260729 ], "splits": [ { "delta_r2": 0.15844577805518212, "delta_rho": 0.05029086996446519, "n_test": 34, "n_train": 136, "r2_base": 0.24057748815493107, "r2_struct": 0.3990232662101132, "rho_base": 0.6130893034049266, "rho_struct": 0.6633801733693918, "scaffold_overlap": [], "seed": 20260710, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.10894368939423504, "delta_rho": -0.02142567109834309, "n_test": 34, "n_train": 136, "r2_base": 0.0753173569506469, "r2_struct": 0.18426104634488194, "rho_base": 0.3374255928820426, "rho_struct": 0.3159999217836995, "scaffold_overlap": [], "seed": 20260711, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": false, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.07229900764275499, "delta_rho": 0.054794018070953965, "n_test": 34, "n_train": 136, "r2_base": 0.03474908347621042, "r2_struct": 0.1070480911189654, "rho_base": 0.3116758043837438, "rho_struct": 0.3664698224546978, "scaffold_overlap": [], "seed": 20260712, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": -0.07932888055232179, "delta_rho": -0.18782560792479713, "n_test": 34, "n_train": 136, "r2_base": 0.009371529716526661, "r2_struct": -0.06995735083579513, "rho_base": 0.22892683918665618, "rho_struct": 0.04110123126185904, "scaffold_overlap": [], "seed": 20260713, "structure_r2_gt_baseline_r2": false, "structure_rho_gt_baseline_rho": false, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.17795748043035797, "delta_rho": 0.15765396639828383, "n_test": 34, "n_train": 136, "r2_base": -0.025625114290944717, "r2_struct": 0.15233236613941326, "rho_base": 0.3148971774546397, "rho_struct": 0.4725511438529235, "scaffold_overlap": [], "seed": 20260714, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.3651105688046139, "delta_rho": 0.5723722889991607, "n_test": 34, "n_train": 136, "r2_base": -0.14781647665809938, "r2_struct": 0.21729409214651452, "rho_base": -0.05458199713591954, "rho_struct": 0.5177902918632412, "scaffold_overlap": [], "seed": 20260715, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.1593569048675234, "delta_rho": 0.26000963937387966, "n_test": 34, "n_train": 136, "r2_base": -0.31410214728889585, "r2_struct": -0.15474524242137244, "rho_base": 0.021809248206604608, "rho_struct": 0.2818188875804843, "scaffold_overlap": [], "seed": 20260716, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.18827443498189356, "delta_rho": 0.03148922432055268, "n_test": 34, "n_train": 136, "r2_base": -0.10861175558073333, "r2_struct": 0.07966267940116023, "rho_base": 0.2930425572131179, "rho_struct": 0.32453178153367057, "scaffold_overlap": [], "seed": 20260717, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.06412327813812702, "delta_rho": 0.04651519519605235, "n_test": 34, "n_train": 136, "r2_base": 0.009225601013266416, "r2_struct": 0.07334887915139343, "rho_base": 0.20013954931621014, "rho_struct": 0.2466547445122625, "scaffold_overlap": [], "seed": 20260718, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.11669355822914607, "delta_rho": 0.2230183252778165, "n_test": 34, "n_train": 136, "r2_base": 0.11300301261267753, "r2_struct": 0.2296965708418236, "rho_base": 0.29368400613333734, "rho_struct": 0.5167023314111538, "scaffold_overlap": [], "seed": 20260719, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.1615774244680297, "delta_rho": 0.1556327648303413, "n_test": 34, "n_train": 136, "r2_base": -0.0004697438215104377, "r2_struct": 0.16110768064651926, "rho_base": 0.2151720464246368, "rho_struct": 0.3708048112549781, "scaffold_overlap": [], "seed": 20260720, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.13393454244880798, "delta_rho": 0.011281368170483141, "n_test": 34, "n_train": 136, "r2_base": 0.12810201588115622, "r2_struct": 0.2620365583299642, "rho_base": 0.305951850465261, "rho_struct": 0.31723321863574416, "scaffold_overlap": [], "seed": 20260721, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.1394537063219743, "delta_rho": 0.10944078731825996, "n_test": 34, "n_train": 136, "r2_base": 0.1597453729206536, "r2_struct": 0.2991990792426279, "rho_base": 0.4096488396211531, "rho_struct": 0.5190896269394131, "scaffold_overlap": [], "seed": 20260722, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.03749212355150777, "delta_rho": -0.020460924496539412, "n_test": 34, "n_train": 136, "r2_base": 0.22792739917286775, "r2_struct": 0.2654195227243755, "rho_base": 0.364431504041778, "rho_struct": 0.3439705795452386, "scaffold_overlap": [], "seed": 20260723, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": false, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.027406566812759103, "delta_rho": 0.1042355837894588, "n_test": 34, "n_train": 136, "r2_base": 0.0790457059498223, "r2_struct": 0.1064522727625814, "rho_base": 0.25539472324636275, "rho_struct": 0.35963030703582155, "scaffold_overlap": [], "seed": 20260724, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.18287532371958493, "delta_rho": 0.1886042686910458, "n_test": 34, "n_train": 136, "r2_base": -0.001996189735690379, "r2_struct": 0.18087913398389455, "rho_base": 0.22568044935707093, "rho_struct": 0.41428471804811673, "scaffold_overlap": [], "seed": 20260725, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.2023991030263118, "delta_rho": 0.3463187369730992, "n_test": 34, "n_train": 136, "r2_base": 0.043956322103222534, "r2_struct": 0.24635542512953434, "rho_base": 0.25719833423563027, "rho_struct": 0.6035170712087294, "scaffold_overlap": [], "seed": 20260726, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.05346556978823658, "delta_rho": -0.037189855887579226, "n_test": 34, "n_train": 136, "r2_base": 0.2499476956003236, "r2_struct": 0.30341326538856017, "rho_base": 0.5136834171866699, "rho_struct": 0.47649356129909065, "scaffold_overlap": [], "seed": 20260727, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": false, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.16469614036100466, "delta_rho": 0.1832511429802354, "n_test": 34, "n_train": 136, "r2_base": 0.057282262855799515, "r2_struct": 0.22197840321680418, "rho_base": 0.2097134137159253, "rho_struct": 0.3929645566961607, "scaffold_overlap": [], "seed": 20260728, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 }, { "delta_r2": 0.12786896886633592, "delta_rho": 0.24574302338227827, "n_test": 34, "n_train": 136, "r2_base": 0.17693494806547427, "r2_struct": 0.3048039169318102, "rho_base": 0.3036752525270744, "rho_struct": 0.5494182759093527, "scaffold_overlap": [], "seed": 20260729, "structure_r2_gt_baseline_r2": true, "structure_rho_gt_baseline_rho": true, "test_scaffolds": 34, "train_scaffolds": 49 } ], "summary": { "delta_r2": { "fraction_wins": 0.95, "mean": 0.12815226446780326, "n_splits": 20, "percentile_2_5_97_5": [ -0.028629543053908364, 0.2878226225599202 ], "wins": 19 }, "delta_rho": { "fraction_wins": 0.8, "mean": 0.1236874572164554, "n_splits": 20, "percentile_2_5_97_5": [ -0.11627362570711862, 0.4649968517867812 ], "wins": 16 } }, "target": "Poucher measured log10 coefficient only; no inferred, OpenPOM, or volatility-derived substantivity labels used.", "test_fraction": 0.2 }