Files
Open_Ultra_SAST/benchmarks/experiments/exp-004-pooled-block-gate/result.json
T

823 lines
22 KiB
JSON

{
"arms": {
"A": {
"label": "today's rule -- Platt map cross-fitted per family, BLOCK only when the calibration check holds, BLOCK point nested on the other folds' calibrated p",
"rule": {
"name": "calibrated_block"
}
},
"B": {
"label": "pooled gate -- one t_block over all families on a per-family Platt score (training folds, inner cross-fit), t_quiet per family at 90% recall, family floor 10 clean held-out flags",
"rule": {
"floor": 10,
"name": "pooled_block_gate",
"normalise": "platt",
"precision": 0.95,
"recall": 0.9,
"select_on": "paired"
}
}
},
"code_commit": "90a2642019df414cb67b5dcea6e4ec235f35650a",
"decision": "inconclusive",
"experiment": "exp-004-pooled-block-gate",
"families": {
"access_control": {
"A": {
"calibration_holds": false,
"ece": 0.0715,
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 28,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 28,
"true_flagged": 0,
"units": 56
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 37,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 38,
"true_flagged": 0,
"units": 66
},
"reason_by_fold": {
"outer-0": "calibration",
"outer-1": "calibration",
"outer-2": "calibration",
"outer-3": "calibration",
"outer-4": "calibration"
},
"thresholds_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
}
},
"B": {
"floor": {
"errors": 0,
"flagged": 0,
"floor": 10,
"may_block": false
},
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 28,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 28,
"true_flagged": 0,
"units": 56
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 37,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 38,
"true_flagged": 0,
"units": 66
},
"quiet_recall": {
"paired": 0.8929,
"pooled": 0.9211
},
"t_block_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
},
"t_quiet_by_fold": {
"outer-0": 0.44900641988143913,
"outer-1": 0.3723384028044276,
"outer-2": 0.4202772526614184,
"outer-3": 0.4481849709797749,
"outer-4": 0.36346174463707154
}
},
"capacity": {
"evaluated": 66,
"excluded_replay_miss": 0,
"not_run": 0,
"paired_positives_by_training_folds": {
"outer-0": 22,
"outer-1": 22,
"outer-2": 22,
"outer-3": 23,
"outer-4": 23
}
}
},
"deserialization": {
"A": {
"calibration_holds": false,
"ece": 0.0725,
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 25,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 25,
"true_flagged": 0,
"units": 50
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 39,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 39,
"true_flagged": 0,
"units": 64
},
"reason_by_fold": {
"outer-0": "calibration",
"outer-1": "calibration",
"outer-2": "calibration",
"outer-3": "calibration",
"outer-4": "calibration"
},
"thresholds_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
}
},
"B": {
"floor": {
"errors": 0,
"flagged": 0,
"floor": 10,
"may_block": false
},
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 25,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 25,
"true_flagged": 0,
"units": 50
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 39,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 39,
"true_flagged": 0,
"units": 64
},
"quiet_recall": {
"paired": 0.88,
"pooled": 0.9231
},
"t_block_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
},
"t_quiet_by_fold": {
"outer-0": 0.17507194490500488,
"outer-1": 0.22766488322270723,
"outer-2": 0.2163745870743959,
"outer-3": 0.3112201878893598,
"outer-4": 0.363243722721776
}
},
"capacity": {
"evaluated": 64,
"excluded_replay_miss": 0,
"not_run": 0,
"paired_positives_by_training_folds": {
"outer-0": 20,
"outer-1": 20,
"outer-2": 20,
"outer-3": 20,
"outer-4": 20
}
}
},
"injection": {
"A": {
"calibration_holds": false,
"ece": 0.1042,
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 41,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 53,
"true_flagged": 0,
"units": 106
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 49,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 78,
"true_flagged": 0,
"units": 149
},
"reason_by_fold": {
"outer-0": "calibration",
"outer-1": "calibration",
"outer-2": "calibration",
"outer-3": "calibration",
"outer-4": "calibration"
},
"thresholds_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
}
},
"B": {
"floor": {
"errors": 0,
"flagged": 0,
"floor": 10,
"may_block": false
},
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 41,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 53,
"true_flagged": 0,
"units": 106
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 49,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 78,
"true_flagged": 0,
"units": 149
},
"quiet_recall": {
"paired": 0.8679,
"pooled": 0.8974
},
"t_block_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
},
"t_quiet_by_fold": {
"outer-0": 0.287983504097999,
"outer-1": 0.17524866355153343,
"outer-2": 0.12918182665453967,
"outer-3": 0.14521040470628938,
"outer-4": 0.23005714687697673
}
},
"capacity": {
"evaluated": 149,
"excluded_replay_miss": 0,
"not_run": 0,
"paired_positives_by_training_folds": {
"outer-0": 38,
"outer-1": 44,
"outer-2": 45,
"outer-3": 43,
"outer-4": 42
}
}
},
"output_encoding": {
"A": {
"calibration_holds": false,
"ece": 0.0736,
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 28,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 28,
"true_flagged": 0,
"units": 56
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 39,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 41,
"true_flagged": 0,
"units": 69
},
"reason_by_fold": {
"outer-0": "calibration",
"outer-1": "calibration",
"outer-2": "calibration",
"outer-3": "calibration",
"outer-4": "calibration"
},
"thresholds_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
}
},
"B": {
"floor": {
"errors": 0,
"flagged": 0,
"floor": 10,
"may_block": false
},
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 28,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 28,
"true_flagged": 0,
"units": 56
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 39,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 41,
"true_flagged": 0,
"units": 69
},
"quiet_recall": {
"paired": 0.8571,
"pooled": 0.9024
},
"t_block_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
},
"t_quiet_by_fold": {
"outer-0": 0.3723166587726901,
"outer-1": 0.5329072757249385,
"outer-2": 0.3785066578507705,
"outer-3": 0.4389011844848602,
"outer-4": 0.43855966643440264
}
},
"capacity": {
"evaluated": 69,
"excluded_replay_miss": 0,
"not_run": 0,
"paired_positives_by_training_folds": {
"outer-0": 23,
"outer-1": 22,
"outer-2": 22,
"outer-3": 22,
"outer-4": 23
}
}
},
"path": {
"A": {
"calibration_holds": false,
"ece": 0.0621,
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 29,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 29,
"true_flagged": 0,
"units": 58
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 37,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 37,
"true_flagged": 0,
"units": 66
},
"reason_by_fold": {
"outer-0": "calibration",
"outer-1": "calibration",
"outer-2": "calibration",
"outer-3": "calibration",
"outer-4": "calibration"
},
"thresholds_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
}
},
"B": {
"floor": {
"errors": 0,
"flagged": 0,
"floor": 10,
"may_block": false
},
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 29,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 29,
"true_flagged": 0,
"units": 58
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 37,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 37,
"true_flagged": 0,
"units": 66
},
"quiet_recall": {
"paired": 0.8621,
"pooled": 0.8919
},
"t_block_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
},
"t_quiet_by_fold": {
"outer-0": 0.4680507207548809,
"outer-1": 0.3955717127313657,
"outer-2": 0.43694703358760895,
"outer-3": 0.4450583743716199,
"outer-4": 0.46396657833309796
}
},
"capacity": {
"evaluated": 66,
"excluded_replay_miss": 0,
"not_run": 0,
"paired_positives_by_training_folds": {
"outer-0": 23,
"outer-1": 23,
"outer-2": 24,
"outer-3": 23,
"outer-4": 23
}
}
},
"untrusted_destination": {
"A": {
"calibration_holds": false,
"ece": 0.1143,
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 31,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 32,
"true_flagged": 0,
"units": 64
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 49,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 50,
"true_flagged": 0,
"units": 87
},
"reason_by_fold": {
"outer-0": "calibration",
"outer-1": "calibration",
"outer-2": "calibration",
"outer-3": "calibration",
"outer-4": "calibration"
},
"thresholds_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
}
},
"B": {
"floor": {
"errors": 0,
"flagged": 0,
"floor": 10,
"may_block": false
},
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 31,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 32,
"true_flagged": 0,
"units": 64
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 49,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 50,
"true_flagged": 0,
"units": 87
},
"quiet_recall": {
"paired": 1.0,
"pooled": 1.0
},
"t_block_by_fold": {
"outer-0": null,
"outer-1": null,
"outer-2": null,
"outer-3": null,
"outer-4": null
},
"t_quiet_by_fold": {
"outer-0": 0.13676746287025585,
"outer-1": 0.11255775703980551,
"outer-2": 0.13763698150768486,
"outer-3": 0.09787938647374639,
"outer-4": 0.13763698150768486
}
},
"capacity": {
"evaluated": 87,
"excluded_replay_miss": 0,
"not_run": 0,
"paired_positives_by_training_folds": {
"outer-0": 26,
"outer-1": 26,
"outer-2": 24,
"outer-3": 26,
"outer-4": 26
}
}
}
},
"hypothesis": "One BLOCK threshold pooled across families on a per-family Platt-normalised score, chosen on training folds only as the lowest one whose Wilson 95% lower bound on flagged precision is >= 0.95, reaches that bound on held-out repositories' paired units, where per-family thresholds (exp-003) could not for lack of flags; a per-family floor (>= 10 held-out flags, none wrong) keeps a family that the pooling masks from blocking.\n",
"manifest_digest": "87fceb6580c6de629d6a6089eb5888cdfc4df0f791576184f987c14c192d85f5",
"min_flagged_for_bound": 73,
"overall": {
"A": {
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 182,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 195,
"true_flagged": 0,
"units": 390
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 220,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 283,
"true_flagged": 0,
"units": 501
}
},
"B": {
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 182,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 195,
"true_flagged": 0,
"units": 390
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 220,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 283,
"true_flagged": 0,
"units": 501
}
}
},
"overall_after_floor": {
"B": {
"paired": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 182,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 195,
"true_flagged": 0,
"units": 390
},
"pooled": {
"coverage": 0.0,
"flagged": 0,
"flagged_precision": null,
"flagged_precision_ci": null,
"flagged_precision_lb": null,
"groups": 220,
"meets_target_lb": false,
"meets_target_point": false,
"positives": 283,
"true_flagged": 0,
"units": 501
}
}
},
"populations": {
"primary": "paired",
"secondary": [
"pooled"
]
},
"primary": [
{
"name": "flagged_precision_lb",
"target": 0.95
}
],
"reason": "B offers no BLOCK on held-out paired units; the incumbent stays",
"resamples": 2000,
"secondary": [
{
"name": "flagged_precision"
},
{
"name": "coverage"
},
{
"name": "flagged"
},
{
"name": "quiet_recall"
}
],
"spend_usd": {
"score": 0.0
},
"target_precision": 0.95,
"units": {
"by_label": {
"0": 218,
"1": 283
},
"evaluated": 501,
"excluded_replay_miss": 0,
"frozen": 501,
"groups": 220,
"not_run": 0,
"paired_evaluated": 390,
"unsure": 146
}
}