{
 "full_121": {
  "name": "full 121 (AI-blinded + human-validated GT)",
  "n": 121,
  "tier": {
   "baseline": {
    "eligible": 59,
    "committed": 35,
    "correct": 29,
    "coverage": [
     35,
     59,
     0.593
    ],
    "committed_correct": [
     29,
     35,
     0.829
    ],
    "overall_correct": [
     29,
     59,
     0.492
    ]
   },
   "guard": {
    "eligible": 59,
    "committed": 26,
    "correct": 22,
    "coverage": [
     26,
     59,
     0.441
    ],
    "committed_correct": [
     22,
     26,
     0.846
    ],
    "overall_correct": [
     22,
     59,
     0.373
    ]
   }
  },
  "dependency": {
   "baseline": {
    "eligible": 71,
    "committed": 71,
    "correct": 14,
    "coverage": [
     71,
     71,
     1.0
    ],
    "committed_correct": [
     14,
     71,
     0.197
    ],
    "overall_correct": [
     14,
     71,
     0.197
    ]
   },
   "guard": {
    "eligible": 71,
    "committed": 48,
    "correct": 21,
    "coverage": [
     48,
     71,
     0.676
    ],
    "committed_correct": [
     21,
     48,
     0.438
    ],
    "overall_correct": [
     21,
     71,
     0.296
    ]
   }
  },
  "terminal": {
   "baseline": {
    "eligible": 59,
    "committed": 48,
    "correct": 16,
    "coverage": [
     48,
     59,
     0.814
    ],
    "committed_correct": [
     16,
     48,
     0.333
    ],
    "overall_correct": [
     16,
     59,
     0.271
    ]
   },
   "guard": {
    "eligible": 59,
    "committed": 36,
    "correct": 25,
    "coverage": [
     36,
     59,
     0.61
    ],
    "committed_correct": [
     25,
     36,
     0.694
    ],
    "overall_correct": [
     25,
     59,
     0.424
    ]
   }
  },
  "opaque": {
   "baseline": {
    "gt_opaque": 8,
    "TP": 8,
    "FN": 0,
    "FP": 23,
    "recall": [
     8,
     8,
     1.0
    ],
    "precision": [
     8,
     31,
     0.258
    ]
   },
   "guard": {
    "gt_opaque": 8,
    "TP": 1,
    "FN": 7,
    "FP": 0,
    "recall": [
     1,
     8,
     0.125
    ],
    "precision": [
     1,
     1,
     1.0
    ]
   }
  },
  "non_upgradeable": {
   "baseline": {
    "total": 50,
    "recognized_immutable": 9,
    "incorrectly_upgradeable": 29,
    "opaque": 11,
    "abstained": 1
   },
   "guard": {
    "total": 50,
    "recognized_immutable": "N/A (guard-directed has no immutable outcome; abstains)",
    "incorrectly_upgradeable": 0,
    "opaque": 0,
    "abstained": 50
   }
  },
  "abstention": {
   "baseline": {
    "committed_tier": 66,
    "IMMUTABLE": 23,
    "AUTHORITY_OPAQUE": 31,
    "ABSTAIN": 1,
    "INFRA_ERROR": 0,
    "n": 121
   },
   "guard": {
    "committed_tier": 26,
    "IMMUTABLE": 0,
    "AUTHORITY_OPAQUE": 1,
    "ABSTAIN": 94,
    "INFRA_ERROR": 0,
    "n": 121
   }
  },
  "paired_tier_over_T1T2T3": {
   "both_correct": 12,
   "baseline_only": 17,
   "guard_only": 10,
   "both_incorrect": 20,
   "discordant": 27,
   "mcnemar_chi2_cc": 1.333
  }
 },
 "upgradeable_71": {
  "name": "71 upgradeable (dev-comparable subset)",
  "n": 71,
  "tier": {
   "baseline": {
    "eligible": 59,
    "committed": 35,
    "correct": 29,
    "coverage": [
     35,
     59,
     0.593
    ],
    "committed_correct": [
     29,
     35,
     0.829
    ],
    "overall_correct": [
     29,
     59,
     0.492
    ]
   },
   "guard": {
    "eligible": 59,
    "committed": 26,
    "correct": 22,
    "coverage": [
     26,
     59,
     0.441
    ],
    "committed_correct": [
     22,
     26,
     0.846
    ],
    "overall_correct": [
     22,
     59,
     0.373
    ]
   }
  },
  "dependency": {
   "baseline": {
    "eligible": 71,
    "committed": 71,
    "correct": 14,
    "coverage": [
     71,
     71,
     1.0
    ],
    "committed_correct": [
     14,
     71,
     0.197
    ],
    "overall_correct": [
     14,
     71,
     0.197
    ]
   },
   "guard": {
    "eligible": 71,
    "committed": 48,
    "correct": 21,
    "coverage": [
     48,
     71,
     0.676
    ],
    "committed_correct": [
     21,
     48,
     0.438
    ],
    "overall_correct": [
     21,
     71,
     0.296
    ]
   }
  },
  "terminal": {
   "baseline": {
    "eligible": 59,
    "committed": 48,
    "correct": 16,
    "coverage": [
     48,
     59,
     0.814
    ],
    "committed_correct": [
     16,
     48,
     0.333
    ],
    "overall_correct": [
     16,
     59,
     0.271
    ]
   },
   "guard": {
    "eligible": 59,
    "committed": 36,
    "correct": 25,
    "coverage": [
     36,
     59,
     0.61
    ],
    "committed_correct": [
     25,
     36,
     0.694
    ],
    "overall_correct": [
     25,
     59,
     0.424
    ]
   }
  },
  "opaque": {
   "baseline": {
    "gt_opaque": 8,
    "TP": 8,
    "FN": 0,
    "FP": 12,
    "recall": [
     8,
     8,
     1.0
    ],
    "precision": [
     8,
     20,
     0.4
    ]
   },
   "guard": {
    "gt_opaque": 8,
    "TP": 1,
    "FN": 7,
    "FP": 0,
    "recall": [
     1,
     8,
     0.125
    ],
    "precision": [
     1,
     1,
     1.0
    ]
   }
  },
  "non_upgradeable": {
   "baseline": {
    "total": 0,
    "recognized_immutable": 0,
    "incorrectly_upgradeable": 0,
    "opaque": 0,
    "abstained": 0
   },
   "guard": {
    "total": 0,
    "recognized_immutable": "N/A (guard-directed has no immutable outcome; abstains)",
    "incorrectly_upgradeable": 0,
    "opaque": 0,
    "abstained": 0
   }
  },
  "abstention": {
   "baseline": {
    "committed_tier": 37,
    "IMMUTABLE": 14,
    "AUTHORITY_OPAQUE": 20,
    "ABSTAIN": 0,
    "INFRA_ERROR": 0,
    "n": 71
   },
   "guard": {
    "committed_tier": 26,
    "IMMUTABLE": 0,
    "AUTHORITY_OPAQUE": 1,
    "ABSTAIN": 44,
    "INFRA_ERROR": 0,
    "n": 71
   }
  },
  "paired_tier_over_T1T2T3": {
   "both_correct": 12,
   "baseline_only": 17,
   "guard_only": 10,
   "both_incorrect": 20,
   "discordant": 27,
   "mcnemar_chi2_cc": 1.333
  }
 },
 "human_confirmed_37": {
  "name": "37 human-confirmed subset",
  "n": 37,
  "tier": {
   "baseline": {
    "eligible": 16,
    "committed": 12,
    "correct": 11,
    "coverage": [
     12,
     16,
     0.75
    ],
    "committed_correct": [
     11,
     12,
     0.917
    ],
    "overall_correct": [
     11,
     16,
     0.688
    ]
   },
   "guard": {
    "eligible": 16,
    "committed": 4,
    "correct": 3,
    "coverage": [
     4,
     16,
     0.25
    ],
    "committed_correct": [
     3,
     4,
     0.75
    ],
    "overall_correct": [
     3,
     16,
     0.188
    ]
   }
  },
  "dependency": {
   "baseline": {
    "eligible": 28,
    "committed": 28,
    "correct": 7,
    "coverage": [
     28,
     28,
     1.0
    ],
    "committed_correct": [
     7,
     28,
     0.25
    ],
    "overall_correct": [
     7,
     28,
     0.25
    ]
   },
   "guard": {
    "eligible": 28,
    "committed": 17,
    "correct": 8,
    "coverage": [
     17,
     28,
     0.607
    ],
    "committed_correct": [
     8,
     17,
     0.471
    ],
    "overall_correct": [
     8,
     28,
     0.286
    ]
   }
  },
  "terminal": {
   "baseline": {
    "eligible": 16,
    "committed": 15,
    "correct": 9,
    "coverage": [
     15,
     16,
     0.938
    ],
    "committed_correct": [
     9,
     15,
     0.6
    ],
    "overall_correct": [
     9,
     16,
     0.562
    ]
   },
   "guard": {
    "eligible": 16,
    "committed": 7,
    "correct": 4,
    "coverage": [
     7,
     16,
     0.438
    ],
    "committed_correct": [
     4,
     7,
     0.571
    ],
    "overall_correct": [
     4,
     16,
     0.25
    ]
   }
  },
  "opaque": {
   "baseline": {
    "gt_opaque": 8,
    "TP": 8,
    "FN": 0,
    "FP": 7,
    "recall": [
     8,
     8,
     1.0
    ],
    "precision": [
     8,
     15,
     0.533
    ]
   },
   "guard": {
    "gt_opaque": 8,
    "TP": 1,
    "FN": 7,
    "FP": 0,
    "recall": [
     1,
     8,
     0.125
    ],
    "precision": [
     1,
     1,
     1.0
    ]
   }
  },
  "non_upgradeable": {
   "baseline": {
    "total": 9,
    "recognized_immutable": 2,
    "incorrectly_upgradeable": 3,
    "opaque": 3,
    "abstained": 1
   },
   "guard": {
    "total": 9,
    "recognized_immutable": "N/A (guard-directed has no immutable outcome; abstains)",
    "incorrectly_upgradeable": 0,
    "opaque": 0,
    "abstained": 9
   }
  },
  "abstention": {
   "baseline": {
    "committed_tier": 17,
    "IMMUTABLE": 4,
    "AUTHORITY_OPAQUE": 15,
    "ABSTAIN": 1,
    "INFRA_ERROR": 0,
    "n": 37
   },
   "guard": {
    "committed_tier": 4,
    "IMMUTABLE": 0,
    "AUTHORITY_OPAQUE": 1,
    "ABSTAIN": 32,
    "INFRA_ERROR": 0,
    "n": 37
   }
  },
  "paired_tier_over_T1T2T3": {
   "both_correct": 2,
   "baseline_only": 9,
   "guard_only": 1,
   "both_incorrect": 4,
   "discordant": 10,
   "mcnemar_chi2_cc": 4.9
  }
 },
 "confusion_baseline_R_B": {
  "T1": {
   "T1": 25,
   "T2": 0,
   "T3": 0,
   "AUTHORITY_OPAQUE": 2,
   "IMMUTABLE": 10,
   "ABSTAIN": 0,
   "INFRA_ERROR": 0
  },
  "T2": {
   "T1": 4,
   "T2": 3,
   "T3": 0,
   "AUTHORITY_OPAQUE": 2,
   "IMMUTABLE": 2,
   "ABSTAIN": 0,
   "INFRA_ERROR": 0
  },
  "T3": {
   "T1": 2,
   "T2": 0,
   "T3": 1,
   "AUTHORITY_OPAQUE": 7,
   "IMMUTABLE": 1,
   "ABSTAIN": 0,
   "INFRA_ERROR": 0
  },
  "AUTHORITY_OPAQUE": {
   "T1": 0,
   "T2": 0,
   "T3": 0,
   "AUTHORITY_OPAQUE": 8,
   "IMMUTABLE": 0,
   "ABSTAIN": 0,
   "INFRA_ERROR": 0
  },
  "UNRESOLVED": {
   "T1": 2,
   "T2": 0,
   "T3": 0,
   "AUTHORITY_OPAQUE": 1,
   "IMMUTABLE": 1,
   "ABSTAIN": 0,
   "INFRA_ERROR": 0
  },
  "NON_UPGRADEABLE": {
   "T1": 24,
   "T2": 5,
   "T3": 0,
   "AUTHORITY_OPAQUE": 11,
   "IMMUTABLE": 9,
   "ABSTAIN": 1,
   "INFRA_ERROR": 0
  }
 },
 "confusion_guard_native": {
  "T1": {
   "T1": 18,
   "T2": 1,
   "T3": 0,
   "AUTHORITY_OPAQUE": 0,
   "IMMUTABLE": 0,
   "ABSTAIN": 18,
   "INFRA_ERROR": 0
  },
  "T2": {
   "T1": 3,
   "T2": 4,
   "T3": 0,
   "AUTHORITY_OPAQUE": 0,
   "IMMUTABLE": 0,
   "ABSTAIN": 4,
   "INFRA_ERROR": 0
  },
  "T3": {
   "T1": 0,
   "T2": 0,
   "T3": 0,
   "AUTHORITY_OPAQUE": 0,
   "IMMUTABLE": 0,
   "ABSTAIN": 11,
   "INFRA_ERROR": 0
  },
  "AUTHORITY_OPAQUE": {
   "T1": 0,
   "T2": 0,
   "T3": 0,
   "AUTHORITY_OPAQUE": 1,
   "IMMUTABLE": 0,
   "ABSTAIN": 7,
   "INFRA_ERROR": 0
  },
  "UNRESOLVED": {
   "T1": 0,
   "T2": 0,
   "T3": 0,
   "AUTHORITY_OPAQUE": 0,
   "IMMUTABLE": 0,
   "ABSTAIN": 4,
   "INFRA_ERROR": 0
  },
  "NON_UPGRADEABLE": {
   "T1": 0,
   "T2": 0,
   "T3": 0,
   "AUTHORITY_OPAQUE": 0,
   "IMMUTABLE": 0,
   "ABSTAIN": 50,
   "INFRA_ERROR": 0
  }
 },
 "baseline_native_immutability": {
  "nonupgradeable_controlModel": {
   "IMMUTABLE_BY_RENUNCIATION": 1,
   "UPGRADEABLE_PROXY": 2,
   "IMMUTABLE_CORE_WITH_GOVERNANCE": 39,
   "IMMUTABLE_NO_GOVERNANCE_FOUND": 8
  },
  "native_IMMUTABLE_star_on_50_nonupg": 48,
  "native_IMMUTABLE_star_on_71_upgradeable_false_positives": 17,
  "note": "baseline native controlModel flags 48/50 non-upgradeable as IMMUTABLE_*; the R_B-derived tier reclassifies 29 of the 50 as T1/T2/T3 because IMMUTABLE_CORE_WITH_GOVERNANCE + a resolvable owner maps to a tier. The false-upgradeable tier count is therefore largely an R_B adapter artifact, not a baseline native miss. Tier is secondary for the baseline per the protocol."
 },
 "guard_abstain_reasons_on_71_upgradeable": {
  "PARTIAL": 12,
  "PRIVILEGED_SINK_UNRESOLVED": 12,
  "GUARD_UNRESOLVED": 10,
  "UNRESOLVED": 9,
  "SOURCE_UNAVAILABLE": 1,
  "OPAQUE": 1
 },
 "mechanism_breakdown_tier_over_T1T2T3": {
  "UUPS/ERC1967": {
   "n": 16,
   "gt_T123": 13,
   "baseline_tier_correct": 9,
   "guard_tier_correct": 4
  },
  "comptroller/unitroller": {
   "n": 5,
   "gt_T123": 5,
   "baseline_tier_correct": 1,
   "guard_tier_correct": 0
  },
  "custom/other": {
   "n": 8,
   "gt_T123": 1,
   "baseline_tier_correct": 1,
   "guard_tier_correct": 1
  },
  "diamond": {
   "n": 7,
   "gt_T123": 7,
   "baseline_tier_correct": 7,
   "guard_tier_correct": 0
  },
  "non-upgradeable": {
   "n": 50,
   "gt_T123": 0,
   "baseline_tier_correct": 0,
   "guard_tier_correct": 0
  },
  "registry/aave-immutable-admin": {
   "n": 2,
   "gt_T123": 2,
   "baseline_tier_correct": 0,
   "guard_tier_correct": 0
  },
  "transparent": {
   "n": 33,
   "gt_T123": 31,
   "baseline_tier_correct": 11,
   "guard_tier_correct": 17
  }
 },
 "development_frame_reference": {
  "source": "DEV_FRAME_RESULTS_v2.md (frozen bc6831de over 103 upgradeable dev deployments)",
  "dependency_recovered": "74/103=0.72",
  "committed_tier": "34/103=0.33",
  "committed_tier_correct": "29/34=0.85",
  "opaque_recall": "4/4=1.0"
 },
 "dev_vs_holdout_guard_directed": {
  "dependency_coverage": {
   "dev": "74/103=0.72",
   "holdout_upgradeable": "48/71=0.68"
  },
  "committed_tier_rate": {
   "dev": "34/103=0.33",
   "holdout_upgradeable": "26/71=0.37"
  },
  "committed_tier_correctness": {
   "dev": "29/34=0.85",
   "holdout_over_GT_T1T2T3": "22/26=0.85"
  },
  "opaque_recall": {
   "dev": "4/4=1.0",
   "holdout": "1/8=0.125"
  },
  "verdict": "committed-tier precision and dependency-recovery coverage generalized; opacity recognition did not; overall tier coverage remained low and did not exceed the fixed-surface baseline on aggregate tier-correctness."
 }
}