{
  "round": 14,
  "evidence_status": "Exact calculations on invented labels. No model inference, human annotation, or Encevra performance benchmark.",
  "checks_passed": 30,
  "checks": [
    "high_agreement_can_reverse_effect",
    "pooled_correction_fixes_overall_mean_not_difference",
    "three_snapshot_label_flips_erase_advantage",
    "four_snapshot_label_flips_reverse_advantage",
    "zero_budget_retains_point",
    "flip_dp_matches_exhaustive_budget_0",
    "flip_dp_matches_exhaustive_budget_1",
    "flip_dp_matches_exhaustive_budget_2",
    "flip_dp_matches_exhaustive_budget_3",
    "flip_dp_matches_exhaustive_budget_4",
    "flip_dp_matches_exhaustive_budget_5",
    "flip_dp_matches_exhaustive_budget_6",
    "unresolved_labels_preserve_full_denominator",
    "same_artifact_coupling_preserves_uncertainty",
    "coupling_reduces_width_sixty_percent",
    "paired_bounds_are_attained",
    "reversing_arms_reverses_bounds",
    "unchanged_labels_have_no_repair_or_damage",
    "equal_marginals_do_not_identify_transitions",
    "zero_denominator_is_explicit",
    "all_correct_has_null_repair_rate",
    "random_audit_estimator_is_exactly_unbiased",
    "finite_population_variance_matches_enumeration",
    "conservative_envelope_covers_this_finite_example",
    "complete_audit_eliminates_sampling_uncertainty",
    "disagreement_only_audit_is_not_representative",
    "empty_cohort_rejected",
    "empty_allowed_set_rejected",
    "nonbinary_label_rejected",
    "duplicate_audit_index_rejected"
  ],
  "differential_judge_bias": {
    "N_tasks": 100,
    "N_snapshot_labels": 200,
    "true_baseline_success": {
      "exact": "4/5",
      "decimal": 0.8
    },
    "true_candidate_success": {
      "exact": "39/50",
      "decimal": 0.78
    },
    "judged_baseline_success": {
      "exact": "4/5",
      "decimal": 0.8
    },
    "judged_candidate_success": {
      "exact": "83/100",
      "decimal": 0.83
    },
    "judge_label_accuracy": {
      "exact": "39/40",
      "decimal": 0.975
    },
    "true_difference": {
      "exact": "-1/50",
      "decimal": -0.02
    },
    "judged_difference": {
      "exact": "3/100",
      "decimal": 0.03
    },
    "pooled_specificity": {
      "exact": "37/42",
      "decimal": 0.8809523809523809
    },
    "pooled_corrected_baseline": {
      "exact": "143/185",
      "decimal": 0.772972972972973
    },
    "pooled_corrected_candidate": {
      "exact": "1493/1850",
      "decimal": 0.807027027027027
    },
    "pooled_corrected_difference": {
      "exact": "63/1850",
      "decimal": 0.03405405405405405
    },
    "flip_sensitivity": {
      "0": [
        {
          "exact": "3/100",
          "decimal": 0.03
        },
        {
          "exact": "3/100",
          "decimal": 0.03
        }
      ],
      "1": [
        {
          "exact": "1/50",
          "decimal": 0.02
        },
        {
          "exact": "1/25",
          "decimal": 0.04
        }
      ],
      "2": [
        {
          "exact": "1/100",
          "decimal": 0.01
        },
        {
          "exact": "1/20",
          "decimal": 0.05
        }
      ],
      "3": [
        {
          "exact": "0",
          "decimal": 0.0
        },
        {
          "exact": "3/50",
          "decimal": 0.06
        }
      ],
      "4": [
        {
          "exact": "-1/100",
          "decimal": -0.01
        },
        {
          "exact": "7/100",
          "decimal": 0.07
        }
      ],
      "5": [
        {
          "exact": "-1/50",
          "decimal": -0.02
        },
        {
          "exact": "2/25",
          "decimal": 0.08
        }
      ]
    }
  },
  "unresolved_pair_bounds": {
    "N": 100,
    "complete_case_difference": {
      "exact": "1/9",
      "decimal": 0.1111111111111111
    },
    "uncoupled": [
      {
        "exact": "0",
        "decimal": 0.0
      },
      {
        "exact": "1/5",
        "decimal": 0.2
      }
    ],
    "six_same_artifact_pairs": [
      {
        "exact": "3/50",
        "decimal": 0.06
      },
      {
        "exact": "7/50",
        "decimal": 0.14
      }
    ],
    "relative_width_reduction": {
      "exact": "3/5",
      "decimal": 0.6
    }
  },
  "same_marginals": {
    "initial_correctness": {
      "exact": "4/5",
      "decimal": 0.8
    },
    "terminal_correctness": {
      "exact": "4/5",
      "decimal": 0.8
    },
    "stable": {
      "repair_rate": [
        {
          "exact": "0",
          "decimal": 0.0
        },
        {
          "exact": "0",
          "decimal": 0.0
        }
      ],
      "damage_rate": [
        {
          "exact": "0",
          "decimal": 0.0
        },
        {
          "exact": "0",
          "decimal": 0.0
        }
      ],
      "repair_denominator_can_be_zero": false,
      "damage_denominator_can_be_zero": false,
      "possible_count_states": 1
    },
    "churn": {
      "repair_rate": [
        {
          "exact": "1",
          "decimal": 1.0
        },
        {
          "exact": "1",
          "decimal": 1.0
        }
      ],
      "damage_rate": [
        {
          "exact": "1/4",
          "decimal": 0.25
        },
        {
          "exact": "1/4",
          "decimal": 0.25
        }
      ],
      "repair_denominator_can_be_zero": false,
      "damage_denominator_can_be_zero": false,
      "possible_count_states": 1
    },
    "independence_implied_repair_rate": {
      "exact": "4/5",
      "decimal": 0.8
    },
    "independence_implied_damage_rate": {
      "exact": "1/5",
      "decimal": 0.2
    }
  },
  "random_audit": {
    "N": 8,
    "m": 4,
    "all_subsets": 70,
    "target": {
      "exact": "-1/8",
      "decimal": -0.125
    },
    "mean_estimate": {
      "exact": "-1/8",
      "decimal": -0.125
    },
    "exact_design_variance": {
      "exact": "1/16",
      "decimal": 0.0625
    },
    "census_envelope": {
      "lower": -0.125,
      "upper": -0.125,
      "radius": 0.0,
      "empty": false
    }
  },
  "biased_audit": {
    "selected_count": 3,
    "estimate_if_naively_treated_as_random": {
      "exact": "-97/100",
      "decimal": -0.97
    },
    "true_difference": {
      "exact": "-1/50",
      "decimal": -0.02
    }
  },
  "conservative_audit_halfwidths": {
    "m24": 1.1088852441549781,
    "m96": 0.5544426220774891,
    "interpretation": "Range-only, 95% audit-selection bound before intersection with [-1,1]; not future-task inference."
  }
}