{
  "generated": "2026-07-01",
  "source": "runs/swarm-study1/events.jsonl (original run; no new API calls)",
  "bootstrap": {
    "resamples": 10000,
    "seed": 20260701,
    "method": "paired item-level percentile"
  },
  "arms": {
    "swarm_round0": {
      "n": 66,
      "accuracy": 0.8485,
      "mean_confidence": 0.8939,
      "overconfidence": 0.0455,
      "ece": 0.0556,
      "mce": 0.0833,
      "brier": 0.0993,
      "auc": 0.7812
    },
    "swarm_postdebate": {
      "n": 66,
      "accuracy": 0.8182,
      "mean_confidence": 0.9293,
      "overconfidence": 0.1111,
      "ece": 0.1111,
      "mce": 0.3333,
      "brier": 0.1279,
      "auc": 0.7407
    },
    "baseline_first3": {
      "n": 66,
      "accuracy": 0.8182,
      "mean_confidence": 0.8939,
      "overconfidence": 0.0758,
      "ece": 0.0758,
      "mce": 0.3333,
      "brier": 0.1162,
      "auc": 0.7639
    },
    "baseline_full": {
      "n": 66,
      "accuracy": 0.8182,
      "mean_confidence": 0.8702,
      "overconfidence": 0.052,
      "ece": 0.109,
      "mce": 0.4286,
      "brier": 0.1,
      "auc": 0.8295
    }
  },
  "baseline_full_auc_coarsened_to_3_levels": 0.7639,
  "round0_unanimity": {
    "unanimous_items": 49,
    "n": 66,
    "share": 0.7424
  },
  "accuracy_discordant_items_swarm_vs_baseline": 2,
  "deltas_ci95": {
    "overconf_postdebate_minus_round0": {
      "mean": 0.0658,
      "lo": 0.0202,
      "hi": 0.1162
    },
    "conf_postdebate_minus_baseline3": {
      "mean": 0.0357,
      "lo": -0.0152,
      "hi": 0.0859
    },
    "overconf_postdebate_minus_baseline3": {
      "mean": 0.0358,
      "lo": -0.0202,
      "hi": 0.0909
    },
    "conf_postdebate_minus_baselinefull": {
      "mean": 0.0594,
      "lo": 0.0158,
      "hi": 0.1058
    },
    "auc_postdebate_minus_baselinefull": {
      "mean": -0.0889,
      "lo": -0.276,
      "hi": 0.08
    },
    "auc_postdebate_minus_baseline3": {
      "mean": -0.0234,
      "lo": -0.2419,
      "hi": 0.196
    },
    "acc_postdebate_minus_baselinefull": {
      "mean": -0.0,
      "lo": -0.0455,
      "hi": 0.0455
    },
    "acc_postdebate_minus_round0": {
      "mean": -0.0304,
      "lo": -0.0758,
      "hi": 0.0
    }
  },
  "per_source_postdebate_vs_baseline": {
    "fresh": {
      "n": 2,
      "swarm_postdebate": {
        "accuracy": 0.0,
        "overconfidence": 0.6667
      },
      "baseline_full": {
        "accuracy": 0.0,
        "overconfidence": 0.4464
      }
    },
    "gsm8k": {
      "n": 20,
      "swarm_postdebate": {
        "accuracy": 0.9,
        "overconfidence": 0.0667
      },
      "baseline_full": {
        "accuracy": 0.95,
        "overconfidence": -0.0205
      }
    },
    "mc": {
      "n": 1,
      "swarm_postdebate": {
        "accuracy": 1.0,
        "overconfidence": -0.3333
      },
      "baseline_full": {
        "accuracy": 1.0,
        "overconfidence": -0.2857
      }
    },
    "mmlu": {
      "n": 20,
      "swarm_postdebate": {
        "accuracy": 1.0,
        "overconfidence": -0.05
      },
      "baseline_full": {
        "accuracy": 0.95,
        "overconfidence": -0.0738
      }
    },
    "priv": {
      "n": 1,
      "swarm_postdebate": {
        "accuracy": 1.0,
        "overconfidence": -0.3333
      },
      "baseline_full": {
        "accuracy": 1.0,
        "overconfidence": -0.2857
      }
    },
    "reason": {
      "n": 2,
      "swarm_postdebate": {
        "accuracy": 1.0,
        "overconfidence": 0.0
      },
      "baseline_full": {
        "accuracy": 1.0,
        "overconfidence": -0.1964
      }
    },
    "triviaqa": {
      "n": 20,
      "swarm_postdebate": {
        "accuracy": 0.6,
        "overconfidence": 0.3167
      },
      "baseline_full": {
        "accuracy": 0.6,
        "overconfidence": 0.2696
      }
    }
  }
}