{
  "complete": true,
  "completed": 40,
  "expected": 40,
  "source_sha256": "5925c8250ff703004a1e0e915bfdabc28e526849df3f77f0aafaeeb90a91ec33",
  "groups": [
    {
      "task": "double_gate_pick_place",
      "policy": "jev",
      "observation_profile": "full_geometry",
      "completed": 10,
      "successes": 0,
      "wilson_95": [
        2.7755575615628914e-17,
        0.27753279986287815
      ],
      "outcome_categories": {
        "task_failure": 7,
        "response_validation_error": 3
      },
      "evaluable": 10,
      "evaluable_wilson_95": [
        2.7755575615628914e-17,
        0.27753279986287815
      ],
      "excluded_from_evaluable": 0,
      "failures": {
        "obstacle_collision": 7,
        "policy_error": 3
      },
      "mean_mean_request_bytes": 25843.09314714049
    },
    {
      "task": "double_gate_pick_place",
      "policy": "jev",
      "observation_profile": "legacy",
      "completed": 10,
      "successes": 0,
      "wilson_95": [
        2.7755575615628914e-17,
        0.27753279986287815
      ],
      "outcome_categories": {
        "task_failure": 10
      },
      "evaluable": 10,
      "evaluable_wilson_95": [
        2.7755575615628914e-17,
        0.27753279986287815
      ],
      "excluded_from_evaluable": 0,
      "failures": {
        "obstacle_collision": 10
      },
      "mean_mean_request_bytes": 9152.701356981917
    },
    {
      "task": "obstacle_pick_place",
      "policy": "jev",
      "observation_profile": "full_geometry",
      "completed": 10,
      "successes": 2,
      "wilson_95": [
        0.05668215145437705,
        0.5098375284633498
      ],
      "outcome_categories": {
        "task_failure": 4,
        "response_validation_error": 4,
        "success": 2
      },
      "evaluable": 10,
      "evaluable_wilson_95": [
        0.05668215145437705,
        0.5098375284633498
      ],
      "excluded_from_evaluable": 0,
      "failures": {
        "obstacle_collision": 4,
        "policy_error": 4
      },
      "mean_mean_request_bytes": 22545.494357146286
    },
    {
      "task": "obstacle_pick_place",
      "policy": "jev",
      "observation_profile": "legacy",
      "completed": 10,
      "successes": 3,
      "wilson_95": [
        0.10779126740630399,
        0.6032218525388472
      ],
      "outcome_categories": {
        "task_failure": 4,
        "success": 3,
        "response_validation_error": 3
      },
      "evaluable": 10,
      "evaluable_wilson_95": [
        0.10779126740630399,
        0.6032218525388472
      ],
      "excluded_from_evaluable": 0,
      "failures": {
        "max_decisions": 2,
        "obstacle_collision": 2,
        "policy_error": 3
      },
      "mean_mean_request_bytes": 9774.940529150064
    }
  ],
  "pairs": [
    {
      "task": "obstacle_pick_place",
      "seed": 0,
      "comparable": true,
      "legacy": {
        "success": false,
        "end_reason": "max_decisions",
        "decisions": 350,
        "outcome_category": "task_failure"
      },
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 136,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "obstacle_pick_place",
      "seed": 1,
      "comparable": true,
      "full_geometry": {
        "success": false,
        "end_reason": "policy_error",
        "decisions": 190,
        "outcome_category": "response_validation_error"
      },
      "legacy": {
        "success": false,
        "end_reason": "max_decisions",
        "decisions": 350,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "obstacle_pick_place",
      "seed": 2,
      "comparable": true,
      "legacy": {
        "success": true,
        "end_reason": "success",
        "decisions": 124,
        "outcome_category": "success"
      },
      "full_geometry": {
        "success": true,
        "end_reason": "success",
        "decisions": 146,
        "outcome_category": "success"
      }
    },
    {
      "task": "obstacle_pick_place",
      "seed": 3,
      "comparable": true,
      "full_geometry": {
        "success": false,
        "end_reason": "policy_error",
        "decisions": 221,
        "outcome_category": "response_validation_error"
      },
      "legacy": {
        "success": true,
        "end_reason": "success",
        "decisions": 181,
        "outcome_category": "success"
      }
    },
    {
      "task": "obstacle_pick_place",
      "seed": 4,
      "comparable": true,
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 174,
        "outcome_category": "task_failure"
      },
      "full_geometry": {
        "success": false,
        "end_reason": "policy_error",
        "decisions": 142,
        "outcome_category": "response_validation_error"
      }
    },
    {
      "task": "obstacle_pick_place",
      "seed": 5,
      "comparable": true,
      "full_geometry": {
        "success": false,
        "end_reason": "policy_error",
        "decisions": 117,
        "outcome_category": "response_validation_error"
      },
      "legacy": {
        "success": false,
        "end_reason": "policy_error",
        "decisions": 112,
        "outcome_category": "response_validation_error"
      }
    },
    {
      "task": "obstacle_pick_place",
      "seed": 6,
      "comparable": true,
      "legacy": {
        "success": false,
        "end_reason": "policy_error",
        "decisions": 109,
        "outcome_category": "response_validation_error"
      },
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 140,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "obstacle_pick_place",
      "seed": 7,
      "comparable": true,
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 153,
        "outcome_category": "task_failure"
      },
      "legacy": {
        "success": false,
        "end_reason": "policy_error",
        "decisions": 163,
        "outcome_category": "response_validation_error"
      }
    },
    {
      "task": "obstacle_pick_place",
      "seed": 8,
      "comparable": true,
      "legacy": {
        "success": true,
        "end_reason": "success",
        "decisions": 220,
        "outcome_category": "success"
      },
      "full_geometry": {
        "success": true,
        "end_reason": "success",
        "decisions": 181,
        "outcome_category": "success"
      }
    },
    {
      "task": "obstacle_pick_place",
      "seed": 9,
      "comparable": true,
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 161,
        "outcome_category": "task_failure"
      },
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 120,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "double_gate_pick_place",
      "seed": 0,
      "comparable": true,
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 143,
        "outcome_category": "task_failure"
      },
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 145,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "double_gate_pick_place",
      "seed": 1,
      "comparable": true,
      "full_geometry": {
        "success": false,
        "end_reason": "policy_error",
        "decisions": 220,
        "outcome_category": "response_validation_error"
      },
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 58,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "double_gate_pick_place",
      "seed": 2,
      "comparable": true,
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 64,
        "outcome_category": "task_failure"
      },
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 130,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "double_gate_pick_place",
      "seed": 3,
      "comparable": true,
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 223,
        "outcome_category": "task_failure"
      },
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 62,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "double_gate_pick_place",
      "seed": 4,
      "comparable": true,
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 59,
        "outcome_category": "task_failure"
      },
      "full_geometry": {
        "success": false,
        "end_reason": "policy_error",
        "decisions": 294,
        "outcome_category": "response_validation_error"
      }
    },
    {
      "task": "double_gate_pick_place",
      "seed": 5,
      "comparable": true,
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 138,
        "outcome_category": "task_failure"
      },
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 59,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "double_gate_pick_place",
      "seed": 6,
      "comparable": true,
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 143,
        "outcome_category": "task_failure"
      },
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 228,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "double_gate_pick_place",
      "seed": 7,
      "comparable": true,
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 165,
        "outcome_category": "task_failure"
      },
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 145,
        "outcome_category": "task_failure"
      }
    },
    {
      "task": "double_gate_pick_place",
      "seed": 8,
      "comparable": true,
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 144,
        "outcome_category": "task_failure"
      },
      "full_geometry": {
        "success": false,
        "end_reason": "policy_error",
        "decisions": 132,
        "outcome_category": "response_validation_error"
      }
    },
    {
      "task": "double_gate_pick_place",
      "seed": 9,
      "comparable": true,
      "full_geometry": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 141,
        "outcome_category": "task_failure"
      },
      "legacy": {
        "success": false,
        "end_reason": "obstacle_collision",
        "decisions": 62,
        "outcome_category": "task_failure"
      }
    }
  ],
  "paired_comparisons": [
    {
      "task": "obstacle_pick_place",
      "comparable_pairs": 10,
      "expected_pairs": 10,
      "both_success": 2,
      "full_only_success": 0,
      "legacy_only_success": 1,
      "both_failure": 7,
      "exact_mcnemar_two_sided_p": 1.0,
      "success_difference_pp": -10.0
    },
    {
      "task": "double_gate_pick_place",
      "comparable_pairs": 10,
      "expected_pairs": 10,
      "both_success": 0,
      "full_only_success": 0,
      "legacy_only_success": 0,
      "both_failure": 10,
      "exact_mcnemar_two_sided_p": 1.0,
      "success_difference_pp": 0.0
    }
  ],
  "denominators": "10 prespecified seeds per task and input. Transport retries pause physics at the same request. Physical and response-validation failures stay in the denominator. Offline replay adds no independent samples.",
  "interpretation": "Legacy remains the default. Full geometry is a simulator-only comparison. Ten paired seeds per task provide an exploratory estimate; report exact paired tests and do not assume improvement.",
  "campaign": "transport-resilient-observation-20260927",
  "date": "2026-09-27",
  "job_id": "ccc2dca0a4a64e148c01e1e2ec70ea80",
  "transport_protocol": {
    "timeout_s": 60,
    "attempt_limit": 20,
    "backoff_s": "5,10,...,30",
    "restart_completed_or_invalid_answer": false
  },
  "independent_trials": 40,
  "offline_independent_new_trials": 0,
  "offline_verified": 40,
  "offline_expected": 40,
  "offline_network_requests": 0,
  "network_requests": 12631,
  "recovered_transport_errors": 113,
  "recorded_http_errors": 38,
  "unresolved_transport_slots": 0,
  "transport_audit": {
    "http_status_counts": {
      "200": 12480,
      "402": 38
    },
    "transport_error_types": {
      "RemoteProtocolError": 46,
      "ConnectError": 64,
      "ReadTimeout": 2,
      "ConnectTimeout": 1
    },
    "account_block_attempts": 38,
    "transient_http_attempts": 0,
    "network_requests": 12631
  },
  "date_timezone": "UTC; completed on September 28 in Asia/Shanghai",
  "finding": "No success-rate improvement observed: single gate 3/10 legacy vs 2/10 full geometry; double gate 0/10 vs 0/10. Small paired sample; not evidence of equivalence or harm.",
  "historical_campaigns_pooled": false,
  "failure_counts": {
    "obstacle_collision": 23,
    "max_decisions": 2,
    "response_validation_error": 10
  }
}
