{
  "schema_version": "1.0.0",
  "primary_contrast": {
    "left": "python-typed",
    "right": "python",
    "direction": "left_minus_right",
    "blocks": 23,
    "estimates": {
      "hidden_test_pass": {
        "mean_difference": 0.13043478260869565,
        "ci95": [
          -0.13043478260869565,
          0.391304347826087
        ]
      },
      "agent_steps": {
        "mean_difference": 1.0,
        "ci95": [
          0.34782608695652173,
          1.6956521739130435
        ]
      },
      "cost_usd": {
        "mean_difference": 0.0005304304347826087,
        "ci95": [
          -0.00023961342391304216,
          0.0011607112173913038
        ]
      },
      "output_tokens": {
        "mean_difference": 207.7826086956522,
        "ci95": [
          36.475000000000286,
          376.4347826086956
        ]
      }
    }
  },
  "study_id": "v1.0-sized-family",
  "repo_revision": "b27e075abffbfca00322637d60e6d533c5ce7d61",
  "rollouts": 119,
  "excluded_infrastructure_failures": [
    {
      "order_index": 93,
      "language": "typescript",
      "attempt": 7,
      "exception_type": "NonZeroAgentExitCodeError"
    }
  ],
  "total_cost_usd": 0.84914395,
  "by_arm": [
    {
      "language": "go",
      "runs": 24,
      "passed": 15,
      "pass_rate": 0.625,
      "pass_rate_ci95": [
        0.4270996229526447,
        0.7884063254095233
      ],
      "total_cost_usd": 0.20011977,
      "mean_agent_steps": 9.583333333333334,
      "mean_input_tokens": 72502.625,
      "mean_output_tokens": 3961.4166666666665,
      "mean_cost_usd": 0.00833832375,
      "mean_agent_steps_when_passed": 7.533333333333333,
      "mean_files_changed": 4.25,
      "mean_lines_added": 296.6666666666667,
      "mean_lines_deleted": 102.33333333333333,
      "mean_patch_lines": 399.0
    },
    {
      "language": "javascript",
      "runs": 24,
      "passed": 18,
      "pass_rate": 0.75,
      "pass_rate_ci95": [
        0.5510055583116656,
        0.8800063384126703
      ],
      "total_cost_usd": 0.1567484,
      "mean_agent_steps": 7.0,
      "mean_input_tokens": 39661.791666666664,
      "mean_output_tokens": 3196.3333333333335,
      "mean_cost_usd": 0.006531183333333333,
      "mean_agent_steps_when_passed": 7.055555555555555,
      "mean_files_changed": 4.333333333333333,
      "mean_lines_added": 60.541666666666664,
      "mean_lines_deleted": 82.375,
      "mean_patch_lines": 142.91666666666666
    },
    {
      "language": "python",
      "runs": 24,
      "passed": 13,
      "pass_rate": 0.5416666666666666,
      "pass_rate_ci95": [
        0.35074865225073715,
        0.7210866638699854
      ],
      "total_cost_usd": 0.1534837,
      "mean_agent_steps": 6.625,
      "mean_input_tokens": 36683.208333333336,
      "mean_output_tokens": 3071.2916666666665,
      "mean_cost_usd": 0.006395154166666667,
      "mean_agent_steps_when_passed": 6.538461538461538,
      "mean_files_changed": 4.333333333333333,
      "mean_lines_added": 109.16666666666667,
      "mean_lines_deleted": 64.04166666666667,
      "mean_patch_lines": 173.20833333333334
    },
    {
      "language": "python-typed",
      "runs": 24,
      "passed": 16,
      "pass_rate": 0.6666666666666666,
      "pass_rate_ci95": [
        0.46706316684738447,
        0.8202780976355061
      ],
      "total_cost_usd": 0.16767614,
      "mean_agent_steps": 7.625,
      "mean_input_tokens": 47593.083333333336,
      "mean_output_tokens": 3309.75,
      "mean_cost_usd": 0.006986505833333333,
      "mean_agent_steps_when_passed": 7.75,
      "mean_files_changed": 127.375,
      "mean_lines_added": 235.91666666666666,
      "mean_lines_deleted": 80.125,
      "mean_patch_lines": 316.0416666666667
    },
    {
      "language": "typescript",
      "runs": 23,
      "passed": 15,
      "pass_rate": 0.6521739130434783,
      "pass_rate_ci95": [
        0.4489033522338618,
        0.8118872415619854
      ],
      "total_cost_usd": 0.17111594,
      "mean_agent_steps": 8.521739130434783,
      "mean_input_tokens": 60795.47826086957,
      "mean_output_tokens": 3386.086956521739,
      "mean_cost_usd": 0.007439823478260868,
      "mean_agent_steps_when_passed": 7.6,
      "mean_files_changed": 7.608695652173913,
      "mean_lines_added": 240.7391304347826,
      "mean_lines_deleted": 88.69565217391305,
      "mean_patch_lines": 329.4347826086956
    }
  ],
  "by_family_and_arm": [
    {
      "task_family": "circuit-breaker",
      "language": "go",
      "runs": 8,
      "passed": 8,
      "pass_rate": 1.0,
      "pass_rate_ci95": [
        0.6755924316585924,
        1.0
      ],
      "total_cost_usd": 0.05015463,
      "mean_agent_steps": 7.5,
      "mean_input_tokens": 39634.5,
      "mean_output_tokens": 3225.125,
      "mean_cost_usd": 0.006269328750000001,
      "mean_agent_steps_when_passed": 7.5,
      "mean_files_changed": 4.0,
      "mean_lines_added": 222.5,
      "mean_lines_deleted": 79.75,
      "mean_patch_lines": 302.25
    },
    {
      "task_family": "circuit-breaker",
      "language": "javascript",
      "runs": 8,
      "passed": 4,
      "pass_rate": 0.5,
      "pass_rate_ci95": [
        0.21521606069626764,
        0.7847839393037324
      ],
      "total_cost_usd": 0.04262199,
      "mean_agent_steps": 5.875,
      "mean_input_tokens": 28836.375,
      "mean_output_tokens": 2650.5,
      "mean_cost_usd": 0.00532774875,
      "mean_agent_steps_when_passed": 5.5,
      "mean_files_changed": 4.0,
      "mean_lines_added": 67.25,
      "mean_lines_deleted": 75.5,
      "mean_patch_lines": 142.75
    },
    {
      "task_family": "circuit-breaker",
      "language": "python",
      "runs": 8,
      "passed": 2,
      "pass_rate": 0.25,
      "pass_rate_ci95": [
        0.07147921209795555,
        0.5907245720727483
      ],
      "total_cost_usd": 0.04358563,
      "mean_agent_steps": 5.75,
      "mean_input_tokens": 28015.25,
      "mean_output_tokens": 2875.875,
      "mean_cost_usd": 0.00544820375,
      "mean_agent_steps_when_passed": 5.5,
      "mean_files_changed": 4.0,
      "mean_lines_added": 97.125,
      "mean_lines_deleted": 60.5,
      "mean_patch_lines": 157.625
    },
    {
      "task_family": "circuit-breaker",
      "language": "python-typed",
      "runs": 8,
      "passed": 4,
      "pass_rate": 0.5,
      "pass_rate_ci95": [
        0.21521606069626764,
        0.7847839393037324
      ],
      "total_cost_usd": 0.04695592,
      "mean_agent_steps": 6.5,
      "mean_input_tokens": 34323.25,
      "mean_output_tokens": 2960.125,
      "mean_cost_usd": 0.00586949,
      "mean_agent_steps_when_passed": 6.25,
      "mean_files_changed": 129.0,
      "mean_lines_added": 209.75,
      "mean_lines_deleted": 75.125,
      "mean_patch_lines": 284.875
    },
    {
      "task_family": "circuit-breaker",
      "language": "typescript",
      "runs": 8,
      "passed": 5,
      "pass_rate": 0.625,
      "pass_rate_ci95": [
        0.30574239265377096,
        0.8631557152608771
      ],
      "total_cost_usd": 0.04464497,
      "mean_agent_steps": 6.5,
      "mean_input_tokens": 37427.75,
      "mean_output_tokens": 2679.75,
      "mean_cost_usd": 0.005580621250000001,
      "mean_agent_steps_when_passed": 6.6,
      "mean_files_changed": 7.0,
      "mean_lines_added": 219.375,
      "mean_lines_deleted": 98.625,
      "mean_patch_lines": 318.0
    },
    {
      "task_family": "expr-eval",
      "language": "go",
      "runs": 8,
      "passed": 0,
      "pass_rate": 0.0,
      "pass_rate_ci95": [
        0.0,
        0.3244075683414076
      ],
      "total_cost_usd": 0.09327932,
      "mean_agent_steps": 13.25,
      "mean_input_tokens": 130996.0,
      "mean_output_tokens": 4858.125,
      "mean_cost_usd": 0.011659915,
      "mean_agent_steps_when_passed": null,
      "mean_files_changed": 4.0,
      "mean_lines_added": 387.125,
      "mean_lines_deleted": 123.75,
      "mean_patch_lines": 510.875
    },
    {
      "task_family": "expr-eval",
      "language": "javascript",
      "runs": 8,
      "passed": 7,
      "pass_rate": 0.875,
      "pass_rate_ci95": [
        0.5291118149439162,
        0.9775825088000281
      ],
      "total_cost_usd": 0.06603834,
      "mean_agent_steps": 8.25,
      "mean_input_tokens": 55100.125,
      "mean_output_tokens": 3742.5,
      "mean_cost_usd": 0.0082547925,
      "mean_agent_steps_when_passed": 8.285714285714286,
      "mean_files_changed": 4.0,
      "mean_lines_added": 44.75,
      "mean_lines_deleted": 76.125,
      "mean_patch_lines": 120.875
    },
    {
      "task_family": "expr-eval",
      "language": "python",
      "runs": 8,
      "passed": 4,
      "pass_rate": 0.5,
      "pass_rate_ci95": [
        0.21521606069626764,
        0.7847839393037324
      ],
      "total_cost_usd": 0.06442355,
      "mean_agent_steps": 7.25,
      "mean_input_tokens": 49970.625,
      "mean_output_tokens": 3355.5,
      "mean_cost_usd": 0.008052943750000001,
      "mean_agent_steps_when_passed": 6.5,
      "mean_files_changed": 4.0,
      "mean_lines_added": 132.375,
      "mean_lines_deleted": 73.375,
      "mean_patch_lines": 205.75
    },
    {
      "task_family": "expr-eval",
      "language": "python-typed",
      "runs": 8,
      "passed": 4,
      "pass_rate": 0.5,
      "pass_rate_ci95": [
        0.21521606069626764,
        0.7847839393037324
      ],
      "total_cost_usd": 0.06787717,
      "mean_agent_steps": 8.375,
      "mean_input_tokens": 64671.5,
      "mean_output_tokens": 3605.75,
      "mean_cost_usd": 0.00848464625,
      "mean_agent_steps_when_passed": 8.75,
      "mean_files_changed": 113.0,
      "mean_lines_added": 250.875,
      "mean_lines_deleted": 78.75,
      "mean_patch_lines": 329.625
    },
    {
      "task_family": "expr-eval",
      "language": "typescript",
      "runs": 8,
      "passed": 3,
      "pass_rate": 0.375,
      "pass_rate_ci95": [
        0.13684428473912286,
        0.694257607346229
      ],
      "total_cost_usd": 0.07863325,
      "mean_agent_steps": 11.5,
      "mean_input_tokens": 96430.25,
      "mean_output_tokens": 4139.25,
      "mean_cost_usd": 0.00982915625,
      "mean_agent_steps_when_passed": 9.666666666666666,
      "mean_files_changed": 7.0,
      "mean_lines_added": 242.125,
      "mean_lines_deleted": 87.875,
      "mean_patch_lines": 330.0
    },
    {
      "task_family": "money-rollup",
      "language": "go",
      "runs": 8,
      "passed": 7,
      "pass_rate": 0.875,
      "pass_rate_ci95": [
        0.5291118149439162,
        0.9775825088000281
      ],
      "total_cost_usd": 0.05668582,
      "mean_agent_steps": 8.0,
      "mean_input_tokens": 46877.375,
      "mean_output_tokens": 3801.0,
      "mean_cost_usd": 0.0070857275,
      "mean_agent_steps_when_passed": 7.571428571428571,
      "mean_files_changed": 4.75,
      "mean_lines_added": 280.375,
      "mean_lines_deleted": 103.5,
      "mean_patch_lines": 383.875
    },
    {
      "task_family": "money-rollup",
      "language": "javascript",
      "runs": 8,
      "passed": 7,
      "pass_rate": 0.875,
      "pass_rate_ci95": [
        0.5291118149439162,
        0.9775825088000281
      ],
      "total_cost_usd": 0.04808807,
      "mean_agent_steps": 6.875,
      "mean_input_tokens": 35048.875,
      "mean_output_tokens": 3196.0,
      "mean_cost_usd": 0.0060110087500000004,
      "mean_agent_steps_when_passed": 6.714285714285714,
      "mean_files_changed": 5.0,
      "mean_lines_added": 69.625,
      "mean_lines_deleted": 95.5,
      "mean_patch_lines": 165.125
    },
    {
      "task_family": "money-rollup",
      "language": "python",
      "runs": 8,
      "passed": 7,
      "pass_rate": 0.875,
      "pass_rate_ci95": [
        0.5291118149439162,
        0.9775825088000281
      ],
      "total_cost_usd": 0.04547452,
      "mean_agent_steps": 6.875,
      "mean_input_tokens": 32063.75,
      "mean_output_tokens": 2982.5,
      "mean_cost_usd": 0.005684315,
      "mean_agent_steps_when_passed": 6.857142857142857,
      "mean_files_changed": 5.0,
      "mean_lines_added": 98.0,
      "mean_lines_deleted": 58.25,
      "mean_patch_lines": 156.25
    },
    {
      "task_family": "money-rollup",
      "language": "python-typed",
      "runs": 8,
      "passed": 8,
      "pass_rate": 1.0,
      "pass_rate_ci95": [
        0.6755924316585924,
        1.0
      ],
      "total_cost_usd": 0.05284305,
      "mean_agent_steps": 8.0,
      "mean_input_tokens": 43784.5,
      "mean_output_tokens": 3363.375,
      "mean_cost_usd": 0.006605381250000001,
      "mean_agent_steps_when_passed": 8.0,
      "mean_files_changed": 140.125,
      "mean_lines_added": 247.125,
      "mean_lines_deleted": 86.5,
      "mean_patch_lines": 333.625
    },
    {
      "task_family": "money-rollup",
      "language": "typescript",
      "runs": 7,
      "passed": 7,
      "pass_rate": 1.0,
      "pass_rate_ci95": [
        0.6456695613241317,
        1.0
      ],
      "total_cost_usd": 0.04783772,
      "mean_agent_steps": 7.428571428571429,
      "mean_input_tokens": 46776.0,
      "mean_output_tokens": 3332.5714285714284,
      "mean_cost_usd": 0.00683396,
      "mean_agent_steps_when_passed": 7.428571428571429,
      "mean_files_changed": 9.0,
      "mean_lines_added": 263.57142857142856,
      "mean_lines_deleted": 78.28571428571429,
      "mean_patch_lines": 341.85714285714283
    }
  ],
  "family_rankings": [
    {
      "task_family": "circuit-breaker",
      "order": [
        [
          "go",
          8,
          8
        ],
        [
          "typescript",
          5,
          8
        ],
        [
          "javascript",
          4,
          8
        ],
        [
          "python-typed",
          4,
          8
        ],
        [
          "python",
          2,
          8
        ]
      ],
      "spread": 0.75
    },
    {
      "task_family": "expr-eval",
      "order": [
        [
          "javascript",
          7,
          8
        ],
        [
          "python",
          4,
          8
        ],
        [
          "python-typed",
          4,
          8
        ],
        [
          "typescript",
          3,
          8
        ],
        [
          "go",
          0,
          8
        ]
      ],
      "spread": 0.875
    },
    {
      "task_family": "money-rollup",
      "order": [
        [
          "python-typed",
          8,
          8
        ],
        [
          "typescript",
          7,
          7
        ],
        [
          "go",
          7,
          8
        ],
        [
          "javascript",
          7,
          8
        ],
        [
          "python",
          7,
          8
        ]
      ],
      "spread": 0.125
    }
  ],
  "paired_contrasts": [
    {
      "left": "javascript",
      "right": "typescript",
      "direction": "left_minus_right",
      "blocks": 23,
      "primary": false,
      "estimates": {
        "hidden_test_pass": {
          "mean_difference": 0.13043478260869565,
          "ci95": [
            -0.13043478260869565,
            0.391304347826087
          ]
        },
        "agent_steps": {
          "mean_difference": -1.565217391304348,
          "ci95": [
            -2.5652173913043477,
            -0.6086956521739131
          ]
        },
        "cost_usd": {
          "mean_difference": -0.0009385439130434781,
          "ci95": [
            -0.0014683964456521732,
            -0.0003960349347826085
          ]
        },
        "output_tokens": {
          "mean_difference": -205.6086956521739,
          "ci95": [
            -384.1358695652174,
            -11.259782608695746
          ]
        }
      }
    },
    {
      "left": "javascript",
      "right": "python",
      "direction": "left_minus_right",
      "blocks": 23,
      "primary": false,
      "estimates": {
        "hidden_test_pass": {
          "mean_difference": 0.2608695652173913,
          "ci95": [
            0.08695652173913043,
            0.43478260869565216
          ]
        },
        "agent_steps": {
          "mean_difference": 0.30434782608695654,
          "ci95": [
            -0.13043478260869565,
            0.7391304347826086
          ]
        },
        "cost_usd": {
          "mean_difference": 3.7837826086956724e-05,
          "ci95": [
            -0.0005885651413043478,
            0.0005358405326086956
          ]
        },
        "output_tokens": {
          "mean_difference": 93.95652173913044,
          "ci95": [
            -67.26304347826085,
            248.8695652173913
          ]
        }
      }
    },
    {
      "left": "javascript",
      "right": "python-typed",
      "direction": "left_minus_right",
      "blocks": 23,
      "primary": false,
      "estimates": {
        "hidden_test_pass": {
          "mean_difference": 0.13043478260869565,
          "ci95": [
            -0.08695652173913043,
            0.34782608695652173
          ]
        },
        "agent_steps": {
          "mean_difference": -0.6956521739130435,
          "ci95": [
            -1.3478260869565217,
            -0.08695652173913043
          ]
        },
        "cost_usd": {
          "mean_difference": -0.000492592608695652,
          "ci95": [
            -0.0009273994891304342,
            -5.953221739130381e-05
          ]
        },
        "output_tokens": {
          "mean_difference": -113.82608695652173,
          "ci95": [
            -316.2173913043478,
            87.4358695652173
          ]
        }
      }
    },
    {
      "left": "javascript",
      "right": "go",
      "direction": "left_minus_right",
      "blocks": 23,
      "primary": false,
      "estimates": {
        "hidden_test_pass": {
          "mean_difference": 0.17391304347826086,
          "ci95": [
            -0.13043478260869565,
            0.43478260869565216
          ]
        },
        "agent_steps": {
          "mean_difference": -2.739130434782609,
          "ci95": [
            -3.6956521739130435,
            -1.7391304347826086
          ]
        },
        "cost_usd": {
          "mean_difference": -0.0019375899999999999,
          "ci95": [
            -0.0025658314999999995,
            -0.0013195595217391318
          ]
        },
        "output_tokens": {
          "mean_difference": -804.3478260869565,
          "ci95": [
            -1038.1315217391304,
            -570.9510869565221
          ]
        }
      }
    },
    {
      "left": "typescript",
      "right": "python",
      "direction": "left_minus_right",
      "blocks": 23,
      "primary": false,
      "estimates": {
        "hidden_test_pass": {
          "mean_difference": 0.13043478260869565,
          "ci95": [
            -0.13043478260869565,
            0.391304347826087
          ]
        },
        "agent_steps": {
          "mean_difference": 1.8695652173913044,
          "ci95": [
            0.9130434782608695,
            2.9565217391304346
          ]
        },
        "cost_usd": {
          "mean_difference": 0.0009763817391304349,
          "ci95": [
            0.00026442205434782615,
            0.0016915068152173912
          ]
        },
        "output_tokens": {
          "mean_difference": 299.5652173913044,
          "ci95": [
            56.64782608695653,
            542.0445652173913
          ]
        }
      }
    },
    {
      "left": "typescript",
      "right": "python-typed",
      "direction": "left_minus_right",
      "blocks": 23,
      "primary": false,
      "estimates": {
        "hidden_test_pass": {
          "mean_difference": 0.0,
          "ci95": [
            -0.2608695652173913,
            0.2608695652173913
          ]
        },
        "agent_steps": {
          "mean_difference": 0.8695652173913043,
          "ci95": [
            0.0,
            1.8695652173913044
          ]
        },
        "cost_usd": {
          "mean_difference": 0.0004459513043478262,
          "ci95": [
            -2.1922010869564954e-05,
            0.0009261893152173916
          ]
        },
        "output_tokens": {
          "mean_difference": 91.78260869565217,
          "ci95": [
            -105.7836956521739,
            303.0880434782608
          ]
        }
      }
    },
    {
      "left": "typescript",
      "right": "go",
      "direction": "left_minus_right",
      "blocks": 23,
      "primary": false,
      "estimates": {
        "hidden_test_pass": {
          "mean_difference": 0.043478260869565216,
          "ci95": [
            -0.17391304347826086,
            0.2608695652173913
          ]
        },
        "agent_steps": {
          "mean_difference": -1.173913043478261,
          "ci95": [
            -1.9565217391304348,
            -0.391304347826087
          ]
        },
        "cost_usd": {
          "mean_difference": -0.0009990460869565218,
          "ci95": [
            -0.0015485381847826093,
            -0.00045156026086956563
          ]
        },
        "output_tokens": {
          "mean_difference": -598.7391304347826,
          "ci95": [
            -802.4402173913044,
            -386.9554347826088
          ]
        }
      }
    },
    {
      "left": "python",
      "right": "python-typed",
      "direction": "left_minus_right",
      "blocks": 23,
      "primary": true,
      "estimates": {
        "hidden_test_pass": {
          "mean_difference": -0.13043478260869565,
          "ci95": [
            -0.391304347826087,
            0.13043478260869565
          ]
        },
        "agent_steps": {
          "mean_difference": -1.0,
          "ci95": [
            -1.6956521739130435,
            -0.34782608695652173
          ]
        },
        "cost_usd": {
          "mean_difference": -0.0005304304347826087,
          "ci95": [
            -0.0011607112173913038,
            0.00023961342391304216
          ]
        },
        "output_tokens": {
          "mean_difference": -207.7826086956522,
          "ci95": [
            -376.4347826086956,
            -36.475000000000286
          ]
        }
      }
    },
    {
      "left": "python",
      "right": "go",
      "direction": "left_minus_right",
      "blocks": 23,
      "primary": false,
      "estimates": {
        "hidden_test_pass": {
          "mean_difference": -0.08695652173913043,
          "ci95": [
            -0.34782608695652173,
            0.17391304347826086
          ]
        },
        "agent_steps": {
          "mean_difference": -3.0434782608695654,
          "ci95": [
            -4.173913043478261,
            -1.9130434782608696
          ]
        },
        "cost_usd": {
          "mean_difference": -0.001975427826086956,
          "ci95": [
            -0.002726711554347825,
            -0.0012530697608695665
          ]
        },
        "output_tokens": {
          "mean_difference": -898.304347826087,
          "ci95": [
            -1186.5239130434782,
            -611.4315217391306
          ]
        }
      }
    },
    {
      "left": "python-typed",
      "right": "go",
      "direction": "left_minus_right",
      "blocks": 23,
      "primary": false,
      "estimates": {
        "hidden_test_pass": {
          "mean_difference": 0.043478260869565216,
          "ci95": [
            -0.21739130434782608,
            0.30434782608695654
          ]
        },
        "agent_steps": {
          "mean_difference": -2.0434782608695654,
          "ci95": [
            -3.130434782608696,
            -0.9565217391304348
          ]
        },
        "cost_usd": {
          "mean_difference": -0.0014449973913043487,
          "ci95": [
            -0.002194654945652174,
            -0.0007242271195652177
          ]
        },
        "output_tokens": {
          "mean_difference": -690.5217391304348,
          "ci95": [
            -964.6978260869564,
            -413.77717391304395
          ]
        }
      }
    }
  ],
  "failure_profile": [
    {
      "language": "go",
      "failed_cases": {
        "undefined-name": 6,
        "precedence-table": 4,
        "bitwise-full-width": 4,
        "parse-errors": 4,
        "shift-range": 3,
        "bindings-and-shadowing": 3,
        "precedence-basics": 2,
        "shift-semantics": 2,
        "hex-and-signed-literals": 2,
        "no-programs": 2,
        "depth-limit": 2,
        "empty-entries": 1,
        "errors-basics": 1,
        "literal-range": 1,
        "division-truncation": 1,
        "rejects-invalid": 1,
        "bindings-basics": 1
      }
    },
    {
      "language": "javascript",
      "failed_cases": {
        "cooldown-boundary": 4,
        "half-open-closes": 4,
        "half-open-reopens": 4,
        "half-open-limit": 4,
        "target-ordering": 1,
        "rejects-invalid": 1,
        "parse-errors": 1
      }
    },
    {
      "language": "python",
      "failed_cases": {
        "cooldown-boundary": 6,
        "half-open-closes": 6,
        "half-open-reopens": 6,
        "half-open-limit": 6,
        "undefined-name": 3,
        "rejects-invalid": 1,
        "depth-limit": 1
      }
    },
    {
      "language": "python-typed",
      "failed_cases": {
        "cooldown-boundary": 4,
        "half-open-closes": 4,
        "half-open-reopens": 4,
        "half-open-limit": 4,
        "undefined-name": 3,
        "depth-limit": 1
      }
    },
    {
      "language": "typescript",
      "failed_cases": {
        "undefined-name": 4,
        "cooldown-boundary": 3,
        "half-open-closes": 3,
        "half-open-reopens": 3,
        "half-open-limit": 3,
        "wraparound-basics": 2,
        "precedence-table": 2,
        "wraparound": 2,
        "depth-limit": 2,
        "literal-range": 1,
        "bindings-basics": 1,
        "bindings-and-shadowing": 1
      }
    }
  ],
  "expr_eval_width_failures": [
    {
      "language": "javascript",
      "runs": 8,
      "width_case_failures": 0,
      "rate": 0.0
    },
    {
      "language": "typescript",
      "runs": 8,
      "width_case_failures": 2,
      "rate": 0.25
    },
    {
      "language": "python",
      "runs": 8,
      "width_case_failures": 0,
      "rate": 0.0
    },
    {
      "language": "python-typed",
      "runs": 8,
      "width_case_failures": 0,
      "rate": 0.0
    },
    {
      "language": "go",
      "runs": 8,
      "width_case_failures": 5,
      "rate": 0.625
    }
  ],
  "decisive_correctness": [
    "JavaScript passed 0.261 more often than Python (95% CI [0.087, 0.435])"
  ],
  "decisive_steps": [
    "JavaScript needed 1.57 fewer agent steps than TypeScript (95% CI [-2.565, -0.609])",
    "JavaScript needed 0.70 fewer agent steps than Python (typed) (95% CI [-1.348, -0.087])",
    "JavaScript needed 2.74 fewer agent steps than Go (95% CI [-3.696, -1.739])",
    "TypeScript needed 1.87 more agent steps than Python (95% CI [0.913, 2.957])",
    "TypeScript needed 1.17 fewer agent steps than Go (95% CI [-1.957, -0.391])",
    "Python needed 1.00 fewer agent steps than Python (typed) (95% CI [-1.696, -0.348])",
    "Python needed 3.04 fewer agent steps than Go (95% CI [-4.174, -1.913])",
    "Python (typed) needed 2.04 fewer agent steps than Go (95% CI [-3.130, -0.957])"
  ],
  "notes": {
    "primary_estimand": "per-family hidden-verifier pass rate by language, reported for all three families separately and never pooled into one score",
    "pooled_by_arm": "the by_arm table pools families that do not agree; it is reported for completeness and is not the estimand",
    "no_drift_table": "money-rollup and circuit-breaker ran against revised instruction text, so their pass rates are not comparable with v0.8 or v0.9",
    "agent_seconds": "not reported: this cohort ran four rollouts concurrently, so elapsed time includes contention",
    "stopping_rule": "fixed at 120 rollouts, declared before the first paid call; no interim analysis informed continuation"
  },
  "leadership": {
    "minimum_spread_runs": 2,
    "spread_runs": {
      "circuit-breaker": 6,
      "expr-eval": 7,
      "money-rollup": 1
    },
    "discriminating_families": [
      "circuit-breaker",
      "expr-eval"
    ],
    "flat_families": [
      "money-rollup"
    ],
    "best_on_every_discriminating_family": [],
    "worst_on_every_discriminating_family": [],
    "reversals": [
      {
        "language": "go",
        "best_on": "circuit-breaker",
        "worst_on": "expr-eval"
      }
    ]
  }
}
