{
  "bench_version": 1,
  "track": "tabular",
  "model": "mlp_tiny",
  "metric_direction": "lower is better",
  "n_seeds": 8,
  "baseline": {
    "best_cfg": {
      "lr": 0.01,
      "k": 0.0,
      "delay": 0
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "k": 0.0,
          "delay": 0
        },
        "mean": 26.43767786026001
      },
      {
        "cfg": {
          "lr": 0.003,
          "k": 0.0,
          "delay": 0
        },
        "mean": 13.834030151367188
      },
      {
        "cfg": {
          "lr": 0.01,
          "k": 0.0,
          "delay": 0
        },
        "mean": 8.090400218963623
      }
    ],
    "full": {
      "mean": 8.471371114253998,
      "std": 0.789625392148929,
      "per_seed": [
        7.42649507522583,
        8.836640357971191,
        7.776073932647705,
        8.322391510009766,
        7.539560794830322,
        8.821172714233398,
        9.681197166442871,
        9.367437362670898
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 15.51313579082489,
    "std": 0.7923130595085129,
    "per_seed": [
      14.976237297058105,
      14.776838302612305,
      16.32500648498535,
      14.312763214111328,
      15.06303882598877,
      16.302553176879883,
      15.731860160827637,
      16.616788864135742
    ],
    "n": 8
  },
  "comparison": {
    "delta_mean": 7.041764676570892,
    "idea_wins": 0,
    "n_pairs": 8,
    "per_seed_diffs": [
      7.549742221832275,
      5.940197944641113,
      8.548932552337646,
      5.9903717041015625,
      7.523478031158447,
      7.481380462646484,
      6.050662994384766,
      7.249351501464844
    ],
    "p_value": 0.0081,
    "mde": 0.793192058324816,
    "mde_rel_pct": 9.363207533077905,
    "verdict": "idea worse (significant)",
    "system_worked": false
  },
  "mechanism_signature": {
    "track_structure": "optimizer on heterogeneous disjoint-worker tabular training",
    "config": {
      "lr": 0.003,
      "k": 0.02,
      "delay": 0
    },
    "predicted_slowdown": 1.0,
    "observed_collective_step_ratio": 0.3575581591576338,
    "relative_error": 0.6424418408423662,
    "final_disagreement": 9.698813028080622e-05,
    "confirmed": false
  },
  "idea_sweep": [
    {
      "cfg": {
        "lr": 0.003,
        "k": 0.02,
        "delay": 0
      },
      "result": {
        "mean": 15.51313579082489,
        "std": 0.7923130595085129,
        "per_seed": [
          14.976237297058105,
          14.776838302612305,
          16.32500648498535,
          14.312763214111328,
          15.06303882598877,
          16.302553176879883,
          15.731860160827637,
          16.616788864135742
        ],
        "n": 8
      }
    },
    {
      "cfg": {
        "lr": 0.003,
        "k": 0.02,
        "delay": 2
      },
      "result": {
        "mean": 15.614838361740112,
        "std": 0.8421123200541285,
        "per_seed": [
          15.206962585449219,
          14.785978317260742,
          16.650211334228516,
          14.41156005859375,
          14.992582321166992,
          16.48356819152832,
          15.700258255004883,
          16.687585830688477
        ],
        "n": 8
      }
    },
    {
      "cfg": {
        "lr": 0.003,
        "k": 0.02,
        "delay": 5
      },
      "result": {
        "mean": 15.973463654518127,
        "std": 0.9573226631534311,
        "per_seed": [
          16.243694305419922,
          14.805825233459473,
          16.784862518310547,
          14.326708793640137,
          15.480900764465332,
          16.745159149169922,
          16.138805389404297,
          17.26175308227539
        ],
        "n": 8
      }
    }
  ],
  "protocol_notes": {
    "architecture_match": true,
    "baseline": "independent local Adam replicas, ensemble prediction",
    "n_train": 800,
    "epochs": 18,
    "workers": 4,
    "lr_union": [
      0.001,
      0.003,
      0.01
    ]
  }
}