{
  "bench_version": 1,
  "track": "poisson_action_value",
  "model": "mlp_tiny",
  "metric_direction": "lower is better",
  "n_seeds": 8,
  "baseline": {
    "best_cfg": {
      "lr": 0.01,
      "n": 8
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "n": 8
        },
        "mean": 0.4739884063601494
      },
      {
        "cfg": {
          "lr": 0.001,
          "n": 16
        },
        "mean": 0.4739884063601494
      },
      {
        "cfg": {
          "lr": 0.001,
          "n": 32
        },
        "mean": 0.4739884063601494
      },
      {
        "cfg": {
          "lr": 0.001,
          "n": 64
        },
        "mean": 0.4739884063601494
      },
      {
        "cfg": {
          "lr": 0.001,
          "n": 128
        },
        "mean": 0.4739884063601494
      },
      {
        "cfg": {
          "lr": 0.003,
          "n": 8
        },
        "mean": 0.2955882400274277
      },
      {
        "cfg": {
          "lr": 0.003,
          "n": 16
        },
        "mean": 0.2955882400274277
      },
      {
        "cfg": {
          "lr": 0.003,
          "n": 32
        },
        "mean": 0.2955882400274277
      },
      {
        "cfg": {
          "lr": 0.003,
          "n": 64
        },
        "mean": 0.2955882400274277
      },
      {
        "cfg": {
          "lr": 0.003,
          "n": 128
        },
        "mean": 0.2955882400274277
      },
      {
        "cfg": {
          "lr": 0.01,
          "n": 8
        },
        "mean": 0.038012176752090454
      },
      {
        "cfg": {
          "lr": 0.01,
          "n": 16
        },
        "mean": 0.038012176752090454
      },
      {
        "cfg": {
          "lr": 0.01,
          "n": 32
        },
        "mean": 0.038012176752090454
      },
      {
        "cfg": {
          "lr": 0.01,
          "n": 64
        },
        "mean": 0.038012176752090454
      },
      {
        "cfg": {
          "lr": 0.01,
          "n": 128
        },
        "mean": 0.038012176752090454
      }
    ],
    "full": {
      "mean": 0.035623283591121435,
      "std": 0.008640293958528833,
      "per_seed": [
        0.03652806580066681,
        0.03965979069471359,
        0.04228668659925461,
        0.03357416391372681,
        0.022456083446741104,
        0.05217588320374489,
        0.02869047038257122,
        0.029615124687552452
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 0.035623283591121435,
    "std": 0.008640293958528833,
    "per_seed": [
      0.03652806580066681,
      0.03965979069471359,
      0.04228668659925461,
      0.03357416391372681,
      0.022456083446741104,
      0.05217588320374489,
      0.02869047038257122,
      0.029615124687552452
    ],
    "n": 8
  },
  "comparison": {
    "delta_mean": 0.0,
    "idea_wins": 0,
    "n_pairs": 8,
    "per_seed_diffs": [
      0.0,
      0.0,
      0.0,
      0.0,
      0.0,
      0.0,
      0.0,
      0.0
    ],
    "p_value": 1.0,
    "mde": 0.0,
    "mde_rel_pct": 0.0,
    "verdict": "no measurable effect",
    "system_worked": false
  },
  "mechanism_signature": {
    "prediction": {
      "quantity": "trained-model candidate action regret",
      "kappa": 2.0,
      "d": 2.0,
      "slope": -1.0
    },
    "observed_slopes_per_seed": [
      -0.7058185791140722,
      -0.5139916507443746,
      -0.48684283734935213,
      -0.5550921083271281,
      -0.3907645786398352,
      -0.27029026113176285,
      -0.4199711072919162,
      -0.6079324563089579
    ],
    "observed_slope_mean": -0.4938379473634249,
    "confirmed": false,
    "note": "Signature uses trained critics; regret is not the primary benchmark metric.",
    "scheduler_pool_sizes": [
      64,
      64,
      64,
      64,
      32,
      64,
      64,
      64
    ],
    "scheduler_pool_mean": 60.0,
    "selected_idea_cfg": {
      "lr": 0.01,
      "epsilon": 0.03
    },
    "custom_track": {
      "name": "poisson_action_value",
      "file": "poisson_action_track.py",
      "domain": "dynamics/control"
    },
    "idea_sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "epsilon": 0.03
        },
        "mean": 0.45789623260498047
      },
      {
        "cfg": {
          "lr": 0.001,
          "epsilon": 0.06
        },
        "mean": 0.45789623260498047
      },
      {
        "cfg": {
          "lr": 0.001,
          "epsilon": 0.12
        },
        "mean": 0.45789623260498047
      },
      {
        "cfg": {
          "lr": 0.003,
          "epsilon": 0.03
        },
        "mean": 0.25725509226322174
      },
      {
        "cfg": {
          "lr": 0.003,
          "epsilon": 0.06
        },
        "mean": 0.25725509226322174
      },
      {
        "cfg": {
          "lr": 0.003,
          "epsilon": 0.12
        },
        "mean": 0.25725509226322174
      },
      {
        "cfg": {
          "lr": 0.01,
          "epsilon": 0.03
        },
        "mean": 0.035623283591121435
      },
      {
        "cfg": {
          "lr": 0.01,
          "epsilon": 0.06
        },
        "mean": 0.035623283591121435
      },
      {
        "cfg": {
          "lr": 0.01,
          "epsilon": 0.12
        },
        "mean": 0.035623283591121435
      }
    ],
    "structural_match": "continuous-action control with a learned Q critic and isolated local action optimum"
  }
}