{
  "schema": "ariasos.public_proof.same_contract_observations.v1",
  "generated_at": "2026-07-14T00:00:00.000Z",
  "measurement_contract": {
    "provider": "xai",
    "model": "grok-4-3",
    "test_item_count": 10,
    "scoring_dimensions": 14,
    "concurrency": 4,
    "comparison_rule": "Only complete runs with the same provider, model, fixed test set, scoring method, item count, and concurrency appear together."
  },
  "observations": {
    "points": [
      {
        "run_id": "2026-05-18T20-53-40-873Z",
        "run_started_at": "2026-05-18T20:53:40.873Z",
        "generated_at": "2026-05-18T20:55:44.898Z",
        "mean_lift": 0.75,
        "baseline_mean": 9.2,
        "aria_mean": 9.95,
        "wins": 6,
        "losses": 4
      },
      {
        "run_id": "2026-05-18T22-36-24-995Z",
        "run_started_at": "2026-05-18T22:36:24.995Z",
        "generated_at": "2026-05-18T22:39:39.217Z",
        "mean_lift": 2.85,
        "baseline_mean": 9.6,
        "aria_mean": 12.45,
        "wins": 10,
        "losses": 0
      }
    ],
    "elapsed_seconds": 6164.122,
    "elapsed_minutes_rounded": 103,
    "best_observed": {
      "mean_lift": 2.85,
      "wins": 10,
      "test_item_count": 10
    }
  },
  "recovery_observation": {
    "excluded_boundary": {
      "run_id": "2026-05-18T22-25-09-821Z",
      "run_started_at": "2026-05-18T22:25:09.821Z",
      "included_in_comparison": false
    },
    "verified_followups": [
      {
        "run_id": "2026-05-18T22-28-55-348Z",
        "run_started_at": "2026-05-18T22:28:55.348Z",
        "mean_lift": 1,
        "elapsed_seconds": 225.527
      },
      {
        "run_id": "2026-05-18T22-30-29-156Z",
        "run_started_at": "2026-05-18T22:30:29.156Z",
        "mean_lift": 1,
        "elapsed_seconds": 319.335
      },
      {
        "run_id": "2026-05-18T22-31-41-158Z",
        "run_started_at": "2026-05-18T22:31:41.158Z",
        "mean_lift": 3,
        "elapsed_seconds": 391.337
      }
    ],
    "seconds_to_first_verified_result": 225.527,
    "seconds_to_strongest_observed_result": 391.337
  },
  "contract_changes": [
    {
      "run_id": "2026-05-22T05-57-57-701Z",
      "test_item_count": 20,
      "concurrency": 2,
      "mean_lift": 1.675,
      "wins": 16,
      "losses": 4,
      "reason_separate": "Expanded 20-item contract; published separately and never connected to the 10-item observations."
    }
  ],
  "claim_boundaries": [
    "The record contains two complete, comparable observations; it is not a smoothed trend, causal claim, or statistical forecast.",
    "A run that did not satisfy the fixed comparison contract is excluded rather than plotted as a quality result.",
    "The later 20-item comparison uses an expanded contract and remains a separate proof artifact.",
    "Live provider output is stochastic; these bounded observations do not guarantee the same score on future prompts."
  ]
}