data/headline.json

{
  "schema_version": 2,
  "generated_at": "2026-05-21T21:28:33+00:00",
  "source": "Built by scripts/build_headline_json.py from data/taxonomy/auto_labels_*.json + docs/STATE.md + docs/figures/phase4_ablation.json.",
  "cells": [
    {
      "cell_id": "y-5cm",
      "axis": "spatial",
      "magnitude": -0.05,
      "mean_tsr_custom": 0.127,
      "std_tsr_custom": 0.009,
      "per_seed_tsr_custom": null,
      "mean_tsr": null,
      "median_tts": null,
      "failure_counts": {
        "success": 19,
        "grasp_failure": 0,
        "approach_failure": 7,
        "recovery_failure": 121,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 3
      },
      "n_rollouts": 150,
      "run_id": "18xb5ob0"
    },
    {
      "cell_id": "y-3cm",
      "axis": "spatial",
      "magnitude": -0.03,
      "mean_tsr_custom": 0.553,
      "std_tsr_custom": 0.025,
      "per_seed_tsr_custom": null,
      "mean_tsr": null,
      "median_tts": null,
      "failure_counts": {
        "success": 83,
        "grasp_failure": 0,
        "approach_failure": 0,
        "recovery_failure": 63,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 4
      },
      "n_rollouts": 150,
      "run_id": "11ugk2a3"
    },
    {
      "cell_id": "y-1cm",
      "axis": "spatial",
      "magnitude": -0.01,
      "mean_tsr_custom": 0.827,
      "std_tsr_custom": 0.034,
      "per_seed_tsr_custom": null,
      "mean_tsr": null,
      "median_tts": null,
      "failure_counts": {
        "success": 124,
        "grasp_failure": 0,
        "approach_failure": 0,
        "recovery_failure": 26,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 0
      },
      "n_rollouts": 150,
      "run_id": "1usoqyez"
    },
    {
      "cell_id": "y+1cm",
      "axis": "spatial",
      "magnitude": 0.01,
      "mean_tsr_custom": 0.72,
      "std_tsr_custom": 0.102,
      "per_seed_tsr_custom": null,
      "mean_tsr": null,
      "median_tts": null,
      "failure_counts": {
        "success": 108,
        "grasp_failure": 0,
        "approach_failure": 0,
        "recovery_failure": 37,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 5
      },
      "n_rollouts": 150,
      "run_id": "p2pltgd8"
    },
    {
      "cell_id": "y+3cm",
      "axis": "spatial",
      "magnitude": 0.03,
      "mean_tsr_custom": 0.553,
      "std_tsr_custom": 0.041,
      "per_seed_tsr_custom": null,
      "mean_tsr": null,
      "median_tts": null,
      "failure_counts": {
        "success": 83,
        "grasp_failure": 0,
        "approach_failure": 1,
        "recovery_failure": 56,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 10
      },
      "n_rollouts": 150,
      "run_id": "miuy4kux"
    },
    {
      "cell_id": "y+5cm",
      "axis": "spatial",
      "magnitude": 0.05,
      "mean_tsr_custom": 0.307,
      "std_tsr_custom": 0.019,
      "per_seed_tsr_custom": null,
      "mean_tsr": null,
      "median_tts": null,
      "failure_counts": {
        "success": 46,
        "grasp_failure": 1,
        "approach_failure": 1,
        "recovery_failure": 89,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 13
      },
      "n_rollouts": 150,
      "run_id": "w6k2wole"
    },
    {
      "cell_id": "nominal",
      "axis": "nominal",
      "magnitude": 0.0,
      "mean_tsr_custom": 0.8,
      "std_tsr_custom": 0.057,
      "per_seed_tsr_custom": null,
      "mean_tsr": null,
      "median_tts": null,
      "failure_counts": {
        "success": 120,
        "grasp_failure": 0,
        "approach_failure": 0,
        "recovery_failure": 0,
        "action_oscillation": 0,
        "timeout": 28,
        "visual_confusion": 0,
        "needs_review": 2
      },
      "n_rollouts": 150,
      "run_id": "cm6uf89g"
    },
    {
      "cell_id": "delay-1step",
      "axis": "temporal",
      "magnitude": 1.0,
      "mean_tsr_custom": 0.753,
      "std_tsr_custom": 0.05,
      "per_seed_tsr_custom": null,
      "mean_tsr": null,
      "median_tts": null,
      "failure_counts": {
        "success": 113,
        "grasp_failure": 0,
        "approach_failure": 0,
        "recovery_failure": 33,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 4
      },
      "n_rollouts": 150,
      "run_id": "ejyfcv2k"
    },
    {
      "cell_id": "delay-3step",
      "axis": "temporal",
      "magnitude": 3.0,
      "mean_tsr_custom": 0.767,
      "std_tsr_custom": 0.068,
      "per_seed_tsr_custom": null,
      "mean_tsr": null,
      "median_tts": null,
      "failure_counts": {
        "success": 115,
        "grasp_failure": 0,
        "approach_failure": 0,
        "recovery_failure": 32,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 3
      },
      "n_rollouts": 150,
      "run_id": "ad3z7eg0"
    },
    {
      "cell_id": "delay-5step",
      "axis": "temporal",
      "magnitude": 5.0,
      "mean_tsr_custom": 0.687,
      "std_tsr_custom": 0.066,
      "per_seed_tsr_custom": null,
      "mean_tsr": null,
      "median_tts": null,
      "failure_counts": {
        "success": 103,
        "grasp_failure": 0,
        "approach_failure": 0,
        "recovery_failure": 45,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 2
      },
      "n_rollouts": 150,
      "run_id": "aycdd8hi"
    }
  ],
  "ablation": [
    {
      "condition_id": "A",
      "label": "Frozen base only",
      "mean_tsr_custom": 0.32,
      "std_tsr_custom": 0.058878405775518984,
      "per_seed_means": [
        0.26,
        0.4,
        0.3
      ],
      "bootstrap_ci": [
        0.26,
        0.4000000000000001
      ],
      "failure_counts": {
        "success": 46,
        "grasp_failure": 1,
        "approach_failure": 1,
        "recovery_failure": 89,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 13
      },
      "run_id": "w6k2wole"
    },
    {
      "condition_id": "B",
      "label": "Residual RL, sparse reward",
      "mean_tsr_custom": 0.18666666666666668,
      "std_tsr_custom": 0.024944382578492942,
      "per_seed_means": [
        0.22,
        0.18,
        0.16
      ],
      "bootstrap_ci": [
        0.16,
        0.22
      ],
      "failure_counts": {
        "success": 28,
        "grasp_failure": 1,
        "approach_failure": 8,
        "recovery_failure": 106,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 7
      },
      "run_id": "o6ukyo53"
    },
    {
      "condition_id": "C",
      "label": "Residual RL, shaped reward",
      "mean_tsr_custom": 0.21333333333333335,
      "std_tsr_custom": 0.0498887651569859,
      "per_seed_means": [
        0.28,
        0.2,
        0.16
      ],
      "bootstrap_ci": [
        0.16,
        0.28
      ],
      "failure_counts": {
        "success": 31,
        "grasp_failure": 1,
        "approach_failure": 3,
        "recovery_failure": 110,
        "action_oscillation": 0,
        "timeout": 0,
        "visual_confusion": 0,
        "needs_review": 5
      },
      "run_id": "43czuigy"
    }
  ],
  "welch_tests": [
    {
      "arm_id": "B",
      "t_statistic": -2.948839123097942,
      "df": 2.695541401273885,
      "p_one_sided": 0.9657149283212578
    },
    {
      "arm_id": "C",
      "t_statistic": -1.9547111097008831,
      "df": 3.8950108459869845,
      "p_one_sided": 0.9378981410804068
    }
  ]
}