{
  "schema_version": 3,
  "status": "ALL_FOUR_ARMS_TRAINED_AND_SAVED",
  "observed_at_utc": "2026-11-01T18:35:53.729624+00:00",
  "observed_at_beijing": "2026-11-02 02:35:53",
  "round": 2,
  "completed_plan_arms": [
    "WT",
    "DT",
    "WDT"
  ],
  "plans": [
    {
      "arm": "WT",
      "plan_input": "wt_plan",
      "core_questions": 31,
      "core_answers_saved": 248,
      "new_candidates": 14,
      "candidate_questions": 45,
      "seed_before": 121,
      "seed_after": 122,
      "seed_answers": 160,
      "old_before": 43,
      "old_after": 50,
      "old_answers": 88,
      "selected_new_questions": 9,
      "allocated_questions": 40,
      "allocated_answers": 320,
      "unselected_candidates": 5,
      "selected_new_kinds": {
        "wide": 9
      }
    },
    {
      "arm": "DT",
      "plan_input": "dt_plan",
      "core_questions": 31,
      "core_answers_saved": 248,
      "new_candidates": 15,
      "candidate_questions": 46,
      "seed_before": 121,
      "seed_after": 122,
      "seed_answers": 160,
      "old_before": 46,
      "old_after": 43,
      "old_answers": 88,
      "selected_new_questions": 9,
      "allocated_questions": 40,
      "allocated_answers": 320,
      "unselected_candidates": 6,
      "selected_new_kinds": {
        "deep_down": 9
      }
    },
    {
      "arm": "WDT",
      "plan_input": "wdt_plan",
      "core_questions": 31,
      "core_answers_saved": 248,
      "new_candidates": 9,
      "candidate_questions": 40,
      "seed_before": 121,
      "seed_after": 125,
      "seed_answers": 160,
      "old_before": 61,
      "old_after": 57,
      "old_answers": 88,
      "selected_new_questions": 9,
      "allocated_questions": 40,
      "allocated_answers": 320,
      "unselected_candidates": 0,
      "selected_new_kinds": {
        "deep_down": 3,
        "wide": 6
      }
    }
  ],
  "allocation": {
    "status": "ALLOCATED",
    "round": 2,
    "saved_at_utc": "2026-11-01T17:11:57+00:00",
    "actual_questions_per_arm": 40,
    "answers_per_arm": 320,
    "seed_questions_per_teacher": 20,
    "old_questions_per_teacher": 11,
    "selected_new_questions_per_teacher": 9,
    "candidate_question_counts": {
      "WT": 45,
      "DT": 46,
      "WDT": 40
    },
    "qualified_new_candidates": {
      "DT": 15,
      "WDT": 9,
      "WT": 14
    },
    "unselected_candidates": {
      "DT": 6,
      "WDT": 0,
      "WT": 5
    },
    "selected_new_kind_counts": {
      "WT": {
        "wide": 9
      },
      "DT": {
        "deep_down": 9
      },
      "WDT": {
        "deep_down": 3,
        "wide": 6
      }
    },
    "training_arms": [
      "WT",
      "DT",
      "WDT",
      "R"
    ],
    "selection_uses_new_task_outcomes": false,
    "source_evidence": [
      {
        "record": "allocation.json",
        "bytes": 4355,
        "sha256": "22de4e6d4522d45b57975fef4a02b20f860b46edeaf306b99bf40396660fcc2c"
      },
      {
        "record": "allocation_capture.json",
        "bytes": 1083,
        "sha256": "4c8a41c2200d08eb62e95d9f2d9c6d42649e20b31c79dd88450d6486cb33921e"
      }
    ]
  },
  "learning_summary": [
    {
      "arm": "WT",
      "round": 2,
      "questions": 40,
      "answers": 320,
      "effective_questions": 19,
      "effective_fraction": 0.475,
      "seed_correct_answers": 122,
      "seed_answers": 160,
      "model_updated": true,
      "whole_round_updates": 1,
      "model_policy_id": "d85d5d82a18ba9f945ca200ebc63c3681e74866ecc5b00f6b1465bd9aa0d6f10",
      "sampling_memory_retries": 0,
      "training_memory_retries": 0,
      "model_saved": true,
      "model_state_bytes": 390190292,
      "model_state_sha256": "b57a830e51690ec7fa0954a473ef4fcecfdf07947a7f0774e926dead76058631",
      "model_saved_at_utc": "2026-11-01T17:32:57+00:00",
      "collection_saved_at_utc": "2026-11-01T17:24:22+00:00",
      "learning_and_save_elapsed_seconds": 515.0,
      "timing_basis": "collection_saved_to_model_complete_marker",
      "worker_exit_code": 0,
      "source_evidence": [
        {
          "record": "progress_completed.json",
          "bytes": 95995,
          "sha256": "06543f95436e0318e7b568581b738fd97a6cf3f72699f5e33d593891a0d36cd3"
        },
        {
          "record": "CAPTURE.json",
          "bytes": 17610,
          "sha256": "2d0f4509138a0b2d9f2df5586bb7a6e83b400142d971e5ad99baeb22c6916f30"
        },
        {
          "record": "complete.json",
          "bytes": 8711,
          "sha256": "ff5d4dafe38e900ce257770d2eacc7319ed7bf5708a2c9f2f55ef091af826976"
        },
        {
          "record": "COMPLETE.json",
          "bytes": 337,
          "sha256": "2e6a0f488fa6258eb217d555040ef437bd0718783ff47164ed2d49950e934eb8"
        }
      ]
    },
    {
      "arm": "DT",
      "round": 2,
      "questions": 40,
      "answers": 320,
      "effective_questions": 25,
      "effective_fraction": 0.625,
      "seed_correct_answers": 122,
      "seed_answers": 160,
      "model_updated": true,
      "whole_round_updates": 1,
      "model_policy_id": "17f3ed0d0c5ecfaadf78f3992d147b0fc856a8e54c8890ba3d4152d23f1297a8",
      "sampling_memory_retries": 0,
      "training_memory_retries": 0,
      "model_saved": true,
      "model_state_bytes": 390190292,
      "model_state_sha256": "8f74c213832422c95ecc4e27630cce4992c170d883f420a939367b5c13a47f60",
      "model_saved_at_utc": "2026-11-01T17:51:01+00:00",
      "collection_saved_at_utc": "2026-11-01T17:39:47+00:00",
      "learning_and_save_elapsed_seconds": 674.0,
      "timing_basis": "collection_saved_to_model_complete_marker",
      "worker_exit_code": 0,
      "source_evidence": [
        {
          "record": "progress_completed.json",
          "bytes": 95995,
          "sha256": "06543f95436e0318e7b568581b738fd97a6cf3f72699f5e33d593891a0d36cd3"
        },
        {
          "record": "CAPTURE.json",
          "bytes": 17610,
          "sha256": "2d0f4509138a0b2d9f2df5586bb7a6e83b400142d971e5ad99baeb22c6916f30"
        },
        {
          "record": "complete.json",
          "bytes": 8719,
          "sha256": "b50877eee545681900439830621f2fe8e4dcbb15842fd2efd3dec2005e727f25"
        },
        {
          "record": "COMPLETE.json",
          "bytes": 337,
          "sha256": "9b7f1fc19147bcb8b00ca51c1cd2c7810173cc33060353bbb9e38f1ef05bc1a3"
        }
      ]
    },
    {
      "arm": "WDT",
      "round": 2,
      "questions": 40,
      "answers": 320,
      "effective_questions": 19,
      "effective_fraction": 0.475,
      "seed_correct_answers": 125,
      "seed_answers": 160,
      "model_updated": true,
      "whole_round_updates": 1,
      "model_policy_id": "f1dd916cad49985cad9da01681c71709f12167da36a5187447fcf49ae6118559",
      "sampling_memory_retries": 0,
      "training_memory_retries": 0,
      "model_saved": true,
      "model_state_bytes": 390190292,
      "model_state_sha256": "f35041c50768fe11d0874657ac5bc4bf6044c195e8f2fe7b8ddc7e2e19cd15c8",
      "model_saved_at_utc": "2026-11-01T18:03:53+00:00",
      "collection_saved_at_utc": "2026-11-01T17:56:47+00:00",
      "learning_and_save_elapsed_seconds": 426.0,
      "timing_basis": "collection_saved_to_model_complete_marker",
      "worker_exit_code": 0,
      "source_evidence": [
        {
          "record": "progress_completed.json",
          "bytes": 95995,
          "sha256": "06543f95436e0318e7b568581b738fd97a6cf3f72699f5e33d593891a0d36cd3"
        },
        {
          "record": "CAPTURE.json",
          "bytes": 17610,
          "sha256": "2d0f4509138a0b2d9f2df5586bb7a6e83b400142d971e5ad99baeb22c6916f30"
        },
        {
          "record": "complete.json",
          "bytes": 8721,
          "sha256": "f5ff9d3025be4b402576f6c902aa5a9a9893138322461a058a18403a8bf4aec0"
        },
        {
          "record": "COMPLETE.json",
          "bytes": 337,
          "sha256": "1c79490dd3cd2138a776922835bbbab4ea248c4e2309a6471f26ba28d525c6e3"
        }
      ]
    },
    {
      "arm": "R",
      "round": 2,
      "questions": 40,
      "answers": 320,
      "effective_questions": 16,
      "effective_fraction": 0.4,
      "seed_correct_answers": 120,
      "seed_answers": 160,
      "model_updated": true,
      "whole_round_updates": 1,
      "model_policy_id": "743c00359419df1fa25555540b9039ec8a29ce47f9b63e426eab7381a340e90b",
      "sampling_memory_retries": 0,
      "training_memory_retries": 0,
      "model_saved": true,
      "model_state_bytes": 390190292,
      "model_state_sha256": "f8c99c99423df3d0e2badc99dd4b40a47418dcbf34f1ca937bcb7893a140dfd4",
      "model_saved_at_utc": "2026-11-01T18:34:21+00:00",
      "collection_saved_at_utc": "2026-11-01T18:28:28+00:00",
      "learning_and_save_elapsed_seconds": 353.0,
      "timing_basis": "collection_saved_to_model_complete_marker",
      "worker_exit_code": 0,
      "source_evidence": [
        {
          "record": "progress_completed.json",
          "bytes": 95995,
          "sha256": "06543f95436e0318e7b568581b738fd97a6cf3f72699f5e33d593891a0d36cd3"
        },
        {
          "record": "CAPTURE.json",
          "bytes": 17610,
          "sha256": "2d0f4509138a0b2d9f2df5586bb7a6e83b400142d971e5ad99baeb22c6916f30"
        },
        {
          "record": "complete.json",
          "bytes": 8701,
          "sha256": "d0fbc3b55cb72a81243611593f00ff94cef1f9f334b18d601d40d5807fd46405"
        },
        {
          "record": "COMPLETE.json",
          "bytes": 337,
          "sha256": "2493d56495039b4597fa2a7ef69c708d667cf40d0df658666f68b3a23ea6d505"
        }
      ]
    }
  ],
  "completed_training_arms": [
    "WT",
    "DT",
    "WDT",
    "R"
  ],
  "all_teacher_plans_complete": true,
  "common_allocation_complete": true,
  "actual_questions_per_arm": 40,
  "answers_per_arm": 320,
  "second_round_training_started": true,
  "second_round_learning_complete": true,
  "saved_answers_by_arm": {
    "WT": 320,
    "DT": 320,
    "WDT": 320,
    "R": 320
  },
  "saved_second_round_answers": 1280,
  "completed_learning_answer_uses": 2272,
  "completed_rounds_physical_answers": 1792,
  "first_round_shared_answer_reuses": 480,
  "third_round_new_answers": 5,
  "physically_saved_answers_total": 1797,
  "answer_count_upper_bound": 4462,
  "remaining_answer_upper_bound": 2665,
  "candidate_counts_are_not_training_allocations": true,
  "score_comparison_scope": "同一批训练题：首轮学习前，与第二轮出题准备时的重新作答；重做发生在第二轮学习开始前。",
  "method_coverage": {
    "WT": {
      "qualified_wide_questions": 14,
      "selected_wide_questions": 9,
      "description": "WT 第二轮得到 14 道合格拓展题，按共同题量选入 9 道。",
      "source_evidence": [
        {
          "record": "wt_plan.json",
          "bytes": 85727,
          "sha256": "cece145730d515d04ef97a9ea9a80d8ebe9358f3caafe1f7444d207cacdf7409"
        },
        {
          "record": "allocation.json",
          "bytes": 4355,
          "sha256": "22de4e6d4522d45b57975fef4a02b20f860b46edeaf306b99bf40396660fcc2c"
        }
      ]
    },
    "DT": {
      "directions": {
        "harder": 1,
        "easier": 2,
        "retain": 28
      },
      "qualified_harder_questions": 0,
      "qualified_easier_questions": 15,
      "selected_harder_questions": 0,
      "selected_easier_questions": 9,
      "task_517_complete_empty_teacher_replies": 4,
      "description": "DT 的 31 道已有题都完成了判断：1 道应出更难题、2 道应出更简单的题、28 道保留。517 已无合格的升难空间，老师四次完整回复都没有给出新题；15 道合格候选及实际选入的 9 道新题，全部是 477 的较简单练习。",
      "source_evidence": [
        {
          "record": "dt_audit.json",
          "bytes": 67423,
          "sha256": "5e6aa00055468352d5547dada2fa42b1817762240244a2433d557240aa431fad"
        }
      ]
    },
    "WDT": {
      "directions": {
        "harder": 2,
        "easier": 3,
        "retain": 26
      },
      "direction_decisions_complete": true,
      "generation_covers_every_eligible_parent": false,
      "unassigned_parent": "477",
      "unassigned_parent_new_questions": 0,
      "qualified_harder_questions": 0,
      "selected_harder_questions": 0,
      "selected_wide_questions": 6,
      "selected_easier_questions": 3,
      "rejected_out_of_rule_harder_candidates": 2,
      "description": "WDT 第二轮对 31 道已有题都完成了难度判断，实际出题覆盖不完整。477 全错并被判为需要更简单的题，却被老师列为未分组，没有得到新增练习；480、517 都从 6/8 到 8/8，但现有规则已无合格的升难空间。第二轮选入 6 道拓展、3 道更简单练习。",
      "source_evidence": [
        {
          "record": "wdt_audit.json",
          "bytes": 74738,
          "sha256": "30c0d59c002afb7174d42c193856e0cc358c390c1fdb50e28124d18acba9ad97"
        },
        {
          "record": "wdt_report.md",
          "bytes": 2267,
          "sha256": "d97f95d4fd4edc8f256622d2503b8a763b147d520e95cbad93bd13790cdc9eef"
        }
      ]
    },
    "new_question_outcomes": {
      "WT": {
        "all_correct": 1,
        "all_wrong": 5,
        "mixed": 3
      },
      "DT": {
        "all_correct": 4,
        "all_wrong": 0,
        "mixed": 5
      },
      "WDT": {
        "all_correct": 6,
        "all_wrong": 0,
        "mixed": 3
      }
    },
    "wdt_new_question_outcomes_by_kind": {
      "deep_down": {
        "all_correct": 2,
        "all_wrong": 0,
        "mixed": 1
      },
      "wide": {
        "all_correct": 4,
        "all_wrong": 0,
        "mixed": 2
      }
    },
    "new_question_outcome_description": "各组新增的 9 道题中，WT 有 3 道、DT 有 5 道、WDT 有 3 道产生学习信号。WT 另有 5 道全部答错；WDT 有 6 道全部答对，其中 4 道是拓展题、2 道是更简单的题。新题的难度还没有完全匹配学生。",
    "harder_practice_benefit_tested_this_round": false,
    "conclusion": "DT、WDT 第二轮都没有选入更难的练习，因此还无法判断这部分练习是否有帮助。"
  },
  "source_evidence": [
    {
      "record": "progress_completed.json",
      "bytes": 95995,
      "sha256": "06543f95436e0318e7b568581b738fd97a6cf3f72699f5e33d593891a0d36cd3"
    },
    {
      "record": "observer_completed.json",
      "bytes": 2328,
      "sha256": "b575c1031b9979eebf6ccd22f904d17c7c37f3f24912bbedd0276d0015307af6"
    },
    {
      "record": "round2_facts.json",
      "bytes": 1638,
      "sha256": "6c0e9ac554c9142742afb35795d1df83c3b353a366741b703d8074fcf1579a33"
    },
    {
      "record": "wt_plan.json",
      "bytes": 85727,
      "sha256": "cece145730d515d04ef97a9ea9a80d8ebe9358f3caafe1f7444d207cacdf7409"
    },
    {
      "record": "dt_plan.json",
      "bytes": 87823,
      "sha256": "c1310a4bf80f483412eaccfbd68256d43e952942b26aa3554500fb057b2a8787"
    },
    {
      "record": "wdt_plan.json",
      "bytes": 78332,
      "sha256": "f339e70107a40881cb1faf599db86eb71644b9986aece148c286d85173ea6ffa"
    },
    {
      "record": "allocation.json",
      "bytes": 4355,
      "sha256": "22de4e6d4522d45b57975fef4a02b20f860b46edeaf306b99bf40396660fcc2c"
    },
    {
      "record": "allocation_capture.json",
      "bytes": 1083,
      "sha256": "4c8a41c2200d08eb62e95d9f2d9c6d42649e20b31c79dd88450d6486cb33921e"
    },
    {
      "record": "wdt_plan_exit.json",
      "bytes": 100,
      "sha256": "78fdd443b48230084d7bcdd132fd69b6966af79838ee09510cb5ba07a30b4971"
    },
    {
      "record": "dt_audit.json",
      "bytes": 67423,
      "sha256": "5e6aa00055468352d5547dada2fa42b1817762240244a2433d557240aa431fad"
    },
    {
      "record": "wdt_audit.json",
      "bytes": 74738,
      "sha256": "30c0d59c002afb7174d42c193856e0cc358c390c1fdb50e28124d18acba9ad97"
    },
    {
      "record": "CAPTURE.json",
      "bytes": 17610,
      "sha256": "2d0f4509138a0b2d9f2df5586bb7a6e83b400142d971e5ad99baeb22c6916f30"
    }
  ]
}