{
  "title": "Fire department incident reporting revision data for AI",
  "source_page": "https://fiund.com/industries/fire-service-report-revision-data",
  "authorship": "Original fictional Fiund editorial exercise. No actual operating records or third-party source assets are included.",
  "example": {
    "title": "Count current report categories after excluding the training entry",
    "context": "A synthetic reporting register uses neutral event categories and an explicit counting rule. No real dispatch, incident, resident, tactics or medical information appears.",
    "records": [
      [
        "FRP1",
        "Practice event register",
        "E1 category A revision 1, nontraining. E2 category A, training. E3 category A, nontraining."
      ],
      [
        "FRP2",
        "Authorized E1 revision 2",
        "Changes E1 category to B. E1 remains nontraining; other events unchanged."
      ],
      [
        "FRP3",
        "Counting rule and old summary",
        "Exclude training entries and use current revision. Earlier summary counted three A entries before correction and training exclusion."
      ]
    ],
    "question": "What current category counts follow from the supplied rule, and why should the older three-A summary not be reused?",
    "answer": "FRP2 makes E1 category B, while E3 remains A. FRP3 excludes E2 because FRP1 labels it training. The current exercise count is therefore A=1 and B=1, covering two nontraining events. The older three-A summary both uses E1's superseded category and includes training E2. Cite the event revisions and exclusion rule instead of treating the cached total as an independent current observation.",
    "failureModes": [
      "Counting both revisions of E1 as separate events.",
      "Retaining the training entry or reusing a summary produced before the stated exclusions."
    ],
    "fields": [
      [
        "Event identity and current revision",
        "Prevents duplicate-event counts."
      ],
      [
        "Training flag",
        "Defines the population being summarized."
      ],
      [
        "Category and summary provenance",
        "Makes the recomputed totals inspectable."
      ]
    ],
    "holdout": "Correct E3's training flag while leaving its category unchanged. Require population membership to update independently of category revision.",
    "boundary": "Use fabricated labels and totals. Exclude actual dispatch, fireground footage, addresses, tactics, medical records and response decisions."
  }
}