{
  "title": "Fiber optic cable data for AI loss and length context",
  "source_page": "https://fiund.com/industries/fiber-cable-loss-length-context-data",
  "authorship": "Original fictional Fiund editorial exercise. No actual operating records or third-party source assets are included.",
  "example": {
    "title": "Normalize only the portion assigned to cable length",
    "context": "Fictional cable cards use toy loss units with a separate connector contribution. No actual fiber is tested.",
    "records": [
      [
        "1",
        "Sample A",
        "10 length units; total loss 9; connector contribution 4."
      ],
      [
        "2",
        "Sample B",
        "5 length units; total loss 5; connector contribution 1."
      ],
      [
        "3",
        "Rule",
        "For this exercise, subtract recorded connector loss and divide by length."
      ]
    ],
    "question": "Which sample has greater normalized cable-only loss?",
    "answer": "A is (9 − 4) ÷ 10 = 0.5 per length unit. B is (5 − 1) ÷ 5 = 0.8, so B is higher by the explicit rule. Comparing total losses alone would reverse the order and mix connector contribution with cable length.",
    "failureModes": [
      "Dividing total loss by length before separating connectors",
      "Comparing unequal lengths by raw total"
    ],
    "fields": [
      [
        "sample_length",
        "Normalizes cable contribution"
      ],
      [
        "connector_loss",
        "Separates the additive contribution"
      ],
      [
        "cable_only_rate",
        "Records the computed basis"
      ]
    ],
    "holdout": "Omit one connector contribution and require refusing that normalized comparison while retaining the raw total.",
    "boundary": "No real fiber loss, installation advice, test setups or compliance thresholds."
  }
}