The Failure Mode

A flood recedes. A single-story house in a mapped floodplain has water two feet above the floor, the drywall is gone, and the mechanical equipment is scrap. Whether that house is repaired where it sits, at the floor elevation it has now, or lifted onto a new foundation above the base flood level, turns on one number: 50 percent. The number does not come from the water depth or from an engineer. It comes from a definition. The definition lives in the eCFR text of 44 CFR 59.1, Definitions , and that one section carries the 50 percent twice. Once for damage, once for improvement. Each time it is 50 percent of the market value of the structure, and each time the section names a different moment at which that market value is fixed. For a house that has just been flooded, the two moments sit on opposite sides of the flood. That is the whole of it. One estimate of $80,000 to put the house back is 40 percent of a $200,000 pre-damage value and 62 percent of a $130,000 damaged-condition valu...

A critical access hospital is a small rural facility that Medicare certifies under its own set of conditions, separate from the ones a general hospital meets. Two of those conditions are numbers, and in the statute they sit inside one clause: 25 beds and 96 hours. The bed figure is a count of objects and behaves like a count of objects. The hour figure does not behave like a clock. Ask what the 96 hours limits and three answers come back, each correct inside its own document. To the condition of participation it is an average across a year of stays. To the contractor that pays the claims it is a quantity derived from a cost report. To the physician writing the admission order it is a forecast recorded before the stay happens. These are not three phrasings of one rule but three tests that share a number. The reading that travels furthest is the one none of the documents supports: that a patient admitted to a critical access hospital has to be out within 96 hours. Where the three rea...

A Medicare beneficiary arrives at an emergency department at 6 p.m. on a Monday and is discharged at 2 p.m. on Wednesday. The wall clock records 44 hours. Four separate federal texts read that same interval and return four different numbers: two midnights, one midnight, one day, and 12 hours. Every one of the four is correctly computed. The divergence is not rounding and it is not a data quality problem. Three of the four rules count calendar boundaries rather than duration, and they do not start counting at the same instant. The fourth counts hours, but only the hours spent in one billing status. The stay is one physical fact; the record of it is four, and they do not reconcile. What follows traces the 44 hours through each text in the order a hospital applies them. Inpatient Status Begins With a Signature The arithmetic depends first on the fact that inpatient status is not a property of a bed or of how sick someone is. It is created by a document. eCFR, 42 CFR 412.3 Admissions...

One Integer, Two Standards Seven is a small number doing a great deal of work in American earthquake response. It is where the U.S. Geological Survey stops claiming that people can tell one level of shaking from the next, and where the same agency's facility-screening software declares damage likely and sends inspectors. Both statements are published, both are correct inside their own documents, and both describe one quantity: Modified Mercalli Intensity, on the same map of the same earthquake. The first appears where the ShakeMap manual explains how its intensity colors were chosen. Discussing a study that compared telephone surveys with field-assigned intensities after the 1994 Northridge earthquake, the U.S. Geological Survey, ShakeMap Manual, Representing Macroseismic Intensity on Maps concludes: "But we know of no solid evidence that one could discriminate intensities higher than VII on the basis of different individuals' descriptions of perceived shaking alone....