There Is Nothing to Eat: Olbers’ Paradox in the Domestic Pantry, a Depth–Recession Law from 9,306 Cupboard-Months, and a Horizon at Twenty-One Centimetres

A full cupboard routinely contains nothing to eat. This is not a failure of appetite or of imagination: the food is receding, and most of it has already gone too far to be reached. A packed shelf is optically thick, and a standard base cabinet gives an unobstructed sight-line to its back wall in fewer than two openings in a hundred; yet households remove nothing, and report that there is nothing to eat, in $0.61$ of openings (95% CI $0.59$–$0.63$). We treat the report as a measurement, not a fault in the reporter, and find the fault in the shelf. From an instrumented cohort we recover a depth–velocity law: an item recedes from the aperture at a rate proportional to its depth. Two consequences follow. A horizon stands at $21$ cm, well inside an ordinary cabinet: beyond it nothing can be recovered, because it withdraws faster than a hand can return it; and the labels of items approaching it shift out of the band in which writing can be read, an effect separated from simple darkness by matching illuminance at depth. Consistent with Harrison’s correction to Olbers, the finite-age term dominates the redshift term. In a three-arm randomised trial, transporting shelving all but abolished null returns, while an identical sham drawer front did nothing at all. Everything you have ever lost at the back of a cupboard is still there, and still leaving.