The Factor Is a Draw: σ Computed, Never Fitted

On 27 October 2026 I said the residue factor is background noise. A draw, not an exact constant. That sentence was carried for five days with no test under it. The computation it owed was simple to state and nobody ran it: take the residues, compute their σ from the records, and price that σ against the error budget. Never fit it. This paper runs it, on the deposited record of the Closure Identity test (Aldred 2026d). The residues are the per-galaxy distances from the frozen line of that test, for each of the seven distinct candidate forms, on the 116 galaxies the two SPARC reductions share. The slope of the line is the integer 4. Its intercept is frozen by the other table's scale. Nothing is fitted to the residues, so their scatter is a computation, not a fit. The computed σ is 0.227 to 0.237 dex across the seven forms. The propagated measurement budget on the same residues is 0.226 dex. The scatter is the budget. No form's centering demands a constant at the registered 3σ level: the largest displacement is the broadest form's, −0.070 dex, a factor of 0.85, at 2.6 times its own standard error. The sentence stands, and now it has a price: the factor is a draw with σ ≈ 0.23 dex, and the draw is no wider than the errors already on the table.