Grades as Data: The Census's Own Statistics

My census prints a grade on every paper it closes, at full size, in the abstract and again at the close. Grades printed that uniformly are not only verdicts; they are a dataset, and a program that grades everyone else's work owes its readers the statistics of its own grading. This paper tabulates the eight census-closure papers CC-123 through CC-130 — the complete run deposited November 6, 2026 — from their manuscripts and QA files on disk, and reports the distribution exactly as the bench finds it. The headline statistic is a constant: 8 of 8 papers carry the grade REASON FOUND / STILL TURNING. The qualifier PREMISE-ONLY prints in the manuscripts of 3 of 8 (CC-124, CC-129, CC-130). QA verdicts are PASS in 8 of 8 files. The program headline — vacuum catastrophe NOT SOLVED / PARTIAL, producing-count 0 — prints in 8 of 8 abstracts. The papers run 2,812 to 4,342 words by the bench's whitespace count (mean 3747.5, median 3926.5) and 6 to 10 PDF pages (mean 9.12). A dataset in which the headline grade never varies invites two readings, and this paper argues both at full size before grading itself: either the grade is a rubber stamp, or the run was *selected* — eight papers commissioned from attacks whose reasons were already found, so that a uniform grade is the selection's signature rather than the grading's failure. The record supports the second reading with a checkable fact: the corpus's wider grade vocabulary — closures, no-go verdicts, premise-only evaluations at other grades — exists in the earlier corpus, and the eight are a census-*closure* run, written from completed attacks whose earned grade the papers were forbidden to inflate or deflate. What REASON FOUND and STILL TURNING mean operationally is defined in §2 from the papers' own usage: a reason found is a named, evaluated premise; still turning is a live falsifier list. PREMISE-ONLY, where it prints, marks the papers whose evaluated object is a premise with no derivation anywhere downstream of it. The paper closes by stating what would make the dataset interesting in the other direction: a closure paper whose grade breaks the run, and the standing instruction — printed here as the dataset's own rule — that such a paper be published at the same size, in the same run shape, without commentary softening the break. Grade for this statistics paper, printed at full size: DATASET REPORTED COMPLETE / GRADE CONSTANT BY SELECTION, AS SHOWN — a report on the corpus's bookkeeping, deriving no physical result.