slug: debski-barrycade-omitted-element-refutation bibkey: debski2026barrycades doi: 10.48550/arXiv.2609.18476 url: https://arxiv.org/abs/2609.18476v1 triage: theorem motivation_gids:
- D5/S3/ArithSums/DebskiBarrycadeOmittedElementRefutation.result
Refutation of the printed barrycade omitted-element relation
Problem
Michał Dębski, Jarosław Grytczuk, Paweł Naroski, Bartłomiej Pawlik, Jakub Przybyło, and Małgorzata Śleszyńska-Nowak, Finite and infinite barrycades, arXiv:2609.18476v1, Section 3.1, Conjecture 2 (2), print:
A2(i) = A1(i) + 1 for every i >= 3.
The paper defines A1(i) as the least positive integer omitted from the
quasi-permutation rho_i and A2(i) as its first entry rho_i(1).
Motivation
Issue #8675 records the source quotation, its full universal scope, the
literature-status search, and the proposed i = 3 counterexample. The result
tests the printed lower endpoint rather than a corrected statement beginning
at i = 4.
Gap
The printed row rho_3 = (4, 3, 1, 5, 6, ...) omits 2 and begins with 4.
Consequently A1(3) = 2 and A2(3) = 4, whereas the printed relation requires
A2(3) = 3. The bounded source and literature searches found no earlier
proof or refutation in the searched scope. This dossier makes no claim of
exhaustive literature coverage or publication priority.
Route
Formalize the paper’s partial-sum sets and greedy row recurrence with
zero-based indices, so Lean’s row 2 is the paper’s rho_3. Strong induction
gives the first two rows in closed form. Least-element arguments for the third
row give row 2 0 = 4, row 2 1 = 3, and row 2 2 = 1, while the tail formula
shows that every positive integer other than 2 occurs in that row. Thus 2 is
its least omitted positive integer, but its first entry is 4. Instantiating the
printed universal statement at i = 3 yields the contradiction 4 = 3.
Falsifier
A faithful proof that rho_3 has a different first entry, contains 2, omits a
positive integer smaller than 2, or does not follow the formalized greedy
recurrence would invalidate this counterexample.
Evidence
- Frozen result:
D5/S3/ArithSums/DebskiBarrycadeOmittedElementRefutation.result. - The result has type
Not claimand is kernel-checked in the pinned Lean toolchain. - The proof derives
row 2 0 = 4,row 2 1 = 3,row 2 2 = 1, and the third-row tail formula on the live path to the contradiction.
Triage
theorem; resolution refuted for the literal lower endpoint in Conjecture 2
(2). The proof shape is content; the four form-(1) escape witnesses are
row_zero, row_one, row_two_tail, and row_two_base. Admission is by the
preregistered external-open-problem resolution in issue #8675.
ASSUMED-UNVERIFIED
The bounded literature search is not exhaustive and does not establish worldwide novelty or priority. Unpublished resolutions remain unverified.