Consecutive Zeckendorf–Niven terms in a 2-AP: a two-parameter family of five-term runs, and six-term runs confined to two Fibonacci shapes
Abstract
A positive integer n is Zeckendorf–Niven when the number s_(Z)(n) of summands in its Zeckendorf representation divides n. Lao, Miller, Rosa, Shiliaev, Tresch, Wong and Zhang recently studied how many consecutive terms of an arithmetic progression of common difference two can be Zeckendorf–Niven. They proved that the only runs of eight or more terms are subsequences of 2,4,…,18, and constructed a five-term run starting at 27+F₁₂₀ⱼ₊₁₇ for every j; between five and seven the maximum outside that block is left undetermined. We prove three things. First, their construction is one line of a two-parameter family: for 8 ≤ a, a+2 ≤ b, 11 ≤ b and Fₐ+F_b ≡ 58 (mod 60), the number n=6+Fₐ+F_b starts a run of five, and the run is exactly five on both sides. Second, every n ≥ 21 that starts a run of six satisfies n+3=F_G or n+5=F_G for some G ≥ 21; beyond n ≥ 21 there is no hypothesis at all — none on the size of n, none on the position of its Zeckendorf digits. Third, neither of the two surviving one-parameter families produces a run of six for G ≤ 501, so no n with 21 ≤ n<F₅₀₀ ≈ 1.39 · 10¹⁰⁴ starts a run of six: throughout that range a run beginning outside 2,4,…,18 has at most five terms, and five occurs. What is left of the question is a single Fibonacci-index divisibility problem, stated in Question [q:residue]. All three results, and every lemma they rest on, are machine-checked in Lean 4.
Open review
This founding-collection manuscript received AI review before publication. Independent human review is open. Submitted reviews enter editorial screening; submitting a review does not change this paper’s status. Contribute an assessment of specific claims, a reproduction, or a correction for editorial screening.
Archived files
- Version 1 · current (opens in a new tab)
Source snapshot 2026-08-30 15:34 UTC
File fingerprint
265a7cb57bb62f224dd7dd63e9ef2c9e626990bf41c0736e6a315ed3d2491e08
Claim ledger
Stated results
Z1candidate2026-08-23
Headline: the source's Theorem 4.4 generalized to a two-parameter family, and its exactness
Z2known2026-08-22
The source's Theorem 4.4 itself, for every j, with the Fibonacci period proved rather than assumed
Z3routine2026-08-22
The nine-term block, the two exhaustive sweeps, and the complete list of six-term runs below 10⁷
Z4candidate2026-08-23
Headline computation: the answer to the source's undetermined maximum is 5
This ledger entry is reported in prose and is not bound to a Lean theorem.Z5routine2026-08-22
Negative controls: the Theorem 4.3 misreading, and four load-bearing hypotheses
Z6routine2026-08-22
Infrastructure: sZ from mathlib, and two proved-equal fast evaluators
Z7known2026-08-22
Zeckendorf gap decomposition: s_Z additive across a gap (sZₐddₒfₛep, sZ_gapₐdd)
Z8routine2026-08-22
Maximum-run theorem with the 10⁷ bound removed: no n >= 21 with a gap at 8 <= G <= 20 starts six
Z9routine2026-08-22
Negative controls for the gap decomposition: separation, index-8 sharpness, shift bound, both directions
ZA1routine2026-08-28
The alternation collapse s_Z(F_(G+2) − w) = s_Z(F_G − w) + 1, and the six-shift profile transport
ZA2routine2026-08-28
The alternating tail: at a first Zeckendorf gap, F_G − 21 ≤ lowPart G n < F_G
ZA3routine2026-08-28
The 21 × 2 base-profile check, and the two w it cannot settle
ZA4candidate2026-08-28
Headline: the gap-index bound removed — a six-term run forces n + 3 or n + 5 to be a Fibonacci number
ZA5candidate2026-08-28
The two surviving families swept to G ≤ 501: no 21 ≤ n < F₅₀₀ ≈ 1.4·10¹⁰⁴ starts six
ZA6routine2026-08-28
Negative controls: both collapse hypotheses, the 21 ≤ n bound, and that the two-family carve-out is real
Provenance
- Generated by
- Machina Mathematica
- Released by
- Korea Superintelligence Labs
- Source context
- Source. Kelly Lao, Steven J. Miller, Nicholas Rosa, Mark Shiliaev, Garrett Tresch, Tony W. H. Wong, Han Zhang, *On Zeckendorf-Niven Numbers and Arithmetic Progressions*, arXiv:2606.24006, June 2026 (NSF DMS-2341670, Polymath Jr).
- Snapshot
- 2026-09-07 03:53 UTC
- Ledger commit
801848d7