AP packing asymptotics for k near √n

Conjecture 5 · arXiv:2603.02786

arXiv Conjecture low confidence— first stated 2026-03-03

Status open high confidence

Conjecture 5 of arXiv:2603.02786 proposes the asymptotic formula $m_k(n) \sim \left(1 - \frac{k^2}{3n}\right) \frac{k}{\ln k} \cdot n$ for every $k$ satisfying $\Omega(\sqrt{n}) = k < \sqrt{n}$. The paper establishes matching upper and lower bounds with this correction factor in Theorem 4 but stops short of closing the gap to a full asymptotic. No follow-up paper resolving or refuting this conjecture was found in the indexed literature as of May 2026, consistent with the paper being only two months old.

Reviewer notes. No follow-up found. The paper is very recent (March 2026). Conjecture 5 addresses the only remaining open asymptotic regime for m_k(n): the intermediate range Omega(sqrt(n)) = k < sqrt(n). The conjectured formula m_k(n) ~ (1 - k^2/(3n)) * k/ln(k) * n interpolates between the small-k result m_k(n) ~ kn/ln(k) and the large-k regime resolved via Theorem 4 assuming Conjecture 1.

Auto-reviewed 2026-05-14 with claude-sonnet-4-6 (web search enabled).

Conjecture. For every $k$ satisfying $\Omega(\sqrt{n})=k<\sqrt{n}$, we have [asymptotic formula for $m_k(n)$ — mathematical conclusion truncated in source].

Context

If Conjecture 1 holds, it also resolves $m_k(n)$ asymptotically for $k\geqslant\sqrt{n}$ via Theorem 4. The only remaining open case is $k<\sqrt{n}$ with $k\not=o(\sqrt{n})$, which this conjecture addresses.

Notes. The mathematical conclusion after 'we have' is truncated in the extracted source; only the quantifier range is visible.

Source paper

Packing arithmetic progressions
Noga Alon, Michał Dębski, Jarosław Grytczuk, Jakub Przybyło · 2026-03-03
https://arxiv.org/abs/2603.02786