4/3 constant in arithmetic progression packing

Conjecture 1 · arXiv:2603.02786

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

Status open high confidence

The source paper (Alon, Dębski, Grytczuk, Przybyło, 2026) proves $m(n)=\Theta(n^{3/2}/\ln n)$ with matching constants 4/3 (lower bound) and 5/3 (upper bound) in Theorem 1, but the exact leading constant remains open. The conjecture that $m(n)\sim\frac{4}{3}\cdot\frac{n^{3/2}}{\ln n}$ is the authors' own expectation stated in the paper. No follow-up work resolving the gap was found in the literature as of May 2026, consistent with the paper being only ~10 weeks old.

Reviewer notes. No follow-up paper found. The conjecture pinning down the constant to 4/3 is very recent (~10 weeks old as of review date). The paper 2605.13628 appeared in search results but is unrelated (it concerns cap sets in finite fields via slice rank). The open gap is between constants 4/3 and 5/3 in the leading coefficient of n^{3/2}/ln n.

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

Conjecture. $m(n)\sim\frac{4}{3}\cdot\frac{n^{3/2}}{\ln n}$.

Context

The paper proves $m(n)=\Theta(n^{3/2}/\ln n)$ with matching constants $4/3$ (lower) and $5/3$ (upper) in Theorem 1. The authors state that closing this gap and pinning down the exact constant requires additional ideas, and suspect $4/3$ is correct.

Also stated in

Source paper

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