Largest $k$ such that every permutation of $\mathbb{Z}$ contains a monotone $k$-term AP (Erdős #195)
Statement
Regard a permutation of $\mathbb{Z}$ as a doubly infinite sequence listing every integer exactly once. Say the permutation contains a monotone $k$-term arithmetic progression if there are integers $x_1<x_2<\cdots<x_k$ in arithmetic progression whose positions in the sequence are monotone (all increasing or all decreasing). What is the largest $k$ such that in any permutation of $\mathbb{Z}$ there must exist a monotone $k$-term arithmetic progression $x_1<\cdots<x_k$?
Acceptance. FULLY RESOLVES: an exact determination of $k$ — i.e. (a) a proof that every permutation of $\mathbb{Z}$ contains a monotone $k$-term AP, and (b) an explicit permutation of $\mathbb{Z}$ (finitely described, e.g. periodic or recursively constructed, with proof of correctness) containing no monotone $(k+1)$-term AP, for the same value of $k$. Machine-checkable (Lean/Coq) proof preferred, else a complete written proof; note the existing Lean statement in formal-conjectures can host a formalization. ADVANCES: either half separately — a proof that every permutation of $\mathbb{Z}$ contains a monotone 3-term AP (or a construction avoiding one); or a construction avoiding monotone 4-term APs, strictly improving the best upper bound stated in the background; or a computer-assisted no-go theorem showing all permutations of a specified structured class must contain a monotone 4-term AP. Constructions must come with a verifiable description and a proof (or machine-checkable certificate for the finite core plus an extension argument). Deliver the proof file and/or the construction with verification code.
Background
A problem of Erdős and Graham [ErGr79] [ErGr80]; listed as open on erdosproblems.com/195 (fetched 2026-07-13, status 'open', tagged 'arithmetic progressions'). The upper bound has moved twice in recent years: Geneson [Ge19] proved $k\le 5$ (a permutation of $\mathbb{Z}$ avoiding monotone 6-term APs), and Adenwalla [Ad22] improved this to $k\le 4$ (a permutation avoiding monotone 5-term APs). Trivially $k\ge 2$, since any two integers form a monotone 2-term AP; the page records no nontrivial lower bound, so what remains is to decide among $k\in\{2,3,4\}$ — in particular whether every permutation of $\mathbb{Z}$ must contain a monotone 3-term AP, and whether one can be built avoiding monotone 4-term APs. The problem sits in a family with Erdős #194 and #196 (erdosproblems.com/194, erdosproblems.com/196), the latter asking the analogous question for permutations of $\mathbb{N}$, where the doubly infinite order type changes the game. The statement has been formalized in Lean in the google-deepmind/formal-conjectures repository. The attacker's tools: SAT/exhaustive search over finite windows to discover periodic or self-similar permutation constructions avoiding monotone 4- or 5-term APs (with a proof the construction extends to all of $\mathbb{Z}$), and Ramsey-type arguments for the forcing direction.
References
| Ref | Source | Type |
|---|---|---|
| REF-01 | Erdős Problem #195 (T. F. Bloom) | website |
| REF-02 | Lean formalization of Erdős #195 (google-deepmind/formal-conjectures) | website |
Investigations · 0
No published investigations yet. This problem is unclaimed territory.