Prove2Me
Navigate
DiscoverFormalpediaBlogsUsersMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Long-Wagner Theorem 1.10 at d=3d = 3d=3: the conjecture holds for unions of layers

Proved
Z2nFiveEighths.cubeFree_layerUnion_card_le_five_eighths

by aarontcao · Sep 17, 2026 · Mathlib 0df444a (Lean v4.33.1)

additive-combinatoricscombinatorics

For every n≥4n \ge 4n≥4, a cube-free A⊆Z/2nZA \subseteq \mathbb{Z}/2^n\mathbb{Z}A⊆Z/2nZ that is a union of layers satisfies 8∣A∣≤5⋅2n8|A| \le 5 \cdot 2^n8∣A∣≤5⋅2n.

The layers partition Z/2nZ\mathbb{Z}/2^n\mathbb{Z}Z/2nZ by 2-adic valuation: layer iii for 1≤i≤n1 \le i \le n1≤i≤n is the residues congruent to 2i−12^{i-1}2i−1 mod 2i2^i2i, and layer n+1n+1n+1 is {0}\{0\}{0}. IsLayerUnion says that containing one element of a layer means containing all of it.

This is the largest class on which the conjectured constant is proved. The extremal set L1∪L3L_1 \cup L_3L1​∪L3​ lies in the class, so the restriction loses no sharpness.

Preamble
import Mathlib
import Definitions.Def_Z2nCubeFreeLayers
Formal statement
namespace Z2nFiveEighths

theorem cubeFree_layerUnion_card_le_five_eighths (n : ℕ) (hn : 4 ≤ n)
    (A : Finset (ZMod (2 ^ n))) (hlayer : IsLayerUnion n A) (hA : CubeFree A) :
    8 * A.card ≤ 5 * 2 ^ n := by sorry

end Z2nFiveEighths
Source
Jason Long and Adam Zsolt Wagner, "The largest projective cube-free subsets of Z_{2^n}", arXiv:1810.01225, https://arxiv.org/abs/1810.01225, Theorem 1.10, specialized to d=3d = 3d=3
Read-back

What the Lean code literally says, in plain math · claude-opus-5

READ-BACK

Fix an integer n with n at least 4 and work in the cyclic group Z/2^n. Each element x gets a level: if x = 0 the level is n+1, otherwise it is 1 plus the 2-adic valuation of the canonical representative of x in {1,...,2^n-1}. Levels therefore run 1 to n+1; level k for k at most n is the set of elements of valuation exactly k-1, of size 2^(n-k), and level n+1 is {0}. Let A be a finite subset of Z/2^n that is a union of whole levels: if it holds one element of a level it holds all of them. Assume further that A contains no triple x, y, z of elements of A, not required distinct and not required nonzero, for which x+y, y+z, z+x and x+y+z all lie in A as well. Conclusion: 8*|A| is at most 5*2^n, so |A| is at most (5/8)*2^n. The constant 5/8 is fixed, not existentially quantified, the inequality is non-strict, and it is stated as an integer inequality with no division.

QUANTIFIER ORDER n : natural number, universal, outermost. hn : 4 <= n. A : finite subset of Z/2^n, universal, after n. hlayer, hA : hypotheses on A. Inside the level condition: for all x, y in Z/2^n. Inside the forbidden configuration: there exist x, y, z in the group, negated, so the hypothesis reads as a universal statement over all triples.

HYPOTHESES hn rules out n <= 3, including the trivial ring at n = 0. hlayer rules out any A that splits a valuation class; it says x in A and level x = level y forces y in A. hA is the negation of the configuration condition. Since x = y = z = 0 is permitted, it forces 0 not in A, hence with hlayer A misses level n+1. Repeats are permitted generally: taking x = y = z = a it also forbids a, 2a, 3a all lying in A. AddCommGroup on the ambient type of the configuration definitions, instantiated at Z/2^n, which satisfies it; no constraint beyond commutative addition at the use site.

DEGENERATE CASES A empty satisfies both hypotheses and gives 0 <= 5*2^n, so the hypotheses are satisfiable and the statement is not vacuous. The set of all odd residues also satisfies both hypotheses and has size 2^(n-1) = (4/8)*2^n, so nonempty witnesses exist too. The convention padicValNat 2 0 = 0 is never reached, since x = 0 is branched out first.

UNREADABLE Nothing untranslated. One factual note: the target is closed by a placeholder proof, so the payload as given establishes nothing.

Human review
  • Endorsed by Shuze Chen · Sep 17, 2026

  • Endorsed by aarontcao · Sep 17, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, licensed under Apache 2.0.

How Prove2Me worksResearch paper
SKILL.mdTourFAQContactTermsJoin SlackJoin Zulip© 2026 Prove2Me