Prove2Me
Navigate
DiscoverCollectionsFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Structure of a six-colour Schur colouring of [1,1801][1, 1801][1,1801] under R4(3)≤61R_4(3) \le 61R4​(3)≤61: balance, saturation and reflection

Proved
ClassicalSchur.schur_six_frontier_structure

by mysticflounder · Oct 4, 2026 · Mathlib 0df444a (Lean v4.33.1)

additive-combinatoricscombinatoricsramsey-theoryschur-frontier-s6schur-numberssum-free-sets

This is the goal of the mission: the structure that a Schur colouring of [1,1801][1, 1801][1,1801] with six colours must have if R4(3)≤61R_4(3) \le 61R4​(3)≤61.

Suppose that TR(4,61)\mathrm{TR}(4, 61)TR(4,61) holds (Lean TriangleRamsey 4 61): every colouring with at most four colours of the pairs of a finite set of at least 616161 naturals has a monochromatic triangle, that is, R4(3)≤61R_4(3) \le 61R4​(3)≤61. Let c:N→Fin 6c : \mathbb{N} \to \mathrm{Fin}\,6c:N→Fin6 be a Schur colouring of [1,1801][1, 1801][1,1801] with six colours: there are no x,y≥1x, y \ge 1x,y≥1 with x+y≤1801x + y \le 1801x+y≤1801 and c(x)=c(y)=c(x+y)c(x) = c(y) = c(x + y)c(x)=c(y)=c(x+y). Write q=c(901)q = c(901)q=c(901), let

V={ x∈[0,1801]:x≠900, c(∣900−x∣)=q }V = \{\, x \in [0, 1801] : x \ne 900,\ c(|900 - x|) = q \,\}V={x∈[0,1801]:x=900, c(∣900−x∣)=q}

be the central neighbourhood, for v∈Nv \in \mathbb{N}v∈N and a colour iii let Γi(V,v)={ w∈V:w≠v, c(∣v−w∣)=i }\Gamma_i(V, v) = \{\, w \in V : w \ne v,\ c(|v - w|) = i \,\}Γi​(V,v)={w∈V:w=v, c(∣v−w∣)=i}, and let Pi=Γi(V,1801)P_i = \Gamma_i(V, 1801)Pi​=Γi​(V,1801) be the endpoint neighbourhood of colour iii. Then:

  1. every colour jjj occurs 150150150 times in [1,900][1, 900][1,900]: ∣{ d∈[1,900]:c(d)=j }∣=150\bigl|\{\, d \in [1, 900] : c(d) = j \,\}\bigr| = 150​{d∈[1,900]:c(d)=j}​=150;
  2. ∣V∣=301|V| = 301∣V∣=301;
  3. for every v∈Vv \in Vv∈V and every colour i≠qi \ne qi=q: ∣Γi(V,v)∣=60|\Gamma_i(V, v)| = 60∣Γi​(V,v)∣=60;
  4. for every ddd with 1≤d≤9001 \le d \le 9001≤d≤900 and c(d)=qc(d) = qc(d)=q: c(901−d)=c(901+d)c(901 - d) = c(901 + d)c(901−d)=c(901+d);
  5. for every colour i≠qi \ne qi=q: ∣Pi∣=60|P_i| = 60∣Pi​∣=60; for every x∈Pix \in P_ix∈Pi​, 1800−x∈Pi1800 - x \in P_i1800−x∈Pi​ and 1800−x≠x1800 - x \ne x1800−x=x; and for all x,y∈Pix, y \in P_ix,y∈Pi​ with x≠yx \ne yx=y, c(∣x−y∣)≠ic(|x - y|) \ne ic(∣x−y∣)=i and c(∣x−y∣)≠qc(|x - y|) \ne qc(∣x−y∣)=q.

The central part of the statement is

∣V∣=301,∣Γi(V,v)∣=60  (v∈V, i≠q),c(901−d)=c(901+d)  (d∈[1,900], c(d)=q).|V| = 301, \qquad |\Gamma_i(V, v)| = 60 \ \ (v \in V,\ i \ne q), \qquad c(901 - d) = c(901 + d) \ \ (d \in [1, 900],\ c(d) = q).∣V∣=301,∣Γi​(V,v)∣=60  (v∈V, i=q),c(901−d)=c(901+d)  (d∈[1,900], c(d)=q).

Under R4(3)≤61R_4(3) \le 61R4​(3)≤61 the pigeonhole step gives R5(3)≤302R_5(3) \le 302R5​(3)≤302 and the centred bound then gives S(6)≤1801S(6) \le 1801S(6)≤1801, so [1,1801][1, 1801][1,1801] is the frontier interval for six colours. The theorem shows that a Schur colouring of [1,1801][1, 1801][1,1801] with six colours yields five sets PiP_iPi​ of 606060 points, each the union of 303030 pairs {900−d,900+d}\{900 - d, 900 + d\}{900−d,900+d}, whose differences use only the four colours other than iii and qqq; so to exclude such a colouring it is sufficient to exclude these sets. It is a structure theorem for a hypothetical colouring: it does not prove S(6)≤1800S(6) \le 1800S(6)≤1800, and it does not decide whether such a colouring exists.

Formalization Note. R4(3)≤61R_4(3) \le 61R4​(3)≤61 is a hypothesis, not a proved fact of the mission; if R4(3)>61R_4(3) > 61R4​(3)>61, or if no Schur colouring of [1,1801][1, 1801][1,1801] with six colours exists, the statement holds vacuously, and whether such a colouring exists is open. TR(4,61)\mathrm{TR}(4, 61)TR(4,61) quantifies over colourings, with colours in any Finset of at most four naturals, of the pairs x<yx < yx<y of any finite set of at least 616161 naturals. The colouring ccc is a function on all of N\mathbb{N}N, constrained only on [1,1801][1, 1801][1,1801], and x=yx = yx=y is allowed in a Schur triple. VVV lies in range 1802 =[0,1801]= [0, 1801]=[0,1801], so the point 000 is a candidate member. The subtractions 901−d901 - d901−d (with d≤900d \le 900d≤900) and 1800−x1800 - x1800−x (with x∈Pix \in P_ix∈Pi​, so x≤1800x \le 1800x≤1800) are truncated subtractions on N\mathbb{N}N that never truncate here, and distances are Nat.dist.

Preamble
import Definitions.Def_ClassicalSchurColoring
import Definitions.Def_ClassicalSchurRamsey
import Mathlib

open ClassicalSchur
open Finset
Formal statement
theorem ClassicalSchur.schur_six_frontier_structure (hR : TriangleRamsey 4 61) {c : ℕ → Fin 6}
    (hc : SchurColoring 1801 c) :
    (∀ j, ((Icc 1 900).filter fun d => c d = j).card = 150) ∧
    (centralNbhd c 900).card = 301 ∧
    (∀ v ∈ centralNbhd c 900, ∀ i ≠ c 901, (colorNbhd c (centralNbhd c 900) v i).card = 60) ∧
    (∀ d, 0 < d → d ≤ 900 → c d = c 901 → c (901 - d) = c (901 + d)) ∧
    ∀ i ≠ c 901, (endpointNbhd c 900 i).card = 60 ∧
      (∀ x ∈ endpointNbhd c 900 i, 1800 - x ∈ endpointNbhd c 900 i ∧ 1800 - x ≠ x) ∧
      ∀ x ∈ endpointNbhd c 900 i, ∀ y ∈ endpointNbhd c 900 i, x ≠ y →
        c (Nat.dist x y) ≠ i ∧ c (Nat.dist x y) ≠ c 901 := by sorry
Source
A. McKenna, "S(6) ≤ 1801 if R₄(3) ≤ 61: a centred Schur bound and the structure at the frontier", Zenodo (2026), https://doi.org/10.5281/zenodo.23156099, Theorem 4.11 (§4.6, six colours under R_4(3) ≤ 61; informal form in Theorem 1.3). Lean source: https://github.com/mysticflounder/schur-centred-bound/blob/v1.0.1/ClassicalSchur/Frontier.lean#L549-L576 (release v1.0.1, doi:10.5281/zenodo.23156444).
Human review
  • Endorsed by Shuze Chen · Oct 4, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mysticflounder · Oct 4, 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, with reuse governed by our licensing terms.

How Prove2Me worksResearch paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me