Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

The involution ϕ\phiϕ is weight-preserving and fixes its pivot

Proved
BunkbedAux.phiExt_involution

by burkh4rt · Sep 12, 2026 · Mathlib 0df444a (Lean v4.33.1)

bunkbedcombinatoricspercolationprobability

The involution used in the robustness lemma of Gladkov--Pak--Zimin is defined by locating the first hyperedge eee along the path ρ\rhoρ whose refined state lies in Λ+2\Lambda^2_+Λ+2​, then applying the half-reflection η\etaη at eee and the level swap ρ\rhoρ to every hyperedge after eee, leaving the hyperedges before eee untouched. This theorem verifies that the resulting map ϕ=phiExt e\phi = \texttt{phiExt}\,eϕ=phiExte really is a weight-preserving involution.

Assume Ψ\PsiΨ is a legal refined configuration, that the state at eee lies in Λ+2\Lambda^2_+Λ+2​, and that no earlier hyperedge has its state in Λ+2\Lambda^2_+Λ+2​ — so eee is the pivot selected by the construction. Then:

  1. ϕ(Ψ)\phi(\Psi)ϕ(Ψ) is again legal;
  2. the state of ϕ(Ψ)\phi(\Psi)ϕ(Ψ) at eee still lies in Λ+2\Lambda^2_+Λ+2​;
  3. no hyperedge before eee has its ϕ(Ψ)\phi(\Psi)ϕ(Ψ)-state in Λ+2\Lambda^2_+Λ+2​;
  4. ϕ(ϕ(Ψ))=Ψ\phi(\phi(\Psi)) = \Psiϕ(ϕ(Ψ))=Ψ;
  5. ∏hwt(ϕ(Ψ)(h))=∏hwt(Ψ(h))\prod_h \texttt{wt}(\phi(\Psi)(h)) = \prod_h \texttt{wt}(\Psi(h))∏h​wt(ϕ(Ψ)(h))=∏h​wt(Ψ(h)).

Points 2 and 3 together say that ϕ\phiϕ fixes its own pivot: the first hyperedge in Λ+2\Lambda^2_+Λ+2​ is the same for Ψ\PsiΨ and for ϕ(Ψ)\phi(\Psi)ϕ(Ψ). This is exactly what makes point 4 meaningful — applying the construction twice recovers Ψ\PsiΨ only because the second application selects the same eee, so that η\etaη is applied twice at eee and the level swap twice beyond it, and both are involutions.

Point 5 is where the refinement of the state space earns its keep. The level swap preserves weights for the trivial reason that a pair's weight is symmetric under transposition; but η\etaη genuinely moves mass between different pairs, and preserves weight only because the refinement split off from (abc,a∣b∣c)(abc, a|b|c)(abc,a∣b∣c) precisely the sub-event of weight Pab∣cPac∣bP_{ab|c}P_{ac|b}Pab∣c​Pac∣b​, which is the weight of (ab∣c,ac∣b)(ab|c, ac|b)(ab∣c,ac∣b).

Mathematical role. Together with the path-surgery lemma, which says that ϕ\phiϕ exchanges the same-level and cross-level connection events, this yields the cancellation: the configurations meeting Λ+2\Lambda^2_+Λ+2​ contribute equally to the two events and drop out of their difference. The robustness lemma is thereby reduced to the configurations that avoid Λ+2\Lambda^2_+Λ+2​ altogether, where the alternative bunkbed hypergraph model and Hollom's computation take over.

Preamble
import Definitions.Def_BunkbedWZSurgery
open Bunkbed Bunkbed.Hyper Finset SimpleGraph
Formal statement
namespace BunkbedAux

/-- **`phiExt e` is a weight-preserving involution that fixes its own pivot.** If `e` is the
first hyperedge whose state lies in `Λ²₊`, then `phiExt e` maps legal configurations to legal
configurations, keeps `e` as the first hyperedge in `Λ²₊`, is an involution, and preserves the
configuration weight. -/
theorem phiExt_involution (P : WZ → ℚ) (e : Fin 6) (Ψ : Fin 6 → WZExt)
    (hleg : ∀ h, WZExt.legal (Ψ h) = true)
    (he : WZExt.inLamSq (Ψ e) = true)
    (hmin : ∀ h : Fin 6, h.val < e.val → WZExt.inLamSq (Ψ h) = false) :
    (∀ h, WZExt.legal (phiExt e Ψ h) = true)
    ∧ WZExt.inLamSq (phiExt e Ψ e) = true
    ∧ (∀ h : Fin 6, h.val < e.val → WZExt.inLamSq (phiExt e Ψ h) = false)
    ∧ phiExt e (phiExt e Ψ) = Ψ
    ∧ (∏ h, WZExt.wt P (phiExt e Ψ h)) = ∏ h, WZExt.wt P (Ψ h) := by
  sorry

end BunkbedAux
Source
N. Gladkov, I. Pak, A. Zimin, The bunkbed conjecture is false, PNAS 122 (2025) no. 24, e2420725122, https://doi.org/10.1073/pnas.2420725122, Section 3.5 (p. 5): the definition of phi and the assertion that it is a weight-preserving involution on X.

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 paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me