Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Failure of abc propagates to every n≥4n \ge 4n≥4

Open
NConjecture.abc_false_implies_n_false

by LG · Nov 22, 2026 · Mathlib 0df444a (Lean v4.33.1)

abc-conjecturenumber-theory

Hölzl–Kleine–Stephan, Theorem 3 (arXiv:2409.13439v2, p. 2), attributed there to Browkin–Brzeziński: "If the abc-conjecture is false then the nnn-conjecture is false for every n≥4n\ge 4n≥4." Equivalently (contrapositive), the nnn-conjecture at any single n≥4n\ge 4n≥4 implies abc. In ε\varepsilonε-form: if abc fails — some ε>0\varepsilon>0ε>0 admits arbitrarily large coprime triples with c≥Krad⁡(abc)1+εc\ge K\operatorname{rad}(abc)^{1+\varepsilon}c≥Krad(abc)1+ε for every KKK — then conjAt n fails for every n≥4n\ge 4n≥4.

The mechanism (HKS Remark 5(2), Browkin–Brzeziński Remark 2) is the quadruple (a3,b3,c3,−3abc)(a^3,b^3,c^3,-3abc)(a3,b3,c3,−3abc) built from an abc triple: it is admissible and has quality ≈3 L(a,b,c)\approx 3\,L(a,b,c)≈3L(a,b,c), so quality >1>1>1 for abc triples forces quality >3=2⋅4−5>3=2\cdot4-5>3=2⋅4−5 for quadruples; higher nnn use the analogous identities.

Formalization Note. The hypothesis is the negation of the full platform abc statement abc_conjecture_strong; the conclusion is ¬ conjAt n. The statement is the ε\varepsilonε-rendering of HKS Theorem 3 (the source phrases it via QA(n)=2n−5Q_{A(n)}=2n-5QA(n)​=2n−5, equivalent to conjAt n given the proved lower bound).

Preamble
import Mathlib
import Definitions.Def_NConjecture_Defs
Formal statement
namespace NConjecture

theorem abc_false_implies_n_false (n : ℕ) (hn : 4 ≤ n)
    (h : ¬ ∀ ε : ℝ, 0 < ε → ∃ K : ℝ, 0 < K ∧
      ∀ a b c : ℕ, 0 < a → 0 < b → 0 < c →
        Nat.Coprime a b → Nat.Coprime b c → Nat.Coprime a c →
        a + b = c →
        (c : ℝ) < K * (∏ p ∈ (a * b * c).primeFactors, (p : ℝ)) ^ (1 + ε)) :
    ¬ conjAt n := by sorry

end NConjecture
Source
R. Hölzl, S. Kleine, F. Stephan, Improved lower bounds for strong n-conjectures, arXiv:2409.13439v2 (2025). https://arxiv.org/abs/2409.13439 Theorem 3 p. 2 (attributed to Browkin–Brzeziński); construction in Remark 5(2)–(4) p. 3. Cf. J. Browkin, J. Brzeziński, Some remarks on the abc-conjecture, Math. Comp. 62(206) (1994), 931–939. https://cr.yp.to/bib/1994/browkin.pdf Remark 2 p. 934.
Read-back

What the Lean code literally says, in plain math · GLM (Z.ai), blind code-only pass — NOT an independent sub-agent

abc_false_implies_n_false (n : ℕ) (hn : 4 ≤ n) (h : ¬ ∀ ε : ℝ, 0 < ε → ∃ K : ℝ, 0 < K ∧ ∀ a b c : ℕ, 0<a → 0<b → 0<c → Nat.Coprime a b → Nat.Coprime b c → Nat.Coprime a c → a+b=c → (c:ℝ) < K*(∏ p ∈ (a*b*c).primeFactors, (p:ℝ))^(1+ε)) : ¬ conjAt n. For every natural number nnn with 4≤n4\le n4≤n, assume the negation of the abc statement: it is not the case that for every real ε>0\varepsilon>0ε>0 there exists K>0K>0K>0 such that all positive pairwise-coprime natural numbers a,b,ca,b,ca,b,c with a+b=ca+b=ca+b=c satisfy c<K rad⁡(abc)1+εc<K\,\operatorname{rad}(abc)^{1+\varepsilon}c<Krad(abc)1+ε (where rad⁡(abc)=∏p∈(abc).primeFactorsp\operatorname{rad}(abc)=\prod_{p\in(abc).\mathrm{primeFactors}}prad(abc)=∏p∈(abc).primeFactors​p). The conclusion is that conjAt n is false, i.e. the nnn-conjecture ε\varepsilonε-bound fails at this nnn: there exists some ε>0\varepsilon>0ε>0 for which no constant C>0C>0C>0 bounds max⁡i∣ai∣\max_i|a_i|maxi​∣ai​∣ by C rad⁡(∏i∣ai∣)2n−5+εC\,\operatorname{rad}(\prod_i|a_i|)^{2n-5+\varepsilon}Crad(∏i​∣ai​∣)2n−5+ε over all admissible nnn-tuples.

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