Skip to content

Erdős Problem #670

Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: open (not automatically the live result).
Subject areas: geometry, distances.

Read this status correctly

The status above is reproduced from GCL's immutable external-catalogue snapshot. It is not a new claim of current mathematical resolution, a source-semantic review, an independent proof check, or GCL certification. A published proof and a GCL replay are separate events.

The question in plain language

Take \(n\) distinct points in Euclidean space and list all their pairwise distances. Suppose no two of those lengths differ by less than one. How wide must the configuration be?

Erdős Problem #670 asks whether the diameter, meaning the longest distance between any two points, is at least \((1+o(1))n^2\). The crucial issue is dimension. The maintained catalogue still regards the conjecture as OPEN when the dimension \(d\) is fixed as \(n\) grows. A 2026 construction by Boon Suan Ho refutes the version where \(d\) can grow with \(n\).

A small example

Four points on the real line at coordinates \(0,1,4,6\) determine the six distances

\[ 1,\quad2,\quad3,\quad4,\quad5,\quad6. \]

Every unordered point-pair has a different length and all six lengths are at least one apart. The diameter is six. This is a small Golomb-ruler fixture, not evidence that diameter six is optimal for an asymptotic quadratic conjecture.

What is known

There are exactly \(N=\binom n2\) unordered point pairs, hence \(N\) different positive distances under the assumed separation condition. Sort them as \(r_1<r_2<\cdots<r_N\). Since \(r_{i+1}-r_i\geq1\),

\[ \mathrm{diam}(A)=r_N\geq r_1+N-1>N-1. \]

Thus elementary counting gives a quadratic lower bound with leading constant one-half, rather than the conjectured one. Improving the coefficient requires geometric restrictions among triangle lengths, not only the fact that the distances occupy distinct unit-spaced intervals.

Fixed dimension versus growing dimension. The source attributes the desired bound for points on the line, \(d=1\), to Erdős's original discussion [Er97f, p.6]. Its exact original proof passage has not been inspected in this audit. For fixed higher dimensions, the source retains OPEN.

In contrast, the primary author preprint by Boon Suan Ho, Erdős's diameter conjecture for separated distances fails in high dimensions, submitted April 16, 2026, gives for every prime power \(q\) a set of \(n=q+1\) points in

\[ \mathbb R^{q^2+q}=\mathbb R^{n^2-n} \]

whose pairwise distances are all at least one apart from one another, yet whose diameter obeys

\[ \mathrm{diam}(A)\leq\left(1-\frac{1}{\pi^2}+o(1)\right)n^2. \]

The leading coefficient is approximately \(0.8987\), strictly below one. That disproves any dimension-uniform version of the conjecture. It does not disprove a claim where \(d=2\), or another fixed \(d\), must be chosen before sending \(n\to\infty\).

The author reports a Lean formalization. This is an external proof-status claim; the GCL cohort has not independently compiled or audited the formal source.

GCL's current state

The authenticated source-audit return for original GCL #1941 explicitly keeps these quantifiers apart. The 2026 author's original abstract and the maintained catalogue were checked; the original Erdős paper and complete Ho proof pages remain uninspected. The issue's disposition is BOUNDED_TASK_DESIGNED.

This is an unaccepted draft. No new fixed-dimensional lower bound is being asserted, and the high-dimensional published counterexample is not represented as a GCL discovery.

What remains

Read the original [Er97f, p.6] passage to determine whether Erdős intended a fixed ambient dimension. For a mathematical successor, first work in exactly two dimensions and seek an improvement beyond the universal counting bound \(\binom n2-1\). Any proposed bound must keep the minimum distance separation and dimension quantifiers explicit.

The Ho construction is relevant for a comparative source audit, including its exact source and optional Lean dependency check, but must not be treated as a planar counterexample. Critical editorial review, protected production merge, deployed MathJax inspection and exact-hash V2 admission remain separate.

Sources and claim boundaries

Boundary: no novel geometric theorem, internally replayed Ho proof, protected merge, deployed browser mathematics validation or editorial admission.

Catalogue and evidence provenance

  • Pinned catalogue source: teorth/erdosproblems at 536408716724.
  • GCL issue: #1941 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
  • GCL triage: STATEMENT_CAPTURE_REQUIRED.
  • Formal-statement coverage: No statement was confirmed in GCL's pinned formal-provider snapshot. Presence does not mean the statement or proof has been checked.
  • External mathematical status: consult current problem page and its dated bibliography.

Back to catalogue · Compendium introduction