Erdős Problem #659
Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: proved (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
How can a large collection of points create very few different distances overall, while still having a surprising amount of distance variety inside every small group of four?
Erdős Problem #659 asks for arbitrarily large \(n\) point sets in the Euclidean plane such that every four-point subset determines at least three different positive distances, but the total number of distinct distances is at most a constant multiple of
The answer is yes, using a carefully stretched version of an integer lattice. Earlier work by Moree and Osburn established the few-distances behavior of such lattices; Benjamin Grayzel's 2026 paper completes the necessary four-point local obstruction using a classification of exceptional distance patterns.
A small example
Take the four points
They form a rectangle of side lengths \(1\) and \(\sqrt2\). Its squared pairwise distances are precisely
There are already three different distances, so this four-point set passes the local requirement. More generally, place points on the stretched lattice
and consider its finite square-indexed box
It contains \(m^2\) points. Distances between lattice points are encoded by the integer quadratic form
Using integer squared distances makes equalities exact: there is no floating-point uncertainty about whether two distances coincide.
An independent complete enumeration of the four-subsets of these boxes for \(m=2,3,4,5,6\) checked respectively 1, 126, 1,820, 12,650 and 58,905 subsets. Every examined subset determined at least three distinct squared distances; none violated the local condition. The global numbers of distinct distances were respectively 3, 8, 14, 22 and 31. This finite experiment supports the construction but cannot establish the infinite theorem.
What is known
Why the global number is small. Each positive squared distance in \(P_m\) has the form
Consequently \(1\leq q\leq3(m-1)^2\). The relevant theorem about how many integers are represented by the positive-definite binary quadratic form \(u^2+2v^2\) is due to Bernays. In the required case, it supplies the estimate that the number of represented integers up to a large cutoff \(T\) is
Taking \(T\) proportional to \(m^2\) proves the published, Bernays-dependent global estimate
To treat every sufficiently large \(n\), choose \(m=\lceil\sqrt n\rceil\) and retain any \(n\) of the \(m^2\) points. Taking a subset cannot increase the global number of distances or introduce a forbidden four-point configuration. Since \(m^2=O(n)\), the same asymptotic upper bound survives.
Why every four-point subset has enough distances. A crucial fact is the classification of planar sets of four distinct points that determine only two distances, described by Antonella Perucca and used by Grayzel. Of the six similarity classes, every one either contains an equilateral triangle, is a square, or is similar to the four-vertex isosceles trapezoid of a regular pentagon.
The stretched lattice excludes all three possibilities:
- No square. A quarter-turn of the lattice vector \((a,b\sqrt2)\) would be \((-b\sqrt2,a)\). To remain in the lattice, its first coordinate would have to be an integer and its second coordinate an integer multiple of \(\sqrt2\). For integers \(a,b\), this forces \(a=b=0\), excluding a nondegenerate square.
- No equilateral triangle. Every nonzero squared lattice distance is an integer \(q\), while the squared area of an equilateral triangle of side length squared \(q\) is \(3q^2/16\). But any triangle formed by lattice vectors has squared area \(d^2/2\) for an integer determinant \(d\). Equality would require \(8d^2=3q^2\), impossible for positive \(q\) by comparing the parity of powers of the prime \(3\).
- No regular-pentagon trapezoid. Its ratio of squared long to short distances is the irrational number \((3+\sqrt5)/2\). Every ratio of two nonzero squared distances in our lattice is rational, as each is a positive integer.
The classification therefore rules out all two-distance four-point configurations. The elementary three exclusions have been reconstructed here; the completeness of Perucca's classification is still an external published theorem, not a newly proved classification by GCL.
The complete argument is attributed to Benjamin Grayzel, Solution to a Problem of Erdős Concerning Distances and Points, January 2026, together with the earlier lattice-distance analysis of Pieter Moree and Robert Osburn, Two-dimensional lattices with few distances, 2006. It uses both analytic number theory and four-point Euclidean geometry. Neither ingredient can simply be omitted.
GCL's current state
The original authenticated RESULT/1 source return records BOUNDED_TASK_DESIGNED and retains the original mathematical qualifications. This is an editorially reviewed explanatory article in the EC-BULK-005-R9 protected batch, effective only when that exact-hash transaction merges. The corresponding maintained Erdős statement and its bibliography were inspected. This editorial decision does not certify any full original theorem proof, independent external formalization, or Lean kernel check.
Preserved source-review qualifications: proved lean conditional on bernays axiom; pinned formal wrapper has sorry; four point classification external. See the immutable 128-case critical review record for the complete source-return caveats.
What remains
For documentary source closure, inspect the exact relevant pages of Grayzel's latest arXiv version, the primary Perucca classification proof and the Bernays theorem as used for the particular form \(u^2+2v^2\). Record whether the six similarity classes are completely and correctly exhausted and how the constant in the represented-number estimate depends on the quadratic form.
For any intended formal certification, replay the precise external Lean project using its locked dependencies. Check the theorem target, all transitive assumptions, the Bernays axiom, and any remaining proof holes. Such a replay would establish precisely what the Lean kernel verified and what remains conditional on an external published theorem. It must not be called a fully unconditional GCL certification merely because the public catalogue says PROVED (LEAN).
The editorial batch additionally requires source-critical review, consolidated protected publication, real-browser MathJax checking and separate exact-hash V2 admission.
Sources and claim boundaries
- T. F. Bloom, Erdős Problem #659, reported resolution and sources, plus public history.
- B. Grayzel, Solution to a Problem of Erdős Concerning Distances and Points, January 2026, latest cited author version. Abstract and theorem summaries inspected; full proof pages and classification not independently reconstructed in this audit.
- P. Moree and R. Osburn, Two-dimensional lattices with few distances, L'Enseignement Mathématique 52 (2006), 361–380. Publication and abstract inspected.
- A. Perucca, Four points, two distances, classification of six similarity classes. Full proof not independently reproduced.
- GCL pinned #659 formal source, specifically its external Bernays-axiom disclosure.
- GCL original issue source report, original issue-bound evidence and finite enumeration results.
- P. Erdős, [Er97e], p.531: original historical problem passage not independently inspected.
Authority boundary: an attributed published mathematical solution, independently checked elementary lattice observations and finite exact fixtures, with Bernays and the classification retained as named external theorems. No novel theorem claim, independently checked entire proof chain, internal Lean certification, protected production merge, browser readback is a protected publication verification gateor editorial admission.
Editorial source decision: R9 critical review, REVIEWED as a source-grounded explanation only. All open mathematical questions, primary-proof fidelity obligations, and external formalization caveats remain separate.
Catalogue and evidence provenance
- Pinned catalogue source: teorth/erdosproblems at
536408716724. - GCL issue: #1930 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
- GCL triage: OUTSIDE_STRICT_OPEN_TRIAGE.
- Formal-statement coverage: A provider statement was included in the pinned snapshot. Presence does not mean the statement or proof has been checked.
- External mathematical status: consult current problem page and its dated bibliography.