Erdős Problem #662 — Consider the triangular lattice with minimal distance between two points 11.

About 29 years old · traced to

Consider the triangular lattice with minimal distance between two points 11. Denote by f(t)f(t) the number of distances from any points ≤t\leq t. For example f(1)=6f(1)=6, f(3)=12f(\sqrt{3})=12, and f(3)=18f(3)=18. Let x1,…,xn∈R2x_1,\ldots,x_n\in \mathbb{R}^2 be such that d(xi,xj)≥1d(x_i,x_j)\geq 1 for all i≠ji\neq j. Is it true that, provided nn is sufficiently large depending on tt, the number of distances d(xi,xj)≤td(x_i,x_j)\leq t is less than or equal to f(t)f(t) with equality perhaps only for the triangular lattice? In particular, is it true that the number of distances ≤3−ϵ\leq \sqrt{3}-\epsilon is less than 11?

References

Progress summary

Refreshed
Open

No proof or counterexample has been found, and the problem’s published wording appears defective.

Erdős, Lovász, and Vesztergombi ask whether sufficiently large one-separated point sets in the plane can have more pairs at distance at most tt than the corresponding triangular-lattice configuration, for every threshold tt. The original formulation appears to contain numerical errors, and its stronger literal reading is described as nonsensical.

Known results

  • Fejes Tóth, 1942: hexagonal packing is optimal for the related densest-packing problem; this does not establish the full threshold-by-threshold extremality claim.
  • The discussion proposes corrected triangular-lattice counts such as f(2)=18f(2)=18, f(7)=30f(\sqrt{7})=30, and f(3)=36f(3)=36, but supplies no proof of the conjectured extremality.

Current status (as of March 2026): The exact short-distance extremality problem remains open; only related packing results and clarification of defects in the formulation are recorded.

Sources

Solutions 0

No solutions have been posted yet.