Erdős Problem #661 — Are there, for all large nn, some points x1,…,xn,y1,…,yn∈R2x_1,\ldots,x_n,y_1,\ldots,y_n\in \mathbb{R}^2 such that the number of distinct distances d(xi,yj)d(x_i,y_j) is o(nlog⁡n)?o\left(\frac{n}{\sqrt{\log n}}\right)?

About 36 years old · traced to

Are there, for all large nn, some points x1,…,xn,y1,…,yn∈R2x_1,\ldots,x_n,y_1,\ldots,y_n\in \mathbb{R}^2 such that the number of distinct distances d(xi,yj)d(x_i,y_j) is o(nlog⁡n)?o\left(\frac{n}{\sqrt{\log n}}\right)?

References

Progress summary

Refreshed
Open

No public proof or counterexample has been found, and the problem remains open.

Erdős Problem #661 asks whether, for every sufficiently large nn, two nn-point sets in the plane can have fewer than n/log⁡nn/\sqrt{\log n} distinct cross-distances in the little-oo sense. A current problem ledger marks it verified open; its opening date is unknown, and no retrieved source reports a resolution.

Current status (as of September 2026): The existence of such configurations remains open; no verified proof, counterexample, or exact advance has been recorded.

Sources

Solutions 0

No solutions have been posted yet.