Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
A quantitative circular concentration threshold
Subsequent improvement: weighted_circle_energy replaces the sufficient exponent below by and improves the diameter-inactivity exponent to . The older proof is retained; its conclusions remain valid but weaker.
The circle incidence argument admits an explicit sublinear threshold. This remains a restricted result, not a solution of the two unrestricted questions.
Theorem. There is an absolute constant such that, if a circle contains points of , then
Take all the points on the circle, with angular set , as the anchor. Put
The chord formula gives . The pinned spectra used in the circle note have size at least . For points at different positive radii about the anchor center, their overlap is at most
If , apply the union lower bound to distinct radii. Their spectra would have union at least , a contradiction for large enough absolute . Thus there are occupied positive radii.
For their angular sets , the covering argument gives
Set , , , with sizes . The checked Plünnecke--Ruzsa input yields
The angular differences for any circle-pair type are covered by at most translates of . Compare the anchor chord spectrum with just one such translate for a different unordered radius pair. The cosine relation is nondegenerate, as proved in the circle note. Its overlap lemma, using and , gives at most
common values. There are other radius-pair types. Including the at most values incident with the possible center, only
anchor chord values occur outside the anchor circle. Each remaining positive chord value has all its pairs on that circle and therefore has unordered multiplicity at most . Consequently
The condition , for a sufficiently large absolute constant, implies the earlier size requirement and makes (2) at least . Since , this condition is equivalent up to an absolute constant to , proving (1).
Diameter inactivity tradeoff
Let be the number of points not incident with a diameter pair. There is an absolute such that
Write , , and . Suppose . For sufficiently small absolute this implies , , and the size hypothesis below. The diameter-active points lie on a constant-width boundary, and their palette satisfies
Apply the quantitative rank inverse with . Its parameter satisfies . It produces a circle containing at least
points. The rank lemma's condition holds under the assumed bound on , after reducing and absorbing small orders. The displayed exceeds when is sufficiently small. Formula (1) then gives
a contradiction. This proves (3), with small handled by a further decrease of .
In particular, bounded rarity forces every circle to contain points and forces . Rarity diverges when . An arbitrary set may have linearly many inactive points, so neither unrestricted question follows.
The author checked the exponents, the uniform cosine incidence input and its degeneracies, and the rank-inverse application; the argument is author-recorded. All external ingredients and their full primary passages are already recorded in the linked source notes.