Alec Kriebel

A four-period counterexample to the conjectured elliptic billiard invariant k107

Read the paper (PDF) · Download PDF · Supporting ZIP · Zenodo record

Plain-English explanation · Read on this page · Collaboration


Abstract

Table 2 of Reznik, Garcia and Koiller's 2020 preprint and 2021 article conjectures the constancy of (A′/A) ∏ᵢ sin(θᵢ/2) when the period is divisible by four. Here A′ is the area of the outer tangent polygon and the angles belong to the original billiard polygon. Two primitive convex four-periodic trajectories in the same confocal caustic family give, exactly, 288/625 and 625/1152. We give a self-contained geometric certificate of the common continuous family and the contradiction. The four-period construction and its area identities are prior work. The same correction appears in Ferudun's contemporary note, which we credit explicitly. No exclusive priority or independent discovery is claimed.

The idea in plain English

Two billiard paths bounce four times in the same ellipse and touch the same inner ellipse. The expression proposed as constant takes a different value on each path.

The precise result

Let A be the area of the billiard quadrilateral and A′ the area of the outer quadrilateral formed from tangents to the billiard ellipse at its vertices. Let θi be the interior angles of the original quadrilateral. Table 2 proposes constancy of (A′/A) ∏i sin(θi/2) for periods divisible by four.

The ellipse with semiaxes 4 and 3 admits a continuous four-periodic family with the same strictly nested confocal caustic. Its two explicit phases have the following products.

Orbit phaseProduct
Axial diamond288/625
Axis-aligned rectangle625/1152

Why the phases are comparable

The proof computes the reflected directions and shows that every side is tangent to the same proper confocal ellipse at a point of the side segment. Along the continuous family both the two areas and the half-angle factors can be calculated directly in the same orthonormal coordinates.

Scope

Both witnesses are primitive convex four-periodic trajectories in the same family. The comparison addresses the product printed as k107. It leaves other area and angle combinations to their own proofs.

Background and prior work

The k107 product appears in Table 2 of Reznik, Garcia and Koiller’s published article and in their preprint. The four-period geometry and area identities precede this note. Ferudun’s preprint gives the same correction and exact witnesses.

Status and AI use

Unrefereed preprint. AI tools assisted substantially in developing, writing and checking the work. Independent human review is welcome.

Read the paper

Open the published PDF · Download PDF · Original PDF on Zenodo

Supporting materials

Download k107_counterexample_support.zip · Original support archive on Zenodo

Collaborate on this result

I welcome researchers in elliptic billiards or classical geometry who would like to check the common-caustic certificate, compare the overlapping proofs, or help develop a revised exposition.

Authorship and responsibilities would be discussed in relation to substantive contributions. Contact Alec.

Cite this paper

Kriebel, Alec (2026). A four-period counterexample to the conjectured elliptic billiard invariant k107. Version 1.0. Zenodo. 10.5281/zenodo.23231145.

Download BibTeX

Licensing

The deposited work is available under Creative Commons Attribution 4.0.