Focal pedal area ratios of closed elliptic billiards
A closed billiard path inside an ellipse produces two sets of perpendicular feet from each focus: one on the path’s chord lines and another on the tangents to the ellipse. This note proves that passing from the first construction to the second multiplies both focal signed areas by the same positive factor. That factor stays fixed as the orbit moves through its family, while the ratio between the two foci can change.
Abstract
Reznik, Garcia and Koiller observed equality between the two focal pedal area ratios of a closed elliptic billiard and its outer tangent polygon. We prove this equality for every primitive convex or star orbit with a nondegenerate elliptical caustic. More precisely, the outer and original signed pedal areas are proportional by a positive constant independent of phase and common to both foci. The proof identifies a common pair of simple poles for the two cyclic area traces and matches their residues on a compact torus. It reconstructs previously used local cancellation arguments within established elliptic-function methods. We also give an exact triangular certificate showing that the common focal ratio need not be constant over the billiard family. Known even and triangular cases and classical methods are credited explicitly. This is an extensively AI-assisted, unrefereed research note, with a reproducible verification supplement.
The precise result
Fix a noncircular ellipse and a strictly nested, nondegenerate confocal elliptical caustic. Let a primitive closed orbit have N ≥ 3 sides and coprime turning number 0 < τ < N/2. Both convex and self-intersecting star orbits are included. The outer polygon is formed by consecutive intersections of the boundary tangents at the orbit’s vertices.
For the two foci F₊ and F₋, write A₊ and A₋ for the original chord-pedal areas, and B₊ and B₋ for the outer tangent-pedal areas. A pedal vertex is the perpendicular foot on the infinite supporting line. The vertices retain traversal order, and area means the signed shoelace sum.
The same C₀ works at both foci and is independent of the phase w within the fixed billiard family. In positive traversal all four signed areas are strictly positive, so the focal ratios are defined at every real phase and satisfy A₊/A₋ = B₊/B₋.
Reversal negates all four areas. A positive integral repetition multiplies all four by the repetition count. The identities and nonvanishing survive both operations. The separate circular regular-star case has coincident foci, both focal ratios equal to one, and C₀ = sec²(πτ/N).
A ratio that can change
The equality of the two focal ratios does not make their common value constant. An exact triangular orbit in x²/21 + y²/16 = 1, with vertices (√21, 0), (−3√21/5, 16/5), (−3√21/5, −16/5), has a confocal caustic with squared semiaxes 189/25 and 64/25. Its multiplier is C₀ = 125/24 and its focal ratio is
Central inversion gives another phase of the same fixed-caustic family and exchanges the foci. Its ratio is the reciprocal, below one. The ratio equality and common multiplier remain valid.
Why the proof works
Canonical Jacobi coordinates turn the two area sums into meromorphic functions of the phase. Local cancellations reduce their possible singularities to the same pair of simple poles. A geometric normal-angle argument proves strict real positivity, including for stars. Matching residues leaves a holomorphic function on a compact torus; the symmetry forces that difference to vanish. This proves the proportionality throughout the family.
The supplement contains exact algebra and finite high-precision real and complex diagnostics. Those diagnostics are not interval-certified bounds and do not establish the universal theorem by sampling. Follow the archive’s reproduction instructions with Python assertions enabled.
Scope and prior work
The theorem concerns signed areas on ordered supporting lines. It excludes hyperbolic or collapsed caustics, two-period diameter walks, unsigned sums of lobe areas and projections clamped to side segments.
The original focal equality is due to Reznik, Garcia and Koiller, observed in April 2020. Their journal Table 7, page 349, labels it k607; the earlier arXiv editions use k605 and k606. Their primitive-even equality proof by central symmetry is prior work. The triangular positive case and classical nonconstancy consequence are also credited as consequences of older results, not separate discoveries.
The canonical coordinates are due to Stachel, with standard Jacobi identities from DLMF. Principal-part and elliptic-function methods are established in Khare, Lakshminarayan and Sukhatme and Roitman, Garcia and Reznik. The specialized local cancellations also appeared in the credited original-foot and outer-foot repository arguments. The supplement explains the triangular deduction from Querret, Fierobe, Garcia–Reznik and Helman and collaborators, including the phase relation needed beyond a locus shape.
Related work includes Alper Ferudun’s Focal Pedal-Antipedal Area Products, Steiner-Centroid Pedal Area Ratios and Outer-Pedal Area Products, all dated 1 October 2026, versions 1.0, 1.0 and 1.1 respectively. These concern related quantities or period classes; earlier even-period multiplier consequences are credited. This note makes no claim to a new observation, new general method, separate triangular novelty or absolute first priority. Historical coverage is bounded, and the old general locus/CAS derivations and every final journal edition have not been independently recertified.
Paper and supporting materials
Zenodo record and DOI · Open the PDF · Download the PDF · Source and verification archive
The embedded paper and downloads are unchanged copies of version 1.0, dated 4 October 2026. Open the Zenodo original.
Citation
Kriebel, Alec. Focal pedal area ratios of closed elliptic billiards. Unrefereed preprint, version 1.0, 4 October 2026. Zenodo. https://doi.org/10.5281/zenodo.23149775. BibTeX.
Collaboration
I welcome substantive mathematical comments, independent verification and contributions toward a journal submission. Authorship would be discussed in relation to actual contributions.