Local rigidity for symplectic billiards

Daniel Tsodikovich

REVIEW 2 major objections 2 minor 19 references

Local rigidity for symplectic billiards

T0 review · 2 major / 2 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper proves that any domain sufficiently close to an ellipse whose symplectic billiard dynamics is rationally integrable must itself be an ellipse.

desk verdict New local rigidity result for symplectic billiards, but the proof of the key contraction lemma contains an arithmetic error that needs fixing; the result looks salvageable. read the letter →

arxiv 2501.08849 v2 pith:AS2KD4LX submitted 2025-01-15 math.DS

Daniel Tsodikovich This is my paper · ORCID

classification math.DS MSC 37C8337E4037J4653A15

keywords symplecticbilliardslocalrigidityrationalintegrabilityellipsesaffinearc-lengthparametrizationtwistmapsFouriercoefficientsperiodicorbits

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Symplectic billiards are a variant of the classic billiard where the reflection law is governed by area rather than length. This paper establishes a local rigidity theorem: for any ellipse and any fixed smoothness bound K, there is a small $C^{1}$ neighborhood such that any domain in that neighborhood whose symplectic billiard map is rationally integrable is in fact an ellipse. The result is the symplectic analogue of the known local rigidity theorem for Birkhoff billiards, and it shows that ellipses are isolated in the class of rationally integrable domains. The proof works by showing that a rationally integrable deformation of an ellipse can be replaced by another ellipse that is strictly closer to the domain, so a minimal-distance argument forces the deformation to disappear.

What carries the argument

The argument is carried out in the affine arc-length parametrization of the ellipse, where the boundary of the deformed domain is written as γ(t)=e_{a,b}(t)+n(t)N(t) with N(t)=e_{a,b}(t) a rescaling of the affine normal. In these coordinates the unperturbed q-periodic orbits are equally spaced, and the implicit function theorem gives explicit control over how they shift under the deformation. The workhorse is a chain of quantitative lemmas: an action expansion (Lemma 3.1), a Fourier-coefficient estimate (Lemma 3.3), and an approximation result (Lemma 3.6) showing that a rationally integrable domain admits another ellipse whose residual deformation has $C^{1}$ norm bounded by a small power of the original deformation. The final minimality step compares distances to the family of ellipses and concludes that the only rationally integrable domains arbitrarily close to an ellipse are ellipses themselves.

What would settle it

Exhibit a non-elliptic domain that is arbitrarily $C^{1}$-close and $C^{127}$-bounded-close to an ellipse and whose symplectic billiard map has invariant curves of q-periodic orbits for every q≥3; the theorem declares such a domain impossible. A more local check of the mechanism is to measure the actions of q-periodic orbits for a small non-elliptic perturbation and test whether inequality (3.2) (Lemma 3.1) holds—if a rationally integrable candidate violates that estimate, the proof's core quantitative step fails.

Watch

Extended reading notes

Core claim

The central claim, Theorem 1, is a rigidity statement: given an ellipse E and any K>0, there exists ε>0 such that every domain Ω that is $C^{127}$ K-close to E and $C^{1}$ ε-close to E with rationally integrable symplectic billiards is an ellipse. Rational integrability means that for every integer q≥3 the map has an invariant curve of q-periodic orbits. The theorem is proved by estimating the action of such periodic orbits: Lemma 3.1 bounds the deviation of the action from the elliptic value by a power of the deformation size times $q^{31}$, Lemma 3.3 turns this into a bound on Fourier coefficients of the deformation function, and Lemma 3.6 uses these bounds to construct a strictly closer ellipse. A compactness argument over the family of nearby ellipses then forces the deformation function to vanish identically.

Load-bearing premise

The proof depends on the assertion, made without proof in Section 3.3, that an affine map sending the ellipse to the unit circle carries the affine normal and the deformation function to those of the circle, so that the existing lemmas for radial deformations of a circle apply verbatim; if this transfer failed, the main estimates would lose their foundation.

Editorial extensions

If this is right

No non-elliptic rationally integrable domain can accumulate on an ellipse: if a sequence of rationally integrable domains converges to an ellipse with the regularity and closeness of the theorem, the domains are eventually ellipses.
Any exotic integrable symplectic billiard—one that is integrable without a full foliation—must stay outside a definite C^1 neighborhood of every ellipse, or must fail to have some rational invariant curve.
Because affine maps preserve symplectic billiards, the rigidity statement transfers to every ellipse once it is proved for a single one, so the entire elliptic family is locally rigid.
The quantitative exponent 7875/7874 shows the distance to the elliptic family is controlled by a superlinear power of the initial deformation size, giving a definite (if very weak) rate of rigidity.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A plausible but unproven strengthening is that the regularity requirement (C^127 and the 7875/7874 exponent) is far from optimal; the method suggests a trade-off between the C^k norm assumed and the resulting power, which could be explored numerically for small k.
The same action-versus-Fourier mechanism may apply to other twist maps whose generating function comes from an area form, such as magnetic twists or higher-dimensional symplectic billiards, since the proof only uses the twist property and the affine structure.
If the affine-normal transfer in Section 3.3 could be proved directly rather than cited, a consequence would be that the rigidity survives under more general centro-affine deformations, not just radial ones; this is a testable extension of the paper's argument.
A numerical experiment could look for rationally integrable near-elliptic domains by searching for periodic orbit families of all small rotation numbers; the theorem predicts that any such family forces the domain to be an ellipse, giving a practical check on the rigidity.

Share X Bluesky LinkedIn Reddit HN

Formalized claims in Lean

Claim #1: The central claim, Theorem 1, is a rigidity statement: given an ellipse E and any K>0, there exists ε>0 such that every domain Ω that is $C^{127}$ K-close to E and $C^{1}$ ε-close to E with rationally integrable symplectic billiards is an ellipse. Rational integrability means that for every integer q≥3 the map has an invariant curve of q-periodic orbits. The theorem is proved by estimating the act

/-- @claim 1 The central claim, Theorem 1, is a rigidity statement: given an ellipse E and any K>0, there exists ε>0 such that every domain Ω that is $C^{127}$ K-close to E and $C^{1}$ ε-close to E with rationally integrable symplectic billiards is an ellipse. Rational integrability means that for every integer q≥3 the map has an invariant curve of q-periodic orbits. The theorem is proved by estimating the act -/
def central_claim : Prop :=

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper proves the first local rigidity theorem for symplectic billiards: a domain sufficiently close to an ellipse that is rationally integrable must itself be an ellipse. That is a natural analogue of Avila–De Simoi–Kaloshin for Birkhoff billiards, and it is genuinely new. The overall strategy is a clean adaptation of the ADK scheme — action estimates for periodic orbits, Fourier coefficient bounds, then a contraction argument that forces the deformation to vanish. Lemmas 3.1–3.3 are worked out in detail with explicit constants, and the exposition is clear.

The soft spot is in Lemma 3.6, and it is load-bearing. The proof chooses α = 1/31 and asserts that 4 − 63α = 2 + α = 63/31. That identity is wrong: 4 − 63/31 = 61/31, not 63/31. The equality would require α = 1/32. As a result the Fourier tail estimate (3.34) gives ||n⊥||_{L2} ≤ C ||n||^{61/62}, which is sublinear. The subsequent Sobolev interpolation then cannot produce a C^1 bound with exponent greater than 1, and the contraction inequality (3.36) that closes the proof in §3.4 is unsupported. This is not a cosmetic typo; it breaks the central mechanism.

That said, the fix is straightforward: take α = 1/32 instead. Then 4 − 63α = 2 + α = 65/32, and the rest of the calculation goes through with the final exponent changed accordingly. I checked the surrounding assumptions: condition (3.31) still holds for small ||n||_{C1}, and the interpolation step would then yield a superlinear bound. So the theorem is very likely true, but the present text has a real gap.

The other concern, noted in the accompanying report, is that Section 3.3 asserts the affine transfer of the deformation structure without proof. That is based on standard facts about affine normals, but the paper would be stronger if it gave a few lines of justification. The heavy reliance on Lemmas 16–19 of ADK is not itself a problem — that is a normal citation pattern, and the results are available.

Who should read this? Anyone working on integrability and rigidity of billiards, symplectic twist maps, or the Birkhoff–Poritsky conjecture in its various forms. The paper deserves a serious referee. My recommendation is to send it to peer review, but to insist that the author fix the arithmetic error and re-verify the interpolation before publication. It is a solid contribution, just not quite ready as written.

Referee Report

2 major / 2 minor

Summary. The paper proves a local rigidity theorem for symplectic billiards: every convex domain that is sufficiently close to an ellipse (C^127-close in a coarse sense and C^1-close in a fine sense) and whose symplectic billiard map is rationally integrable must itself be an ellipse. The proof closely follows the strategy of Avila, De Simoi, and Kaloshin for Birkhoff billiards: it estimates the action of periodic orbits, derives Fourier bounds for the deformation function, constructs a nearby ellipse that is even closer, and uses a minimality argument to conclude the deformation vanishes.

Significance. If the proof is valid, the result is a significant contribution to the rigidity theory of symplectic billiards, establishing the first local rigidity statement near ellipses for this system and showing that the integrability obstruction is strong enough to force the domain to be an ellipse. The paper is carefully structured, provides explicit neighborhood and norm bounds in the implicit-function-theorem step, and makes the dependence on the ellipse's normalized area explicit throughout; it also names the imported lemmas from [2] rather than hiding them. These features make the argument verifiable in principle. However, the proof as written contains arithmetic and sign errors in a load-bearing interpolation step, which currently invalidate the central contraction estimate.

major comments (2)

[Section 3.3, Eqs. (3.32)–(3.34)] The claimed identity '4 − 63α = 2 + α = 63/31' for α = 1/31 is arithmetically false. With α = 1/31, the sum over low frequencies in (3.32) has exponent 4 − 63/31 = 61/31, while the high-frequency sum in (3.33) has exponent 2 + 1/31 = 63/31. The low-frequency term therefore dominates, and (3.34) should read ∥n⊥∥_{L2} ≤ C ∥n∥^{61/62}, which is sublinear. Since 61/62 < 1, the subsequent Sobolev interpolation in (3.35) cannot yield a C^1 bound with exponent greater than 1, and the contraction estimate (3.36) in Section 3.4 — the key step of the minimality argument — is not established. This is a load-bearing error, not a cosmetic one; the theorem may be repairable by taking α = 1/32, but the present text does not contain a valid derivation.
[Section 3.3, Eq. (3.35)] The interpolation inequality as stated has the exponent ε^{j/(j−127)}. For j = 2 this exponent is −125/2, which is negative. With the choice ε = ∥n⊥∥_{C^1}^{7875/7874}, the second term becomes ε^{−125/2} ∥n⊥∥_{L2}, which is not bounded by a constant times ∥n∥^{7875/7874}; it diverges as ∥n∥ → 0 unless ∥n⊥∥_{L2} is exponentially small, which is not shown. The standard Gagliardo–Nirenberg inequality would require a positive exponent (ε^{j/(127−j)}), so this appears to be a sign error. As written, however, the displayed formula does not support the C^1 bound of n⊥ that the proof needs.

minor comments (2)

[Section 3.3, first paragraph] The assertion that an affine map sending the ellipse E to the unit circle carries the affine normal of E to the affine normal of the circle, so that Lemmas 16–19 of [2] apply verbatim, is stated without proof. This is a standard fact in affine differential geometry, but a brief justification or a more precise reference would make the transfer rigorous and would avoid a gap in the derivation of Lemma 3.4.
[Throughout] There are several typos, including 'percise' in the introduction and 'deonminator' in the proof of Lemma 3.2; these should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the proof adapts external, non-self-cited lemmas and uses the integrability hypothesis as an input, not as the conclusion.

full rationale

The derivation chain is not circular. The dynamical estimates in Lemmas 3.1–3.3 are obtained directly from the symplectic billiard equations (2.1) and (3.8), together with the rational integrability hypothesis of Definition 2.1; the action-constancy of q-periodic invariant curves is used as a hypothesis to bound Fourier coefficients, which is the intended direction of the argument. The later low-frequency approximation lemmas are imported from Avila–De Simoi–Kaloshin [2], which is an external source with no author overlap, and they are geometric/estimative statements rather than reformulations of Theorem 1. The Section 3.3 transfer from an ellipse to the unit circle by an affine map is a standard affine-normal fact, and even if its proof is only sketched, it is a supporting geometric lemma rather than a circular reduction. No equation in the paper is equivalent to the conclusion that a rationally integrable near-ellipse is an ellipse; in particular, the rational integrability assumption is never derived from the conclusion. There is no fitted parameter renamed as a prediction, and no load-bearing self-citation chain. The skeptic's arithmetic objection to Lemma 3.6, concerning the claimed identity 4−63α=2+α for α=1/31, is a correctness risk rather than a circularity, and it does not change the circularity verdict. Overall, the central claim has independent content and the mild external reliance does not constitute circular reasoning.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The proof introduces no fitted parameters and no new mathematical or physical entities. It relies on the rational-integrability hypothesis, a C^127 regularity assumption, and two external inputs: the geometric approximation lemmas of [2] and the affine-invariance of the normal. All constants are universal or made explicit; none are fitted to the target statement.

assumptions (5)

domain assumption Rational integrability: for all q >= 3 there exists an invariant curve of rotation number 1/q consisting entirely of q-periodic orbits.
This is the hypothesis of Theorem 1 and is used in Lemma 3.3 to ensure that the action is constant on each invariant curve.
domain assumption The deformation function n is C^127 smooth with small C^1 norm, satisfying the smallness condition (3.31).
Needed for the Sobolev interpolation inequality (3.35) and for controlling high-order derivatives of the deformation in Lemma 3.6.
standard math Lemmas 16-19 of Avila, De Simoi and Kaloshin (Annals of Mathematics 2016), concerning radial deformations of a circle and approximation by ellipses, are valid and applicable.
These published lemmas are imported verbatim in Section 3.3 and are used to prove Lemma 3.4, which is essential for the approximation argument.
domain assumption Affine invariance of symplectic billiards and of the affine normal and affine arc-length parametrization.
Used in Section 3.3 to transfer the circle-based geometric lemmas to an arbitrary ellipse, and in the introduction to state that all ellipses behave the same way.
standard math Standard twist map theory: action is constant on invariant curves, and all orbits on an invariant curve with rational rotation number are periodic.
Invoked in Section 2.1 and used in Lemma 3.3; cited to [3,11,12].

how reviews work

0 comments

Cite this review

Pith. "Pith review of Local rigidity for symplectic billiards." pith.science (2026). https://pith.science/paper/AS2KD4LX

@misc{pith2026250108849,
  author       = {Pith},
  title        = {Pith review of: Local rigidity for symplectic billiards},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AS2KD4LX}},
  note         = {Machine review of arXiv:2501.08849}
}

read the original abstract

We show a local rigidity result for the integrability of symplectic billiards. We prove that any domain which is close to an ellipse, and for which the symplectic billiard map is rationally integrable must be an ellipse as well. This is in spirit of the result of Avila, De Simoi, and Kaloshin for Birkhoff billiards.

Figures

Figures reproduced from arXiv: 2501.08849 by the authors.

**Figure 1.** The symplectic billiard collision law. 2.1 Symplectic Billiards Symplectic billiards were introduced by Albers and Tabachnikov in [1]. In the plane (which is the only case which is of interest for us in this work), it is defined as follows. Given a smooth convex curve γ(t), three points γ(t1), γ(t2), and γ(t3) are consecutive points of a symplectic billiard orbit if and only if the tangent at γ(t2) is positively par… view at source ↗

**Figure 2.** Same domain (Ω, in red) seen as a deformation of two different [PITH_FULL_IMAGE:figures/full_fig_p020_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

19 extracted references · 19 canonical work pages

[2]

Avila, J

A. Avila, J. De Simoi, and V. Kaloshin. An integrable deformation of an ellipse of small eccentricity is an ellipse. Annals of Mathematics , pages 527–558, 2016

work page 2016
[1]

Albers and S

P. Albers and S. Tabachnikov. Introducing symplectic billiards. Ad- vances in mathematics (New York. 1965) , 333:822–867, 2018

work page 1965
[3]

V. Bangert. Mather Sets for Twist Maps and Geodesics on Tori . Dy- namics reported. Wiley, 1988

work page 1988
[4]

Baracco and O

L. Baracco and O. Bernardi. Totally integrable symplectic billiards are ellipses. Advances in Mathematics , 454:109873, 2024

work page 2024
[5]

Area spectral rigidity for axially symmetric symplectic billiard tables

L. Baracco, O. Bernardi, and A. Nardi. Area spectral rigidity for axially symmetric and Radon domains. arXiv preprint arXiv:2410.12644, 2024

work page Pith review arXiv 2024
[6]

Baracco, O

L. Baracco, O. Bernardi, and A. Nardi. Bialy-Mironov type rigidity for centrally symmetric symplectic billiards. Nonlinearity, 37(12):125025, nov 2024

work page 2024
[7]

Bialy and A

M. Bialy and A. E. Mironov. The Birkhoff-Poritsky conjecture for centrally-symmetric billiard tables. Annals of Mathematics , 196(1):389 – 413, 2022

work page 2022
[8]

De Simoi, V

J. De Simoi, V. Kaloshin, Q. Wei, and H. Hezari. Dynamical spec- tral rigidity among Z2-symmetric strictly convex domains close to a cir- cle (Appendix B coauthored with H. Hezari). Annals of Mathematics , 186(1):277–314, 2017

work page 2017

Show all 19 references

[9]

Fierobe, A

C. Fierobe, A. Sorrentino, and A. Vig. Deformational spectral rigidity of axially-symmetric symplectic billiards. arXiv preprint arXiv:2410.13777, 2024

2024 arXiv
[10]

Gilbarg and N

D. Gilbarg and N. S. Trudinger. Elliptic partial differential equations of second order, volume 224. Springer, 1977

1977
[11]

C. Gole. Symplectic twist maps: global variational techniques . Advanced Series In Nonlinear Dynamics. World Scientific Publishing Co. Pte. Ltd, Singapore, 2001. 27

2001
[12]

M. Herman. Existence et non existence de tores invariants par des diff´ eomorphismes symplectiques.S´ eminaire´Equations aux d´ eriv´ ees par- tielles (Polytechnique) dit aussi ”S´ eminaire Goulaouic-Schwartz”, 1987-

1987
[13]

Huang, V

G. Huang, V. Kaloshin, and A. Sorrentino. Nearly circular domains which are integrable close to the boundary are ellipses. Geometric and Functional Analysis, 28:334–392, 2018

2018
[14]

Jindal, D

A. Jindal, D. Chatterjee, and R. Banavar. Estimates of the size of the domain of the implicit function theorem: a mapping degree-based approach. Mathematics of Control, Signals, and Systems, 36(1):139–175, Sept. 2023

2023
[15]

Kaloshin and A

V. Kaloshin and A. Sorrentino. On the local Birkhoff conjecture for convex billiards. Annals of Mathematics , 188(1):315–380, 2018

2018
[16]

I. Koval. Local strong Birkhoff conjecture and local spectral rigidity of almost every ellipse. arXiv preprint arXiv:2111.12171 , 2021

2021 arXiv
[17]

Kulkarni, D

D. Kulkarni, D. Schmidt, and S.-K. Tsui. Eigenvalues of tridiago- nal pseudo-Toeplitz matrices. Linear Algebra and its Applications , 297(1):63–80, 1999

1999
[18]

V. F. Lazutkin. KAM theory and semiclassical approximations to eigen- functions, volume 24. Springer Science & Business Media, 2012

2012
[19]

Sapiro and A

G. Sapiro and A. Tannenbaum. On affine plane curve evolution. Journal of functional analysis , 119(1):79–120, 1994. 28

1994

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [2]

Avila, J

A. Avila, J. De Simoi, and V. Kaloshin. An integrable deformation of an ellipse of small eccentricity is an ellipse. Annals of Mathematics , pages 527–558, 2016

work page 2016

[2] [1]

Albers and S

P. Albers and S. Tabachnikov. Introducing symplectic billiards. Ad- vances in mathematics (New York. 1965) , 333:822–867, 2018

work page 1965

[3] [3]

V. Bangert. Mather Sets for Twist Maps and Geodesics on Tori . Dy- namics reported. Wiley, 1988

work page 1988

[4] [4]

Baracco and O

L. Baracco and O. Bernardi. Totally integrable symplectic billiards are ellipses. Advances in Mathematics , 454:109873, 2024

work page 2024

[5] [5]

Area spectral rigidity for axially symmetric symplectic billiard tables

L. Baracco, O. Bernardi, and A. Nardi. Area spectral rigidity for axially symmetric and Radon domains. arXiv preprint arXiv:2410.12644, 2024

work page Pith review arXiv 2024

[6] [6]

Baracco, O

L. Baracco, O. Bernardi, and A. Nardi. Bialy-Mironov type rigidity for centrally symmetric symplectic billiards. Nonlinearity, 37(12):125025, nov 2024

work page 2024

[7] [7]

Bialy and A

M. Bialy and A. E. Mironov. The Birkhoff-Poritsky conjecture for centrally-symmetric billiard tables. Annals of Mathematics , 196(1):389 – 413, 2022

work page 2022

[8] [8]

De Simoi, V

J. De Simoi, V. Kaloshin, Q. Wei, and H. Hezari. Dynamical spec- tral rigidity among Z2-symmetric strictly convex domains close to a cir- cle (Appendix B coauthored with H. Hezari). Annals of Mathematics , 186(1):277–314, 2017

work page 2017