{"id":"91628c84-3452-454d-bbae-3b4455eb59e3","arxiv_id":"2501.08849","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Near any ellipse, rational integrability of the symplectic billiard map forces the domain to be an ellipse.","lead":"This paper proves that a convex domain whose symplectic billiard map is rationally integrable and whose boundary is close to an ellipse must itself be an ellipse. It is the symplectic-billiard analogue of a known rigidity theorem for ordinary billiards.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Arithmetic error in Lemma 3.6: with α=1/31 the claimed identity 4−63α = 2+α is false, so the Fourier-tail estimate is sublinear and the contraction argument in §3.4 is unsupported as written.","rationale":"The reader flagged the affine-invariance transfer from Section 3.3 as the weakest assumption. That concern is reasonable but less decisive than the arithmetic inconsistency I found. The central claim of the paper stands or falls on Lemma 3.6's contraction estimate, which feeds directly into the minimal-ellipse argument in §3.4. The displayed identity that calibrates the Fourier exponent is simply false for the stated α = 1/31. Recomputing the sums shows the tail is only sublinearly small in the C^1 norm, so the interpolation step cannot produce the superlinear bound (3.27) as written. I do not see an immediate way around this within the text; however, the structure strongly suggests a fixable parameter error (α = 1/32 is the natural value satisfying the intended equality), and the method of [2] is robust enough that a corrected exponent should restore the proof. Therefore I recommend a conditional acceptance rather than outright rejection: the result is plausible and likely correct, but the manuscript as written contains a load-bearing gap in the proof of Lemma 3.6.","tokens_in":18322,"tokens_out":23004,"duration_ms":228348,"concrete_test":"Recompute (3.32)–(3.34) literally with α = 1/31: verify that 4 − 63α = 61/31 and 2 + α = 63/31, and hence the L2 bound for n⊥ has exponent 61/62, not 63/62. Then substitute δ = 61/62 into the interpolation step with ε optimized in (3.35) and check whether the resulting C^1 bound for n⊥ has exponent > 1. If it does not, test the proposed correction α = 1/32 with the j = 1 version of (3.35); if that gives a superlinear exponent, the argument is repairable, otherwise Lemma 3.6 fails.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The proof of Lemma 3.6 depends on a superlinear bound for the high-frequency tail n⊥, namely ||n⊥||_{L2} ≤ C ||n||^{63/62}, so that Sobolev interpolation can produce a C^1 bound of order ||n||^{1+δ} and hence the contraction (3.36). This exponent is obtained by asserting, after (3.32)–(3.33), that “4 − 63α = 2 + α = 63/31” for the chosen α = 1/31. But 4 − 63·(1/31) = 61/31, not 63/31; the equality 4 − 63α = 2 + α would require α = 1/32. Consequently, with α = 1/31 the low-frequency contribution is actually ≤ C ||n||^{61/31}, which dominates the tail contribution ≤ C ||n||^{63/31}, so (3.34) should read ||n⊥||_{L2} ≤ C ||n||^{61/62}. Since 61/62 < 1, the subsequent interpolation in (3.35) cannot yield a C^1 bound with exponent > 1; the inequality (3.27) and the final contradiction in §3.4 rely on exactly such a superlinear exponent. This is an internal inconsistency in a load-bearing step, not merely a cosmetic typo: as written, the proof of Lemma 3.6 does not establish the contraction that the closing argument needs. The theorem may well be salvageable by correcting α to 1/32 and redoing the interpolation, but the present text lacks a valid derivation.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":18616,"tokens_out":12475,"duration_ms":102918,"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":[{"comment":"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":"Section 3.3, Eqs. (3.32)–(3.34)"},{"comment":"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.","section":"Section 3.3, Eq. (3.35)"}],"minor_comments":[{"comment":"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.","section":"Section 3.3, first paragraph"},{"comment":"There are several typos, including 'percise' in the introduction and 'deonminator' in the proof of Lemma 3.2; these should be corrected.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The arithmetic error in Lemma 3.6 is serious but appears to be repairable: changing α from 1/31 to 1/32 makes the two exponents coincide and yields a superlinear L2 bound, after which the interpolation step can likely be fixed. I therefore recommend major revision rather than rejection. The author should be asked to rederive the exponents in Lemma 3.6 carefully and to correct the sign in Eq. (3.35)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","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.\n\nThe 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.\n\nThat 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.\n\nThe 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.\n\nWho 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.","headline":"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.","tokens_in":19197,"tokens_out":3969,"would_cite":true,"duration_ms":35609,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37C83","37E40","37J46","53A15"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that any domain sufficiently close to an ellipse whose symplectic billiard dynamics is rationally integrable must itself be an ellipse.","keywords":["symplectic billiards","local rigidity","rational integrability","ellipses","affine arc-length parametrization","twist maps","Fourier coefficients","periodic orbits"],"falsifier":"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.","tokens_in":18082,"feed_emoji":"🎱","tokens_out":8085,"duration_ms":71825,"temperature":0.7,"pith_summary":"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.","feed_headline":"Near an ellipse, integrable symplectic billiards are still ellipses","feed_subtitle":"A new proof shows any rationally integrable perturbation of an ellipse must be the ellipse itself.","key_machinery":"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.","core_discovery":"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.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"Introduces symplectic billiards, establishes the twist-map and generating-function formulation, and proves the suitability of affine arc-length parametrization; the system under study.","marker":"[1]"},{"why":"Supplies the rigidity strategy and the Lemmas 16-19 for radial deformations of a circle that the paper transfers to ellipses; the main method being mimicked.","marker":"[2]"},{"why":"Provides the theory of twist maps used to assert that orbits on invariant curves are locally maximizing, a background needed for the action-constant argument.","marker":"[11]"},{"why":"Gives the explicit implicit-function-theorem domain-size estimates used to quantify the neighborhood where the parametrization of periodic orbits is controlled.","marker":"[14]"},{"why":"Provides the eigenvalues of the tridiagonal matrix appearing as the Jacobian of the periodicity map, used to control the inverse derivative.","marker":"[17]"},{"why":"Defines the affine normal and affine arc-length parametrization used to write the deformation of the ellipse.","marker":"[19]"},{"why":"Supplies the Sobolev interpolation inequalities used in Lemma 3.6 to upgrade L^2 Fourier bounds to C^1 bounds on the non-elliptic part.","marker":"[10]"}],"fun_headline_variants":["Near an ellipse, integrable symplectic billiards must be elliptic","Local rigidity: integrable symplectic billiards near an ellipse are ellipses","Rational integrability near an ellipse forces symplectic billiards to stay elliptic","Symplectic billiard ellipses: local rigidity from rational integrability","Close to an ellipse, rational integrability pins symplectic billiards to elliptic"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["Near an ellipse, integrable symplectic billiards must be elliptic","Local rigidity: integrable symplectic billiards near an ellipse are ellipses","Rational integrability near an ellipse forces symplectic billiards to stay elliptic","Symplectic billiard ellipses: local rigidity from rational integrability","Close to an ellipse, rational integrability pins symplectic billiards to elliptic"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001723,"raw_usage":{"total_tokens":6724,"prompt_tokens":768,"completion_tokens":5956,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":384,"completion_tokens_details":{"reasoning_tokens":5851}},"tokens_in":384,"tokens_out":5956,"duration_ms":37420,"temperature":1.0,"reasoning_tokens":5851,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:17:06.874223+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[{"cited_title":"Albers and S","cited_arxiv_id":null,"evidence_quote":"Introduces symplectic billiards, establishes the twist-map and generating-function formulation, and proves the suitability of affine arc-length parametrization; the system under study."},{"cited_title":"Avila, J","cited_arxiv_id":null,"evidence_quote":"Supplies the rigidity strategy and the Lemmas 16-19 for radial deformations of a circle that the paper transfers to ellipses; the main method being mimicked."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the theory of twist maps used to assert that orbits on invariant curves are locally maximizing, a background needed for the action-constant argument."},{"cited_title":"Jindal, D","cited_arxiv_id":null,"evidence_quote":"Gives the explicit implicit-function-theorem domain-size estimates used to quantify the neighborhood where the parametrization of periodic orbits is controlled."},{"cited_title":"Kulkarni, D","cited_arxiv_id":null,"evidence_quote":"Provides the eigenvalues of the tridiagonal matrix appearing as the Jacobian of the periodicity map, used to control the inverse derivative."},{"cited_title":"Sapiro and A","cited_arxiv_id":null,"evidence_quote":"Defines the affine normal and affine arc-length parametrization used to write the deformation of the ellipse."},{"cited_title":"Gilbarg and N","cited_arxiv_id":null,"evidence_quote":"Supplies the Sobolev interpolation inequalities used in Lemma 3.6 to upgrade L^2 Fourier bounds to C^1 bounds on the non-elliptic part."}],"review_version":1}