{"id":"33c166e2-ed67-4845-bc9a-909f6a2eb803","arxiv_id":"2507.08446","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"At high energies, every real-analytic convex Kepler billiard is non-integrable for all but at most one position of the center unless the table is an ellipse with the center at a focus.","lead":"Mathematicians prove that in a high-energy Kepler billiard, a particle moving under a central attraction inside a convex analytic table is generically chaotic: only an elliptic table with the attraction at a focus can be integrable, apart from possibly one exceptional center location. The proof combines symbolic dynamics from shadowed broken-line trajectories with a rigidity theorem classifying 'focal points of the second kind'.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"High-energy Jacobi-length asymptotics are the load-bearing link: uniform remainder and containment in Lemmas 2.2/2.3 must hold for every endpoint pair used in the shadowing chains, or the degree-persistence argument in Lemmas 3.9/3.10 collapses.","rationale":"The reader's CONDITIONAL verdict seems right. I read the full construction: non-integrability hinges on producing a Bernoulli subsystem by degree persistence, and that persistence depends entirely on the high-energy asymptotics of the Jacobi length. If those estimates are correct, the periodic-orbit construction and the diagonalization to a full shift are standard and plausible, though the diagonalization is only sketched by reference to [3]. The rigidity section is comparatively self-contained; Lemma 4.9's unique continuation is terse but has a clear mechanism, and the computations in Lemmas 4.5 and 4.8 are checkable. I did not find an explicit contradiction or counterexample in the text. Therefore the appropriate action is to keep the CONDITIONAL verdict and ask the authors to expand the asymptotic estimates and their use in Lemmas 3.9/3.10, rather than to reject. The paper also has independent support: Panov's integrability for elliptic focus centers, the explicit string construction of non-elliptic focal domains, and the numerical simulations. My proposed check is deliberately targeted at the one step where the text is too compressed to verify from the page: the uniform remainder control for the exact endpoint pairs used in the shadowing chains.","tokens_in":25927,"tokens_out":36614,"duration_ms":438088,"concrete_test":"Re-derive Lemma 2.3 for the exact antipodal pair (ξP,ξQ) used in Section 3.2.2, computing first and second derivatives of L(za/c)-sqrt(h)f_a/c with respect to endpoints and checking uniform boundedness on the full set \\tilde K_δ, including across det(p0,p1)=0. Then, for the periodic word s=(m,M,m,M), compute the topological degree of ∇\\tilde W_s on ∂U_s for h=10^3 and h=10^6, using the explicit remainders; if the degree differs from the product of the indices of the limiting critical point, or if any zero of the linear homotopy to sqrt(h)∇(2|γ|+Ψ_a) appears on ∂U_s, the shadowing construction in Lemma 3.10 fails. This directly tests the chain that produces symbolic dynamics.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 1.4 rests on Lemmas 3.9 and 3.10: the true generating function for a periodic word is written as sqrt(h) times a perimeter functional plus a remainder, and critical points of the limiting functional (with nonzero index) are shown to persist. This is valid only if Lemmas 2.2 and 2.3 give uniform remainder control for every endpoint pair used in the concatenations, and if the direct, indirect, and anticlockwise Keplerian arcs remain inside the domain. Lemma 2.2's C^2 uniformity is asserted with a short proof sketch and a reference to [3]; Lemma 2.3, proved in Appendix A, establishes first-derivative convergence but leaves the remainder at O(1), not o(1), and does not exhibit a uniform C^2 bound across the switching locus det(p0,p1)=0 where f_a and f_c change branches. The topological-degree argument needs |∇W_s - sqrt(h)∇Ψ_s| to be uniformly smaller than sqrt(h) times the minimal modulus of ∇Ψ_s on the boundary of the isolating neighborhoods; if the O(1) remainder from Lemma 2.3 or the constants from shrinking neighborhoods interact badly, zeros could appear on the boundary or the degree could change. The same asymptotics are the only justification that the periodic critical points correspond to genuine billiard trajectories rather than arcs that exit Omega. The focal-point counting, ellipse rigidity, and Proposition 4.10 are largely independent of these estimates, so this is the single most load-bearing unresolved step in the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies analytic integrability of Kepler billiards in strictly convex real-analytic planar domains at high energies. The main result (Theorem 1.1) states that, apart from at most one position of the attraction center (two for an ellipse, namely its foci), the Kepler billiard is not analytically integrable; the mechanism is the construction, for all centers of the first kind or non-focal centers, of a subsystem of the first-return map that is semi-conjugate to a Bernoulli shift (Theorem 3.4 and Corollary 3.11), yielding positive topological entropy and no analytic first integral. The shadowing is based on high-energy asymptotics of the Jacobi length of direct/indirect/clockwise/anticlockwise Keplerian arcs, with limiting perimeter functionals Psi, Psi_a, Psi_c; critical points are propagated by topological degree. The second half of the paper classifies focal points of the second kind, showing that there are at most two, with equality only for ellipses at their foci, and constructs an infinite-dimensional family of non-elliptic domains with one focal point, supported by numerics.","tokens_in":26198,"tokens_out":17860,"duration_ms":198008,"significance":"If the central arguments are completed, this is a substantial contribution to the Keplerian analogue of the Birkhoff-Poritsky conjecture. It introduces a clean three-case shadowing strategy, uses topological-degree persistence in an essential way, and separates a rigidity problem about focal points that is interesting in its own right. The paper is honest about the dependence on [3] for the first-kind case and about the limitation to high energies; the exceptional focal case is explicitly identified rather than hidden. The main weaknesses are that two load-bearing analytic steps are compressed (uniform asymptotic remainder estimates and the passage from an invariant-graph identity to ellipsoidal rigidity) and that the claimed construction of non-elliptic focal domains is not actually proved. The result is, in my assessment, likely correct in broad outline, but the manuscript needs a substantial revision before it can be accepted.","major_comments":[{"comment":"The proof of Lemma 2.3 does not fully establish the uniform first-order asymptotic used in Lemma 3.10. The statement requires g_a and g_c to be C^1 and uniformly bounded in p0,p1, and the derivation in Appendix A obtains the leading term of \\nabla_{p1} L(z_a) from an expansion y = y0 + sqrt(h') g with y0 piecewise defined at <w0,w1>=0. The passage \"by regularity of the involved functions\" from this expansion to uniform C^1 bounds for the remainder is not justified near that switching locus, where the chosen square-root branch w1 = +/- sqrt(p1) changes; and the uniform control for h' -> 0 is only sketched. Since the degree-persistence argument in Lemma 3.10 requires the remainder in (15) to be uniformly O(1) (i.e., o(sqrt(h))) on the whole product of the neighborhoods I_m, I_M, I_P, I_Q and for every periodic word s, this is a load-bearing point. Please provide a complete proof or a precise reference for the uniform C^1 estimate across det(p0,p1)=0.","section":"Section 2, Lemma 2.3 and Appendix A"},{"comment":"The construction of non-elliptic domains with a focal point of the second kind is not proved. After defining gamma as a level set of f(x)=|x-c|+d(x,tau), the proof asserts that \"clearly\" phi_c is constant because all chords orthogonal to T have the same length; this does not follow, since phi_c is defined by two consecutive reflections on gamma and involves the three distances |gamma(xi)-c|, |gamma(xi)-gamma(xi')| and |gamma(xi')-c|, not the chord lengths of tau. The intermediate claim that gamma has a unique periodic trajectory through c is also insufficient. This gap affects the advertised infinite-dimensional family and the sharpness discussion of Theorem 1.1, so it should either be fixed with a real proof or the corresponding claims should be weakened.","section":"Section 4.1, Proposition 4.10"},{"comment":"The rigidity step at the end of Lemma 4.9 and in the proof of Theorem 4.1 is too compressed. From the identity delta_+(S^1) = delta^c_-(S^1) the manuscript concludes that partial Omega is an ellipse with foci 0 and c, saying that the family of invariant lines consists of the two pencils through the foci and draws triangles of equal perimeter. This is a nontrivial characterization and no proof is supplied. Because Theorem 4.1 is what limits the exceptional center positions in Theorem 1.1, this step is load-bearing; it should be stated as a lemma with a complete argument or an explicit reference.","section":"Section 4, Lemma 4.9 and proof of Theorem 4.1"}],"minor_comments":[{"comment":"The notation \"focal points of the second kind\" is used in the abstract before its definition in Definitions 3.1 and 3.3; consider adding a forward reference.","section":"Abstract and Section 3"},{"comment":"The derivative formula (15) is stated only for k=1,2; the analogous formula for the third coordinate u_{i,3} is needed for the critical-point equation and should be included.","section":"Section 3.2.2, Lemma 3.10"},{"comment":"There are several typos (\"critival\" in Appendix B, \"anergy\" in Corollary 1.5, \"world\" in Section 3.3), and the reference to Figure 15 in Section 4.1 does not mention the numerical integration method; these should be corrected.","section":"Throughout"},{"comment":"The function B in the proof is introduced without specifying its domain and regularity; a brief justification (or a local version via the implicit function theorem) would improve readability.","section":"Section 3.1, Lemma 3.5"}],"recommendation":"major_revision","confidential_remarks":"To the editor: the main theorem is likely to be correct, but two proofs are too compressed for a journal of this level: the uniform asymptotic estimates behind Lemma 3.10 and the rigidity conclusion of Lemma 4.9/Theorem 4.1. The reader's stress-test concern about O(1) versus o(1) in Lemma 2.3 is somewhat overstated because topological-degree persistence only needs a uniformly bounded remainder; nevertheless the missing uniform C^1 control across det(p0,p1)=0 is a genuine presentation gap. Additionally, Proposition 4.10 is advertised as a theorem but appears unproved. I recommend major revision and, if the authors can supply the missing estimates, the result would be a strong contribution. There is substantial overlap with [3] for the first-kind case, but the non-focal construction and the focal-point rigidity are new; the overlap is properly acknowledged."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper is worth taking seriously. The main result—high-energy analytic non-integrability for non-elliptic Kepler billiards except possibly one center position—is a real step beyond the earlier first-kind construction in [3] and beyond Panov's integrability example. The three symbolic-dynamics constructions, especially the new ones for second-kind and non-focal centers via punctured Birkhoff triangles, are clever and, as far as I can tell, sound. The focal-point rigidity section is genuinely elegant: the Hessian computation, the chord argument, and the use of invariant graphs to force ellipticity are convincing and largely independent of the high-energy asymptotics. The construction of an infinite-dimensional family of non-elliptic tables with a focal point of the second kind is a nice bonus and makes Theorem 1.1's exception sharp.\n\nThe stress-test note names the right soft spot. The whole non-integrability theorem rests on Lemmas 2.2 and 2.3 giving uniform asymptotic control of the Jacobi length for every endpoint pair used in the concatenations. Lemma 2.2 is mostly a reference to [3] with a sketch. Lemma 2.3, proved in the appendix, gives C^1 convergence but leaves the remainder at O(1), not o(1), and the uniformity across the switching locus det(p0,p1)=0 is not fully written out. The degree-persistence argument in Lemmas 3.9 and 3.10 needs the remainder to be uniformly smaller than sqrt(h) times the minimal modulus of the limiting gradient on the isolating neighborhoods; that quantitative condition is not checked. Containment inside the domain is also asserted by uniform convergence, but for direct arcs it is stated only when |p0-p1| >= epsilon, and the concatenations in Lemma 3.10 use short transfer arcs that need the same control.\n\nI do not see an actual contradiction, and the authors may well have the estimates in hand. But this is the load-bearing step, and the paper currently treats it too casually. Lemma 4.9's unique-continuation step is quick, though plausible. These are addressable gaps rather than fatal flaws.\n\nMy take: the central claim is likely correct, the rigidity half is solid, and the paper deserves a serious referee. I would send it out and ask the referee to push hard on Lemmas 2.2 and 2.3, and to make the isolating neighborhoods in Lemmas 3.9 and 3.10 explicit. If that section tightens up, this becomes a strong paper.","headline":"Genuinely new and likely correct, but the high-energy asymptotic estimates that carry the non-integrability theorem are compressed and need referee scrutiny.","tokens_in":26778,"tokens_out":2223,"would_cite":true,"duration_ms":27504,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37C83","34C28","37J51","37J46"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper establishes that high-energy analytic Kepler billiards are integrable only on elliptic tables with the attraction center at a focus.","keywords":["Kepler billiards","Birkhoff-Poritsky conjecture","analytic integrability","symbolic dynamics","topological entropy","focal points of the second kind","high-energy regime","shadowing method"],"falsifier":"Find a real-analytic strictly convex non-elliptic domain admitting two distinct focal points of the second kind; the polynomial $P(t)$ derived in Lemma 4.5 constrains such points to one orthogonal chord and forces the domain to be an ellipse, so exhibiting such a pair would refute Theorem 4.1 and the rigidity half of Theorem 1.1. Equivalently, exhibit an analytic first integral for a high-energy Kepler billiard in a non-elliptic analytic table with a non-focal center.","tokens_in":25694,"feed_emoji":"🪐","tokens_out":4815,"duration_ms":54491,"temperature":0.7,"pith_summary":"This paper asks whether a mechanical billiard with an inverse-distance gravitational force can be analytically integrable. It argues that, at high energies, the answer is essentially no for strictly convex analytic tables, unless the table is an ellipse and the attracting center sits at a focus. The proof constructs symbolic dynamics—a coding of trajectories by a full shift on two symbols—by shadowing nearly degenerate triangular billiard paths through the center. If correct, the results give a partial affirmative answer to the Keplerian analogue of the Birkhoff-Poritsky conjecture, up to one exceptional center position in non-elliptic domains.","feed_headline":"High-energy Kepler billiards are almost never integrable","feed_subtitle":"Non-elliptic analytic tables force chaotic symbolic dynamics and positive entropy; only elliptic focus tables can stay integrable.","key_machinery":"The argument is carried by the high-energy asymptotics of the Jacobi length: a direct Keplerian arc between two boundary points has length $\\sqrt{h}\\lvert p_0-p_1\\rvert + O(\\mu/\\sqrt{h})$, and an indirect arc through the center has length $\\sqrt{h}(\\lvert p_0\\rvert + \\lvert p_1\\rvert) + O(\\mu/\\sqrt{h})$ plus a logarithmic term, with variants for antipodal pairs. These expansions turn the multi-valued generating function of the Kepler billiard into $\\sqrt{h}$ times a perimeter function $\\Psi(\\xi,\\eta)=\\lvert\\gamma(\\xi)\\rvert+\\lvert\\gamma(\\eta)\\rvert+\\lvert\\gamma(\\eta)-\\gamma(\\xi)\\rvert$, whose critical points are punctured Birkhoff triangles. The persistence of these critical points under the high-energy perturbation is controlled by topological degree, and shadowing periodic words in the resulting symbolic system is extended to all bi-infinite sequences by a diagonal argument.","core_discovery":"The central discovery is a rigidity theorem: for a strictly convex bounded planar domain with real-analytic boundary, at sufficiently high energy the Kepler billiard map cannot be analytically integrable unless the boundary is an ellipse and the center of attraction is one of its foci. For non-elliptic domains, the theorem allows at most one exceptional center position, and for ellipses the only exceptions are the two foci. This is proved by showing that, for every non-exceptional center, the high-energy first-return map contains a subsystem semi-conjugate to a Bernoulli shift, which implies positive topological entropy and the absence of analytic first integrals. A companion geometric rigidity result says that real-analytic non-elliptic domains admit at most one focal point of the second kind, while ellipses are the only domains admitting two, namely their classical foci.","pith_inferences":["The same triangle-shadowing mechanism may extend to finite energies if quantitative bounds on the Jacobi-length remainders can be obtained; the paper only claims the high-energy regime.","The numerical simulations reported for a focal point of the second kind suggest that chaotic behavior persists even where the proof does not apply, so the exceptional one-center case may be a limitation of the method rather than true integrability.","The geometric part of the result—at most two focal points of the second kind, with two forcing an ellipse—stands independently of Keplerian dynamics and can be tested directly on the Birkhoff billiard map.","A natural testable extension is to push the shadowing construction to the zero-energy case, which the paper notes is conjugate to an ordinary Birkhoff billiard on the double cover of the domain."],"forward_implications":["If the central claim is correct, any strictly convex real-analytic non-elliptic table at high energy has positive topological entropy for the Kepler billiard map for all but possibly one center position.","At those centers, the first-return map admits no analytic first integral, so the system is not analytically integrable in the sense used here.","The only analytically integrable high-energy Kepler billiards among analytic convex tables are elliptic tables with the center at a focus, matching Panov's known integrable examples.","The rigidity theorem implies that a table with two focal points of the second kind must be an ellipse, giving a sharp geometric obstruction to the symbolic-dynamics construction.","The paper exhibits an infinite-dimensional family of non-elliptic analytic tables with a focal point of the second kind, so the exceptional case in Theorem 1.1 is real and not vacuous."],"supporting_citations":[{"why":"Provides the positive half of the rigidity picture: elliptic Kepler billiards are integrable at all energies when the center is at a focus.","marker":"[29]"},{"why":"Supplies the first symbolic-dynamics construction for centers of the first kind and the asymptotic length and gradient estimates used throughout.","marker":"[3]"},{"why":"Introduces the degenerate-billiard shadowing method in celestial mechanics that the paper adapts to Kepler billiards.","marker":"[8]"},{"why":"Gives the periodic and chaotic shadowing framework for second-species trajectories that underpins the construction of periodic Kepler billiard trajectories.","marker":"[9]"},{"why":"Establishes the classical existence of direct and indirect Keplerian arcs between two points, the starting point for the multi-valued generating functions.","marker":"[19]"},{"why":"Provides the topological-degree product and homotopy properties used to prove persistence of critical points in the shadowing lemmas.","marker":"[26]"},{"why":"Supplies the classical implication that positive topological entropy of a subsystem precludes analytic integrability.","marker":"[17]"},{"why":"Supplies the Katok-style argument adapted to show absence of analytic first integrals from the presence of a chaotic subsystem.","marker":"[24]"},{"why":"Provides the constant-width domain and string construction used to build the infinite-dimensional family of non-elliptic tables with a focal point of the second kind.","marker":"[23]"}],"fun_headline_variants":["Kepler billiards: integrable only if elliptic with focal center","Elliptic focus is only integrable Kepler billiard at high energy","Analytic Kepler billiards: integrability forces elliptic focus","High-energy Kepler billiards: chaos unless elliptic focus","Only elliptic focus domains keep Kepler billiards integrable"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper relies on high-energy asymptotic expansions of the Jacobi length holding uniformly with controlled remainders and with the Keplerian arcs staying inside the strictly convex domain, so that the true generating function can be replaced by $\\sqrt{h}$ times a perimeter function without losing critical points.","fun_headline_variants_meta":{"raw":{"variants":["Kepler billiards: integrable only if elliptic with focal center","Elliptic focus is only integrable Kepler billiard at high energy","Analytic Kepler billiards: integrability forces elliptic focus","High-energy Kepler billiards: chaos unless elliptic focus","Only elliptic focus domains keep Kepler billiards integrable"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000725,"raw_usage":{"total_tokens":3252,"prompt_tokens":951,"completion_tokens":2301,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":567,"completion_tokens_details":{"reasoning_tokens":2214}},"tokens_in":567,"tokens_out":2301,"duration_ms":16402,"temperature":1.0,"reasoning_tokens":2214,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:21:23.126913+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a real-analytic strictly convex non-elliptic domain admitting two distinct focal points of the second kind; the polynomial $P(t)$ derived in Lemma 4.5 constrains such points to one orthogonal chord and forces the domain to be an ellipse, so exhibiting such a pair would refute Theorem 4.1 and the rigidity half of Theorem 1.1. Equivalently, exhibit an analytic first integral for a high-energy Kepler billiard in a non-elliptic analytic table with a non-focal center.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the positive half of the rigidity picture: elliptic Kepler billiards are integrable at all energies when the center is at a focus."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the first symbolic-dynamics construction for centers of the first kind and the asymptotic length and gradient estimates used throughout."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the degenerate-billiard shadowing method in celestial mechanics that the paper adapts to Kepler billiards."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the periodic and chaotic shadowing framework for second-species trajectories that underpins the construction of periodic Kepler billiard trajectories."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the classical existence of direct and indirect Keplerian arcs between two points, the starting point for the multi-valued generating functions."},{"cited_title":"Krawcewicz and J","cited_arxiv_id":null,"evidence_quote":"Provides the topological-degree product and homotopy properties used to prove persistence of critical points in the shadowing lemmas."},{"cited_title":"Hasselblatt and A","cited_arxiv_id":null,"evidence_quote":"Supplies the classical implication that positive topological entropy of a subsystem precludes analytic integrability."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Katok-style argument adapted to show absence of analytic first integrals from the presence of a chaotic subsystem."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the constant-width domain and string construction used to build the infinite-dimensional family of non-elliptic tables with a focal point of the second kind."}],"review_version":1}