{"id":"08b12bed-6665-436a-b054-b08d688da45f","arxiv_id":"2504.21274","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For CM abelian varieties satisfying a simplicity hypothesis, Selmer ranks in p-th twists obey the symplectic or unitary distribution D^epsilon_q, giving unsolvability of most twisted Fermat curves along a fan structure.","lead":"Selmer ranks in p-th twists of CM abelian varieties are shown to follow symplectic or unitary probability distributions along a fan-structure stratification of characters. This yields a large lower bound on the proportion of twisted Fermat equations X^p+Y^p=delta that have no nontrivial solutions over number fields containing the p-th roots of unity.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The distribution theorem rests on Theorem 4.1, imported from the companion paper [33]; if the metabolic/unitary classification does not cover the omega-not-in-O case, local Kummer images need not be maximal isotropic and the Markov model collapses.","rationale":"The reader's weakest_assumption identifies the imported local quadratic-space theorem as the load-bearing premise, and I agree. My independent reading confirms that Theorem 4.1 is used at every critical juncture: local maximal isotropy (Prop. 4.2, 4.3), the counting of ramified isotropic lines (Prop. 4.2, 4.11), the rank-jump counts (Prop. 4.14), and the Markov transition probabilities (Theorem 5.7). Without it, Theorem 1.4 has no foundation. I found no internal contradiction stronger than this dependence. The stray factor '8 d(B,k)' in equation (7.1) appears to be a typo; the surrounding argument uses the correct expression and does not affect the conclusion. The paper is explicit that the densities are fan-structure densities rather than natural densities, so that is a stated limitation rather than a hidden flaw. The proof of Theorem 1.1 is sketched but plausible, relying on Northcott finiteness as in [9]; it is not the most fragile point. Therefore the verdict should remain CONDITIONAL: the main theorem is plausible and the framework is coherent, but the unresolved dependency on [33] must be verified before full acceptance. Since my concern is the same as the reader's, no verdict adjustment is needed.","tokens_in":34971,"tokens_out":9032,"duration_ms":94664,"concrete_test":"Independently re-derive Theorem 4.1 from [33, Proposition 4.4 and Theorem 5.4] under the exact hypotheses (CM),(H0)-(H3), and check whether the proof of [33, Theorem 5.4] requires a symmetric isogeny on A_chi (or on A) of degree prime-to-p whose Rosati involution restricts to complex conjugation on O. In the omega-not-in-O case, the paper says such an isogeny need not exist; if [33] implicitly assumes it, then Proposition 4.3's maximal-isotropy claim is unsupported for that case. If [33] instead covers omega-not-in-O via the Bloch-Kato pairing, verify explicitly that the reduction of that pairing to theta^lambda_p, as asserted in equation (4.6), is correct, since Proposition 4.3 depends entirely on this identification.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central distribution Theorem 1.4 depends on Theorem 4.1: for every place v, the local Tate pairing h_v makes H^1(F_v,T) a metabolic symplectic (CMS) or unitary (CMU) k-space. This is stated as a special case of [33, Proposition 4.4 and Theorem 5.4] and is not proved here. Every subsequent step inherits this: Proposition 4.2 uses it to assert H^1_ur(F_v,T) is maximal isotropic and |I_v^ram|=p; Proposition 4.3 uses it, together with a Bloch-Kato argument for omega not in O, to show each local Kummer image L_{A,v,chi} is maximal isotropic; Proposition 4.14 uses maximal isotropy to conclude the localized subspace V has dimension 1 and to count the +1/0/-1 transitions; Theorem 5.7 encodes these counts in the Markov operator M^epsilon. If Theorem 4.1 fails or requires extra hypotheses, the transition probabilities are wrong and the limiting distribution D_q^epsilon does not follow. The paper itself flags (Section 1.4) that for omega not in O there is generally no symmetric isogeny on A_chi of degree prime-to-p with Rosati involution restricting to complex conjugation on O, so the quadratic structure on H^1(F_v,A_chi[p]) is not available. Proposition 4.3's workaround via Bloch-Kato reduces the pairing to theta^lambda_p through the identifications in (4.6); that reduction is asserted but not fully expanded, and its validity is exactly what needs checking against [33]. If [33, Theorem 5.4] implicitly assumes a polarization of degree prime-to-p on the variety being twisted, then Theorem 4.1 does not transfer to A_chi in the omega-not-in-O case, and the proof of maximal isotropy fails there.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the p-Selmer ranks of p-th twists A_chi of a CM abelian variety A over a number field F. Under hypotheses (CM), (H0)-(H3) and either F=F_{d-1} != F_d with d>=2 or F_1=F_d, the main theorem (Theorem 1.4) asserts that, along a \"fan-structure\" stratification of the character group, the ranks of Sel_p(A_chi) modulo the Kummer image of torsion obey the symplectic or unitary distribution D_q^epsilon. The proof follows the Markov-model strategy of Klagsbrun-Mazur-Rubin: local Kummer images are maximal isotropic subspaces of quadratic cohomology spaces, and the rank transitions are governed by a Markov operator whose invariant distribution is D_q^epsilon. Applications include a lower bound for the proportion of delta in F^times/F^{times p} for which the twisted Fermat equation X^p+Y^p=delta has no nontrivial F-solution (Theorem 1.1), and results on rank growth in cyclic extensions (Theorems 1.6-1.8). The paper is structured and explicit about its dependence on the companion article [33] for the local quadratic-space structure.","tokens_in":35308,"tokens_out":6853,"duration_ms":74277,"significance":"If the companion result [33] is correct, this is a substantial contribution: it gives closed-form distributions for Selmer ranks in a large family of CM twists, with explicit numerical values and a concrete arithmetic application to twisted Fermat curves. The Markov architecture is clearly explained, no free parameters enter the distributions, and the paper is transparent about the main external input and about the omega-not-in-O obstruction. The applications to Fermat curves and cyclic extension rank growth are natural and nontrivial. The significance is, however, conditional on the cited quadratic-space structure theorem and on the sketched Diophantine passage in Theorem 1.1.","major_comments":[{"comment":"Theorem 4.1 is the load-bearing input that makes the local quadratic spaces metabolic symplectic or unitary, and it is only cited from [33, Proposition 4.4 and Theorem 5.4]; no proof is given in this manuscript. This structure is used immediately in Proposition 4.2 to identify unramified cohomology as maximal isotropic and to obtain |I_v^ram|=p, and that count propagates through Proposition 4.11, Proposition 4.14 and Theorem 5.7 into the Markov transition probabilities and hence into the limiting distribution in Theorem 7.9 and Corollary 7.10. In particular, the manuscript must clarify that Theorem 4.1 applies when omega is not in O: Section 1.4 notes that in this case A_chi generally does not admit a symmetric isogeny of degree prime to p with Rosati involution restricting to complex conjugation, and Proposition 4.3 uses a Bloch-Kato orthogonality argument for the maximal isotropy of L_{A,v,chi}. The counting of isotropic lines in Proposition 4.2(3), however, still depends directly on Theorem 4.1. If [33, Theorem 5.4] implicitly assumes the existence of such a polarization on the twisted variety, then the omega-not-in-O branch would not be covered by the cited theorem and the distribution theorem would not follow. Please add a proof, or a precise statement of the applicable hypotheses, or an explicit reduction of the omega-not-in-O case to the cited theorem.","section":"§1.1, proof of Theorem 1.1"},{"comment":"The passage from rank-zero Jacobians to nonsolubility of X^p+Y^p=delta is the arithmetic content of the Fermat application, but the manuscript only says \"by the same argument using Northcott's finiteness theorem as in [9]\". Since the setting here is an arbitrary number field containing mu_p, rather than the totally real fields of [9], a complete proof or a precise theorem quoted with its hypotheses is needed to justify that all but finitely many delta in Pi_1 give no nontrivial F-solution. This is a load-bearing step for Theorem 1.1 and Corollary 1.2.","section":"§1.1, proof of Theorem 1.1"}],"minor_comments":[{"comment":"The displayed equation has a stray factor 8d(B,k) on the left-hand side; as printed the equality is false. The subsequent argument uses the identity without this factor, so this appears to be a typo, but it should be corrected.","section":"§7, Eq. (7.1)"},{"comment":"The word \"otherwsie\" should be \"otherwise\".","section":"§5, Eq. (5.2)"},{"comment":"\"Rostai involution\" should be \"Rosati involution\".","section":"§1.4"},{"comment":"The word \"fan-structue\" should be \"fan-structure\", and in the Mathematics Subject Classification line \"Primiary\" should be \"Primary\".","section":"§1.2, after Definition 1.3"}],"recommendation":"major_revision","confidential_remarks":"The central distribution theorem depends on the author's own companion preprint [33], and the present manuscript does not contain enough information for a referee to verify Theorem 4.1. I recommend that the editor require a copy of [33] or an appendix with the proof of the relevant structure theorem before a final decision. The rest of the paper appears coherent and potentially publishable once that input is verified."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the quick read. The genuinely new result is a symplectic/unitary analogue of the Klagsbrun–Mazur–Rubin Markov model for p-Selmer ranks in p-th twists of CM abelian varieties, with explicit limiting distributions D_q^epsilon. The Fermat-curve application is a nice byproduct, giving a quantitative 'large unsolvability' statement along fan-structure densities. The paper is organized well and the Markov-chain transfer is done carefully, with effective Chebotarev and the fan-structure machinery handled in detail.\n\nThe soft spot is not novelty; it's the foundation. Theorem 4.1 — the claim that local Tate pairings make H^1(F_v,T) metabolic symplectic or unitary — is imported from the author's own companion [33] and not proved here. Everything downstream, including maximal isotropy of local Kummer images and the transition probabilities in Theorem 5.7, rests on it. The paper does flag the omega-not-in-O issue and offers a Bloch–Kato workaround in Proposition 4.3, which is more than the stress-test note credits; but the reduction in (4.6) is asserted rather than fully expanded. If [33] is correct and transfers to the twisted varieties, the argument holds. I couldn't verify that from this preprint.\n\nTwo smaller things. The Diophantine passage in Theorem 1.1 ('same argument using Northcott as in [9]') is a one-paragraph sketch; it's probably fine but deserves expansion. And (7.1) has a stray factor of 8d(B,k) on the left side that should be deleted; it's visibly a typo, not a mathematical issue. Also note the densities are fan-structure densities, not natural densities; the paper says so, but casual readers will miss that.\n\nOverall: this is a serious paper that probably has the shape of the truth. The right referee will need the companion [33] in hand and should check Theorem 4.1 and the omega-not-in-O transfer carefully. If that passes, this is a publishable result in arithmetic statistics. Definitely send it to peer review; it deserves a real referee, not a desk reject.","headline":"First symplectic/unitary Selmer-rank distribution for CM twists, but the load-bearing local quadratic-space structure is imported from the companion paper; a serious referee is needed.","tokens_in":35874,"tokens_out":2972,"would_cite":true,"duration_ms":27322,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":null,"created_at":"2026-08-16T05:09:24.982386+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}