{"id":"5a190b77-1228-459a-aae6-e9f84474eada","arxiv_id":"1908.00455","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A cohomological identity on strata of rational Hurwitz spaces yields a new differential recursion that determines all genus 0 double Hurwitz numbers.","lead":"This paper derives a new recursion for genus 0 double Hurwitz numbers from cohomology classes of multisingularity loci in Hurwitz spaces. The recursion computes all such numbers from explicit initial series and is expected to generalize to settings where closed formulas are unknown.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The recursion's geometric coefficient rests on Theorem 3.3(d), whose proof is delegated to Lemma 2.1 of [10] and only sketched in §3.2; an error in the asserted multiplicity σ1···σℓ would propagate into every computed Hurwitz number.","rationale":"The reader's weakest_assumption identifies exactly the same point: Theorem 3.3(d), the vanishing-order computation. This is the correct load-bearing concern because the relation (3) and the recursion of Theorem 1.7 depend on the coefficient σ1···σℓ; if that coefficient is wrong, the entire computation of genus 0 double Hurwitz numbers shifts. The manuscript itself signals the risk by saying the proof is 'rather concise' and refers to Lemma 2.1 of the authors' prior paper [10]. The local calculation is not merely routine: it involves a root-of-unity action, branch counting, and the assertion that the λ1-th derivative equals c^K times a nonzero factor on each branch, so there is room for a subtle error. I do not see evidence that the theorem is false; the examples are consistent and the overall structure is coherent, so the concern is not a rejection but a request for verification. The reader's CONDITIONAL verdict already captures this, so no change is needed. The structural issues (duplicated proof sections, sketchy Section 6) are secondary and do not affect the central claim as directly.","tokens_in":22016,"tokens_out":13161,"duration_ms":144176,"concrete_test":"Test Theorem 3.3(d) directly for the first case where the branch count is nontrivial: take J = {1,2}, λJ = (2,2), σ = (2,2), so K = 2 and σ1σ2/K = 2. In the normal form (4), set r1 = r2 = 1, choose generic ūi, āij, and βj, and compute the order of vanishing of f^(2)(x1) in the smoothing parameter c separately on each of the two root-of-unity branches. If every branch has order exactly 2, the total coefficient in (3) is 4 and this concern is resolved. As an independent cross-check, compute h(3,3) from the reduced recursion of Theorem 4.1 and compare it with the Schur/Frobenius formula e^H = Σλ e^{w(λ)β} sλ(p)sλ(q) given in §1.1; any mismatch would indicate a wrong multiplicity in the recursion.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The recursion in Theorem 1.7 is built on the cohomological identity (3). Its only geometric input is Corollary 3.4, and the coefficient σ1···σℓ in (3) comes entirely from Theorem 3.3(d). The proof of (d) in §3.2 is explicitly a sketch: 'The proof repeats that of Lemma 2.1 in [10] with a slight modification due to supplementary marked points. We give here a rather concise presentation, referring to [10] for more details.' The calculation uses the local normal form (4) and asserts that, under ui = ζi c^{r_i} ūi and aij = ζi^j c^{j r_i} āij, the λ1-th derivative equals c^K g^(λ1)(x1) on each of the σ1···σℓ/K branches, giving total order σ1···σℓ. This multiplicity is load-bearing: every coefficient produced by the recursion, and hence every genus 0 double Hurwitz number derived from it, is linear in these σ-products. A wrong branch count, a missing factor, or an extra vanishing of g^(λ1)(x1) on a branch would shift all subsequent hλ. The published examples do exercise some boundary terms, but they do not isolate the delicate case where the branch count σ1···σℓ/K is nontrivial, so the central claim remains dependent on an unverified local computation.","agreement_with_reader":"agree"},"referee_report":null,"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Let me give you the quick take on Kazarian–Lando–Zvonkine. The paper has a genuinely new idea: it derives a recursion for genus-0 double Hurwitz numbers from a cohomological identity on multisingularity strata in Hurwitz spaces, rather than from the usual cut-and-join or Frobenius machinery. The central recursion (Theorem 1.7) and the identity (3) appear to be new, and the derivation is coherent. The paper also proves string and dilaton equations for the descendant potential, and the worked examples (h_(2), h_(3), h_(2,2), etc.) are consistent with known numbers. That is a solid piece of work.\n\nThe soft spots are real but manageable. The one that matters is Theorem 3.3(d), the vanishing order σ1···σℓ of the section f^(λ1)(x1) along boundary strata. That multiplicity is load-bearing: every coefficient produced by the recursion is linear in these products. But the proof is explicitly a sketch and defers to Lemma 2.1 of the authors' earlier paper [10]. The local parametrization in §3.2 is plausible, but the published examples do not isolate the case where the branch count σ1···σℓ/K is nontrivial, so the recursion is not independently checked in that delicate regime. I don't think it's wrong, but a referee should insist on a self-contained proof or a precise cite to [10] with the necessary details. There's also a duplicated section header (\"Proof of the recursion\" appears as 3.3 and 3.4), which is a drafting slip, not a mathematical issue. Theorem 6.1 about the KP hierarchy is more of a remark; it's restricted to smooth fibers and the paper itself admits the pushforward computation is incomplete. I'd treat that as a side claim, not a load-bearing one.\n\nThe citation pattern is fine: the reliance on [10] and [3] is legitimate given the subject. The paper is honest about the higher-genus obstacle. On the whole, the central argument holds up, conditional on the deferred lemma. I'd send it to a serious referee. If you work on Hurwitz numbers, it's worth a close read; I'd probably cite it for the approach.","headline":"New geometric recursion for genus-0 double Hurwitz numbers, credible and worth refereeing, but the key multiplicity lemma is delegated to a prior paper and needs scrutiny.","tokens_in":22862,"tokens_out":2965,"would_cite":true,"duration_ms":28020,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":null,"created_at":"2026-08-14T15:55:34.451266+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}