{"id":"93a0cf62-1ed2-481d-82ea-23c4c3e5526c","arxiv_id":"2502.00259","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An explicit computation of the stringy Chow ring for quotients by diagonalizable groups and for weighted blowups, with finite generation criteria.","lead":"The authors compute the stringy Chow ring, a ring of intersection classes associated to singular spaces with group actions, for quotient stacks and for weighted blowups. This gives explicit formulas for the ring structure and a condition under which the ring is finitely generated, extending earlier results that were only known over the complex numbers.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: Proposition 2.4.1 and the product rules built on it survive the stress-test; the only genuine limitation is the explicitly-stated regularity of the Rees algebra.","rationale":"The reader's acceptance at moderate confidence is consistent with my independent pass. The strongest claim, Proposition 2.4.1, is indeed the engine of the paper, and I traced it through Lemma 2.4.2, the projection formula, the Cadman root-stack computation, the fractional-sum equivalence used in the weighted-blowup product, and the sign conventions around e−1=−t; I found no internal inconsistency. The reader's identified weakest assumption, Assumption 3.1.1 (with the stronger form 3.1.4), is a real restriction, but it is stated precisely and used exactly where smoothness of the weighted blowup and the Chow-ring presentations of [AOA23, QR22] are imported. That makes it a scope condition rather than a hidden failure. The main soft spot is the comparison K(X) ≅ IIμ(X), especially the Artin-approximation reduction in Proposition 2.3.3, which is terse and relies on standard but nontrivial stack-theoretic facts; however, the proof is detailed and I found no concrete gap. The only concrete blemish is the typo in Lemma 2.4.2 where 'a2/r1' should read 'a2/r2'. Since no load-bearing concern landed, I recommend no change to the reader's ACCEPT verdict.","tokens_in":28603,"tokens_out":31928,"duration_ms":342949,"concrete_test":"Recompute Lemma 2.4.2 for a non-cyclic diagonalizable H, e.g. H=μ4×μ2 with ζ the injection into the first factor and η the injection into the second factor, and with a character θ having argθ(ζ)+argθ(η)>1; verify directly that dim H1(Cζ,η,kθ)=1 exactly when that sum exceeds 1. This exercises the only part of the obstruction computation that is not literally cyclic, and if it fails, Proposition 2.4.1 and the multiplication rule (25) would need revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I checked the chain on which the central claim rests. Lemma 2.4.2's reduction of R1π∗f∗Eθ to Eθ ⊗ H1(Cζ,η,kθ) is sound for diagonalizable H, and the subsequent computation of H1(P1,p∗kθ) correctly yields a nonzero obstruction only when argθ(ζ)+argθ(η)>1; the apparent 'a2/r1' typo does not affect the conclusion. The step in Proposition 3.2.4 identifying the normal-bundle factor for ev3 with those a satisfying argζa+argηa=1 is also valid: for H=μp, p|a(c+d) together with ζa or ηa nontrivial forces the fractional parts {ac/p}+{ad/p}=1, so the condition is not a hidden mismatch. Sign conventions around t and e−1=−t agree with (19) and with j∗ being multiplication by −t. The substantive restriction is Assumptions 3.1.1 and 3.1.4, which are explicit: they are exactly what is needed for smoothness of the weighted blowup and for the [AOA23, QR22] presentations. This is a scope condition, not an internal inconsistency. The least-secure points are the terse Artin-approximation reduction in Proposition 2.3.3 and the affine-local verification in Lemma 3.2.1, but I found no concrete counterexample or missing hypothesis there.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper computes the stringy Chow ring A*_st(X) of smooth tame Deligne-Mumford stacks X=[X~/G] with X~ smooth and G diagonalizable, over a base field k that is not assumed to be algebraically closed. The central technical result is Proposition 2.4.1, which identifies the obstruction sheaf on the sector II_mu(zeta,eta) as a product of equivariant top Chern classes of character summands of the normal bundle. The authors then specialize to weighted blowups Bl_I•(X) of a smooth variety along a smooth center, imposing explicit regularity Assumptions 3.1.1 and 3.1.4, and obtain an explicit presentation of A*_st(Bl_I•(X)) (Propositions 3.2.4 and 3.3.2) together with a finite-generation criterion over A*(X) (Theorem 3.3.5).","tokens_in":28852,"tokens_out":22951,"duration_ms":218864,"significance":"If correct, the paper gives a substantial generalization of the complex-number computations of BCS05, GHK07, and JT10 to arbitrary base fields, and it provides an explicit, computable description in the weighted-blowup case. The obstruction-formula proof is carried out from first principles, and the regularity assumptions delimiting the weighted-blowup results are stated explicitly rather than hidden. The finite-generation theorem with explicit relations (29) is a concrete structural statement that should be useful for further examples. The paper would be a valuable reference for work on stringy Chow rings and weighted blowups.","major_comments":[],"minor_comments":[{"comment":"In the proof of Lemma 2.4.2, the second denominator in the displayed formula for b is printed as r1 twice ('-a2/r1' and 'a1/r1 + a2/r1 > 1'); it should be r2. The conclusion is unaffected once this typo is fixed, but the current text is confusing.","section":"Lemma 2.4.2"},{"comment":"Theorem 1.1.1 in the introduction states the base-change criterion using all subgroups H of G, while Corollary 2.4.3 states the criterion using all sectors I_mu(zeta); these are formally different conditions and the two statements should be reconciled.","section":"Section 1.1 and Corollary 2.4.3"},{"comment":"The reduction to complete local rings with separably closed residue field is only sketched: the fully-faithfulness assertion is attributed to finiteness of Isom schemes and essential surjectivity to Artin approximation. Since this isomorphism is the foundation for Theorem 2.3.1 and hence for the obstruction computation, a few more sentences explaining the descent and the Artin approximation step would make the proof substantially easier to verify.","section":"Proposition 2.3.3"},{"comment":"The product in (25) is taken over all a in Z, whereas the classes e_a were introduced only for a > 0; the convention that e_a = 1 for all a <= 0 except e_{-1} = -t makes this harmless, but an explicit sentence recording this convention would prevent confusion.","section":"Proposition 3.2.4, eq. (25)"},{"comment":"The phrase 'generated by ker i* and prod_{zeta in mu_{bk}} e_k' appears to contain a notation error; the intended relation should describe the kernel of the composition A*(X) -> A*(Y) -> A*(I_mu(zeta)) in terms of the weights or indices defining the closed immersion I_mu(zeta) -> Y, and it should be restated.","section":"Theorem 3.3.5, proof"},{"comment":"The same symbol Y is used for the center in X and for the exceptional divisor in X, which makes statements such as 'N' is a subbundle of N_{Y/X}' and the diagram in the proof hard to parse on first reading. Please use different letters, for example Z for the center.","section":"Lemma 3.3.3"}],"recommendation":"minor_revision","confidential_remarks":"The paper is well within the scope of the journal and the central results appear sound. The only concerns are local typos and the terseness of the Artin approximation reduction in Proposition 2.3.3. The self-citation [OW24] is used only for a technical factorization statement in Lemma 2.3.7 and does not create circularity."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here is my take on arXiv:2502.00259. It is a solid computational paper in intersection theory, and the main results look right to me. The genuinely new content is the stringy Chow ring for [X/G] with diagonalizable G over arbitrary fields, built on a new uniqueness statement for G-covers of P1 with prescribed gerbe monodromy at two points (Corollary 2.3.8), the explicit obstruction sheaf formula in Proposition 2.4.1, and the weighted blowup product formulas plus the finite generation criterion in Section 3. The weighted blowup part is the practical payoff: the stringy product becomes multiplication of explicit polynomials in t with an explicit class C_{ζη}(t), which is exactly what someone computing examples needs.\n\nThe paper is honest about its debts: the framework follows BCS05 and the Chow ring presentations come from AOA23 and QR22. The arbitrary-field generalization is real and nontrivial—it is not a cosmetic base change, since the cyclotomic inertia stack has a different structure over non-closed fields. The proof of Proposition 2.4.1 via Lemma 2.4.2 is clean: the H1 computation on the root stack reduces to a line bundle degree check, and the condition arg ζ + arg η > 1 falls out naturally. I also checked the ζη = 1 case, including the e−1 = −t factor, and the stress-test analysis is correct.\n\nThe soft spots are real but minor. The Artin approximation reduction in Proposition 2.3.3 is terse; a referee should ask for more detail there. There is an evident typo in Lemma 2.4.2 ('a2/r1' should be 'a2/r2'), which does not affect the conclusion but should be fixed. Section 3 works only in characteristic zero; this is stated in Assumption 3.1.1 but not in the abstract, which could mislead readers. The regularity assumptions 3.1.1 and 3.1.4 are explicit and structural—they are exactly what is needed for smoothness of the weighted blowup and for importing the [AOA23] and [QR22] presentations—so this is a scope restriction, not a hidden hypothesis. The one thing I could not fully verify by reading alone is the stack-theoretic identification of K(X) with IIμ(X) over non-algebraically closed fields; the argument is lengthy and delicate, but I found no concrete gap.\n\nWho is this for? Anyone computing stringy Chow rings of quotient stacks or weighted blowups, and people working on orbifold Chow rings over non-closed fields. It deserves a serious referee; with small corrections, I would be happy to see it published.","headline":"Solid computational paper: arbitrary-field stringy Chow rings and explicit weighted blowup formulas, with minor presentation gaps; deserves peer review.","tokens_in":29421,"tokens_out":2371,"would_cite":true,"duration_ms":23447,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14C15","14A20","14E05","14M25"],"pacs":[],"model":"deepseek-v4-flash","headline":"The stringy Chow ring of a quotient stack [X/G] and of weighted blowups is computed explicitly, with product formulas controlled by a character-sum condition.","keywords":["stringy Chow ring","cyclotomic inertia stack","obstruction sheaf","weighted blowup","diagonalizable group scheme","Deligne-Mumford stack","finite generation","Chow ring"],"falsifier":"Take the root-stack example of Section 3.4, $X=\\operatorname{Spec} k[x]$ with $J_\\bullet=((x),b)$, where formula (32) says $e_{\\zeta^{n_1}}\\star e_{\\zeta^{n_2}}=(-t)e_{\\zeta^{n_1+n_2}}$ when $n_1,n_2>0$ and $n_1+n_2\\le b$, and $e_{\\zeta^{n_1+n_2}}$ otherwise; compute the same product on $[\\mathbb A^1/\\mu_b]$ directly from the definition (8) of the stringy product and compare the two answers. If they differ, the character-sum rule in (25) is wrong.","tokens_in":28323,"feed_emoji":"📐","tokens_out":10651,"duration_ms":99175,"temperature":0.7,"pith_summary":"The paper computes the stringy Chow ring $A^*_{\\mathrm{st}}(\\mathcal X)$ — the Chow-group analogue of orbifold cohomology — for any smooth tame Deligne-Mumford stack of the form $\\mathcal X=[\\tilde X/G]$ with $G$ diagonalizable, over a field that need not be algebraically closed. It gives a complete description of the cyclotomic inertia stack, the obstruction sheaf that governs multiplication, and the resulting ring structure. It then specializes to weighted blowups $\\mathrm{Bl}_Y X$ of a smooth variety along a smooth center, writing $A^*_{\\mathrm{st}}$ very explicitly in terms of the Chow ring of the center and certain Chern classes $e_a$. The payoff is a finite-generation criterion: under a stronger regularity assumption, $A^*_{\\mathrm{st}}$ is generated over $A^*(X)$ by the fundamental classes of the twisted sectors exactly when the restriction map $A^*(X)\\to A^*(Y)$ is surjective, with explicit relations (29).","feed_headline":"Product formula computes stringy Chow rings of weighted blowups","feed_subtitle":"One character-sum condition controls the obstruction sheaf; finite generation follows when a restriction map is surjective.","key_machinery":"The load-bearing mechanism is the cyclotomic inertia stack $I^\\mu(\\mathcal X)=\\coprod_{\\zeta:\\mu_r\\to G}[\\tilde X^\\zeta/G]$, whose sectors are smooth closed substacks even when the ordinary inertia stack is only a twisted form; together with the obstruction sheaf $R^1\\pi_*f^*T_{\\mathcal X}$ on the double cyclotomic inertia, whose top Chern class is computed by Proposition 2.4.1. The proof of that computation rests on an explicit model of the universal twisted stable map: the moduli stack $K(\\mathcal X)$ is $\\mathrm{II}_\\mu(\\mathcal X)$, and on a sector the universal curve is $[\\tilde X^H/G]\\times [C_{\\zeta,\\eta}/H]$, where $C_{\\zeta,\\eta}$ is the unique $k$-curve obtained by rooting $\\mathbb P^1$ at three points with prescribed monodromy at two of them (Corollary 2.3.8). For weighted blowups, these ingredients are combined with the known presentations $A^*(\\mathcal Y)=A^*(Y)[t]/(\\prod_a e_a)$ and $A^*(\\mathcal X)=A^*(Y)[t]\\cdot t\\oplus A^*(X)/\\langle(P(t)-P(0))\\alpha,-i^*\\alpha\\rangle$, where $e_a$ is the $\\mathbb G_m$-equivariant top Chern class of the weight-$a$ summand of the weighted normal bundle.","core_discovery":"On each sector $\\mathrm{II}_\\mu(\\zeta,\\eta)=[\\tilde X^H/G]$ of the double cyclotomic inertia stack, the obstruction sheaf $R^1\\pi_*f^*T_{\\mathcal X}$ that defines the stringy product is identified as a vector bundle whose top Chern class is the product over characters $\\theta$ of the subgroup $H$ generated by $\\zeta$ and $\\eta$ of the equivariant top Chern class $c^G_{\\mathrm{top}}((N_{\\tilde X^H/\\tilde X})_\\theta)$, taken for those $\\theta$ with $\\arg\\theta(\\zeta)+\\arg\\theta(\\eta)>1$ (Proposition 2.4.1). Here $\\arg$ of a homomorphism $\\mu_r\\to \\mathbb G_m$ is the rational number $b/r$ from its Cartier dual. From this formula the paper derives the multiplication rule (25) for $A^*_{\\mathrm{st}}$ of a weighted blowup: sectors multiply with twist factors $C_{\\zeta\\eta}(t)=\\prod e_a$ over integer weights $a$ with $\\arg\\zeta^a+\\arg\\eta^a\\ge 1$ and $\\zeta^a$ or $\\eta^a\\ne 1$. Under Assumption 3.1.4 and in characteristic zero, Theorem 3.3.5 gives the full presentation $A^*_{\\mathrm{st}}(\\mathrm{Bl}_Y X)=A^*(X)[t,e_\\zeta]/\\mathcal J$ when $A^*(X)\\to A^*(Y)$ is surjective, and Proposition 3.3.2 shows the ambient subring $A^*_{\\mathrm{st}}(\\mathrm{Bl}_Y X)^{\\mathrm{amb}}$ is always a finitely generated $A^*(X)$-algebra with the same style of relations.","pith_inferences":["Because the obstruction formula is expressed purely in characters and equivariant Chern classes, the same presentation should extend to iterated weighted blowups and to root stacks along normal crossing divisors, where the Rees algebra condition is often automatic.","The clean separation of the base field from the combinatorics suggests that stringy Chow rings of toric and quasi-toric stacks could be used to compute arithmetic intersection numbers over number fields: the sectors are constant, and only the field of definition of cycles changes.","The ambient subring construction points to a natural filtration of $A^*_{\\mathrm{st}}(\\mathrm{Bl}_Y X)$ by the number of exceptional classes needed; one could test whether module-finiteness over $A^*(X)$ holds under assumptions weaker than surjectivity of $i^*$.","The '$\\arg\\theta(\\zeta)+\\arg\\theta(\\eta)>1$' cutoff is a discrete analogue of the age condition in orbifold cohomology; the same inequality may be the right correction term when comparing stringy Chow rings to orbifold cohomology after cycle class maps."],"forward_implications":["For a weighted blowup satisfying Assumption 3.1.4 with $A^*(X)\\to A^*(Y)$ surjective, $A^*_{\\mathrm{st}}(\\mathrm{Bl}_Y X)$ is completely described by the explicit generators and relations (29), so stringy products can be computed by polynomial arithmetic in $A^*(Y)[t]$.","Over any field extension $K/k$, there is a canonical ring map $A^*(X_K)\\otimes_{A^*(X)}A^*_{\\mathrm{st}}(X)\\to A^*_{\\mathrm{st}}(X_K)$, an isomorphism whenever the corresponding maps on sectors are; in particular for toric Deligne-Mumford stacks the stringy Chow ring is invariant under field extension (Corollary 2.4.4).","The obstruction formula (12) implies the product of two twisted-sector classes $e_\\zeta$, $e_\\eta$ lands in the sector of the product homomorphism $\\zeta\\eta$, with twist factors $e_a$ accumulating only when the two arguments sum to at least 1.","The equivalence in Lemma 3.3.3 shows that surjectivity (or finite generation) of $A^*(X)\\to A^*(Y)$ is detected on the exceptional divisor of the weighted blowup, a statement that holds for ordinary blowups as well."],"supporting_citations":[{"why":"Defines the stringy Chow ring and its product via twisted stable maps and the obstruction sheaf, the object computed throughout.","marker":"[AGV02]"},{"why":"Supplies the toric Deligne-Mumford formula over $\\mathbb C$ that Proposition 2.4.1 generalizes, along with the toric sector decomposition.","marker":"[BCS05]"},{"why":"Gives the Chow ring presentations (18) and (20) for weighted blowups used to write $A^*_{\\mathrm{st}}$ explicitly.","marker":"[AOA23]"},{"why":"Provides the regularity criteria and weighted blowup constructions behind Assumptions 3.1.1 and 3.1.4.","marker":"[QR22]"},{"why":"Contains the complex-number version of the unique $\\mathbb P^1$ cover with prescribed monodromy, generalized to arbitrary fields in Corollary 2.3.8.","marker":"[FG03]"},{"why":"Gives an earlier integral computation for toric quotient stacks over $\\mathbb C$, a special case of the present framework.","marker":"[JT10]"},{"why":"Provides the cohomological analogue for torus quotients over $\\mathbb C$, which the obstruction formula mirrors in Chow theory.","marker":"[GHK07]"},{"why":"Used to identify $R^1\\pi_*$ of the character bundle as a trivial bundle whose rank is $H^1(C_{\\zeta,\\eta},k_\\theta)$.","marker":"[HR17]"},{"why":"Gives the root-stack description of $C_{\\zeta,\\eta}$ and its tautological line bundles used in Lemma 2.3.4 and the age computation.","marker":"[Cad06]"},{"why":"Supplies the rigidity/factorization fact used to conclude injectivity of the comparison map $\\Phi$ on isomorphism classes.","marker":"[OW24]"}],"fun_headline_variants":["Weighted blowup stringy Chow product made explicit","Character sums give stringy Chow ring of weighted blowups","Explicit ring structure for weighted blowup stringy Chow","Stringy Chow ring of weighted blowups via character-sum twist","One twist factor describes weighted blowup stringy ring"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"For the weighted-blowup results, the paper assumes the Rees algebra defining the blowup is locally generated by a regular sequence in the declared degrees and has smooth center $V(I_1)$; if this regularity fails, the explicit sector decomposition and product formulas for $A^*_{\\mathrm{st}}(\\mathrm{Bl}_Y X)$ are not established.","fun_headline_variants_meta":{"raw":{"variants":["Weighted blowup stringy Chow product made explicit","Character sums give stringy Chow ring of weighted blowups","Explicit ring structure for weighted blowup stringy Chow","Stringy Chow ring of weighted blowups via character-sum twist","One twist factor describes weighted blowup stringy ring"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001086,"raw_usage":{"total_tokens":4552,"prompt_tokens":971,"completion_tokens":3581,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":587,"completion_tokens_details":{"reasoning_tokens":3501}},"tokens_in":587,"tokens_out":3581,"duration_ms":22651,"temperature":1.0,"reasoning_tokens":3501,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T19:38:23.060050+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the root-stack example of Section 3.4, $X=\\operatorname{Spec} k[x]$ with $J_\\bullet=((x),b)$, where formula (32) says $e_{\\zeta^{n_1}}\\star e_{\\zeta^{n_2}}=(-t)e_{\\zeta^{n_1+n_2}}$ when $n_1,n_2>0$ and $n_1+n_2\\le b$, and $e_{\\zeta^{n_1+n_2}}$ otherwise; compute the same product on $[\\mathbb A^1/\\mu_b]$ directly from the definition (8) of the stringy product and compare the two answers. If they differ, the character-sum rule in (25) is wrong.","supporting_citations":[],"review_version":1}