{"id":"458ecdcd-3d04-435f-b716-b22f701b2816","arxiv_id":"2506.16221","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The irreducible components of genus-0 stable map spaces to toric varieties are indexed by stable decorated trees that maximize a combinatorial invariant d_GGG.","lead":"This paper gives a combinatorial method to list the irreducible components of moduli spaces of genus-0 maps to any smooth projective toric variety, and it works out the first such description for a target that is not projective space. Generalists may care because these moduli spaces underlie Gromov-Witten invariants, and a component description is a step toward splitting curve-counting invariants.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Claimed dimension 7 for M_{0,0}(P^2,2) conflicts with the standard expected-dimension formula and with the paper's own cone computation.","rationale":"The reader identified the black-box open embedding (Eq. 37) as the weakest assumption and asked for explicit verification of the numerical checks in Sections 6.3-6.4. My reading found a more specific, independently checkable problem: the asserted dimensions in Theorem 6.3.1 are off by 2. The dimension of M_{0,0}(P^2,2) is a standard fact (5, not 7), and the same value follows from the paper's own abelian-cone dimension formula if the stack dimension of P is computed correctly. This error does not invalidate Theorem A, but it directly affects the paper's headline application and reinforces the need for explicit, reproducible computations in the examples. The open embedding concern remains a live but less concrete risk; the dimension arithmetic is the load-bearing defect that can be settled immediately by recomputation.","tokens_in":30221,"tokens_out":55689,"duration_ms":548287,"concrete_test":"Recompute dim M_{0,0}(P^2,2) using the standard formula d(N+1)+N-3 = 5, and independently compute dim S_{GGG0} from the paper's equations: dim P = -3 - 2 = -5, rank F = 3+3+3+1 = 10, so dim M_{GGG0} = 5. If both give 5, then Theorem 6.3.1's dimension assertion is false and the example decomposition must be rechecked: verify whether M_{GGG0} and M_{GGG1} still have equal dimensions and whether their union is the full moduli space.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 6.3.1 states dim(MGGG0) = dim(M_{0,0}(P^2,2)) = vdim(M_{0,0}(P^2,2)) = 7. The standard formula for the moduli of genus-0 maps to P^N gives vdim = d(N+1)+N-3, which for N=2,d=2 is 5, not 7. This is not a matter of convention: the space of conics in P^2 has dimension 5. Independently, the paper's own abelian-cone framework gives the same value. The base P = Pic^st_{0,0,X,2ℓ,σ} has stack dimension -3 - 2 = -5 (M_{0,0} is a PGL_2-gerbe over a point, and two BG_m factors from the two independent line bundles), and the generic rank of F is sum_ρ h^0(P^1,O(β·D_ρ)) = h^0(2)+h^0(2)+h^0(2)+h^0(0) = 3+3+3+1 = 10. Hence dim S_{GGG0} = dim P + rank F = 5, and since M_{GGG0} is open in S_{GGG0}, dim M_{GGG0} = 5. The claimed dimension 7 therefore contradicts the paper's own setup. This is a concrete, checkable arithmetic error in the main worked example; it does not by itself refute Theorem A, but it shows the numerical checks in Sections 6.3-6.4 are unreliable as presented.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a combinatorial description of the irreducible components of the moduli space of genus-0 stable maps to a smooth projective toric variety X. The main theorem (Theorem A, Theorem 5.4.2) states that the irreducible components are the closures of the loci M_GGG indexed by stable decorated marked trees GGG satisfying two conditions: the locus is nonempty, and a numerical inequality d_GGG >= d_GGG' holds for every edge-contraction GGG' of GGG. This is deduced from a general structure theorem (Theorem D) for irreducible components of an abelian cone over a smooth Noetherian Artin stack, applied to the cone S_{0,n,X,beta,sigma} containing M_{0,n}(X,beta) as an open substack. The paper also gives applications to M_{0,0}(Bl_p P^2, 2l) and M_{0,0}(Bl_p P^2, 3l), to quasimaps, and to the contraction morphism of [CR24].","tokens_in":30511,"tokens_out":41457,"duration_ms":458327,"significance":"If the main theorem is correct, it gives a genuinely combinatorial and checkable criterion for irreducible components of genus-0 stable map spaces to toric varieties, extending the projective-space story and providing a framework that could be used for Gromov-Witten/quasimap comparisons. The general abelian-cone theorem, proved via Nakayama-type dimension bounds and an elementary topological lemma, is a useful stand-alone result. The paper is largely self-contained in its central derivation, with the main external input being the open embedding of the stable map space into the abelian cone. However, the worked examples contain concrete numerical errors and rely on several unchecked computations; these problems do not by themselves disprove Theorem A but they make the applications unreliable as currently written.","major_comments":[{"comment":"The claimed dimensions in Theorem 6.3.1 are inconsistent with the paper's own setup. For X=Bl_pP^2 we have K_X=-3l+e, so the virtual dimension of M_{0,0}(X,2l) is (-K_X).(2l)+(dim X-3)=6-1=5, not 7. The abelian-cone computation in Sections 6.2-6.3 gives the same value: for sigma=sigma_{0,2}, the base Pic^st_{0,0,X,2l,sigma} has dimension -3-2=-5, and the generic rank of F is h^0(2)+h^0(2)+h^0(2)+h^0(0)=3+3+3+1=10, hence dim S_GGG0 = -5+10 = 5. The equality in the proof with dim M_{0,0}(P^2,2) is also false, since that space has dimension 2(2+1)+2-3=5, not 7. The stated dimension 7 for M_GGG0 and M_GGG1 in Theorem 6.3.1 should be corrected to 5, and the subsequent uses of this dimension should be rechecked.","section":"Sections 6.3 and 6.4"},{"comment":"The numerical verification of the inequalities d_GGG >= d_GGG' is the load-bearing content of the examples, but almost all of it is deferred to 'one can check' statements. For instance, after Lemma 6.3.2 the text says 'For i in {2,3,4}, one can check that d_GGGi - d_GGG0 < 0', and in Section 6.4 it says 'One can compute that d_GGG2 - d_GGG0 = 1 and d_GGGi - d_GGG0 = 0 for all other graphs in Figure 6'. These are finite but nontrivial computations of h^0 of line bundles on reducible genus-0 curves and of the edge-contraction poset. The paper should provide the actual values, for example a table listing h^0(GGG,L_rho), i_GGG, #E(GGG), and d_GGG-d_GGG0 for every graph in Table 2 and Figure 6, together with the contraction relations. The dimension error in Theorem 6.3.1 shows that the numerical assertions in this section are not reliable as they stand.","section":"Sections 6.3 and 6.4"},{"comment":"The definition of Airred_1(X) in Section 6.1 and its use in Section 6 are not mathematically precise. The text defines Airred_1(X) as the set of classes 'that can be represented by an irreducible curve' and asserts the equality Airred_1(X)=Z_{\\ge0}e \\cup (Z_{\\ge0}s+Z_{\\ge0}l). As written, this equality is false: 2s and 2e are not classes of irreducible curves (for 2s, adjunction gives 2p_a-2=(2s)^2+K_X.(2s)=-4, so p_a=-1). What is evidently intended is that a class such as 2s is allowed because it is the pushforward of a degree-2 cover of an irreducible curve of class s. This distinction matters because the condition M_GGG nonempty in Theorem 5.4.2 is controlled by the 'irreducible' decorated trees of Definition 3.3.5. Please state the intended definition precisely and prove the characterization used to enumerate Table 2 and Figure 6.","section":"Sections 3.3 and 6.1"},{"comment":"Theorem 5.4.2 depends on the open embedding M_{0,n}(X,beta) \\hookrightarrow S_{0,n,X,beta,sigma} in Equation (37). The paper cites [CL12] but does not state the precise hypotheses or verify that the stability condition in Equation (26) is equivalent to the stability of the corresponding stable maps. Since the component criterion would collapse if this embedding were not open, the paper should either prove this toric case using Cox's presentation, or give a precise reference containing the statement for smooth projective toric targets. As it stands, this is a central black box in the proof of Theorem A.","section":"Section 5.2, Equation (37)"}],"minor_comments":[{"comment":"The statement says 'with GGG0,...,GGG4 in Gamma^st_{0,0}(Bl_pP^2,2l)' but the theorem is about 3l; this should be corrected.","section":"Theorem 6.4.1"},{"comment":"The caption refers to 'Bl_pP^3' but the target throughout Section 6.4 is Bl_pP^2.","section":"Figure 6 caption"},{"comment":"There are repeated misstatements that Section 6.4 describes components of M_{0,0}(Bl_pP^2,2l); the passage should refer to 3l.","section":"Introduction, Section 1.3 and Section 1.4"},{"comment":"The equality dim M_GGG0 = dim M_{0,0}(P^2,2) is a typo if the intended space is M_{0,0}(Bl_pP^2,2l); as written it is false and should be corrected.","section":"Proof of Theorem 6.3.1"},{"comment":"The remark begins 'As in Remark 5.3.5', which is a self-reference; it should refer to the preceding remark or to Notation 5.3.1.","section":"Remark 5.3.5"},{"comment":"The last column uses an unexplained symbol (printed as 'Ę') for inclusions; the entry for GGG8 appears to conflict with the surrounding text and should be clarified.","section":"Table 2"}],"recommendation":"major_revision","confidential_remarks":"The central theorem appears plausible and the abelian-cone argument is well structured, but the worked examples contain a demonstrable dimension error and numerous unverified numerical assertions. I would not reject based on Theorem A alone, but the applications must be corrected and the numerical checks made explicit before the paper is publishable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here is the short version: Theorem A gives a combinatorial description of the irreducible components of M_{0,n}(X,β) for smooth projective toric X, via a study of abelian cones over Artin stacks. The strategy is clean and, as far as I can tell, the main proof is sound. But the paper's main worked example contains an arithmetic error: it claims dim M_{0,0}(P^2,2) = 7, while the actual dimension is 5. The paper's own framework gives 5: the base Picard stack has dimension -5, the generic rank of the cone is 10, so the main component of S has dimension 5. Thus the dimension statements in Theorem 6.3.1, and the relative claims in Section 6.4, are not reliable as written.\n\nWhat is genuinely new: the abelian cone theorem (Theorem D) and the decorated-tree stratification. The authors reduce a geometric component problem to a finite, purely combinatorial check on decorated trees. That is a real step forward. For targets other than projective space, no general component description existed; the blow-up examples are the first non-P^n computations, and the claimed decomposition into two components for 2ℓ (or five for 3ℓ) may well be correct. The paper is also honest about the literature: the only self-citation used outside of Section 7 is [CR24], and that is not part of the main proof.\n\nWhere it is soft: besides the dimension error, the crucial numerical checks in Sections 6.3 and 6.4 are asserted with \"one can check\" rather than demonstrated. For the advertised applications, those checks are the whole content. A referee should ask for the checks to be written out, or for a small script or table that makes them reproducible. There is also a black-box reliance on the open embedding M_{0,n}(X,β) into the abelian cone following [CL12]. I do not doubt that result, but the paper's main theorem inherits its validity from it, so it should be stated more explicitly.\n\nWho this is for: researchers in Gromov-Witten theory and toric moduli. The main theorem deserves attention; the examples as they stand do not.\n\nRecommendation: send to peer review, but require a revision that fixes the dimension computation and supplies the missing finite checks. The core idea is good enough that a careful referee should engage.","headline":"A genuinely new combinatorial criterion for irreducible components, but the flagship example has a concrete dimension error that needs correction before the applications are credible.","tokens_in":31074,"tokens_out":12092,"would_cite":false,"duration_ms":117856,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14N35","14M25","14D22","14H10"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that irreducible components of genus-0 stable maps to smooth projective toric varieties are classified by finite decorated trees.","keywords":["moduli of stable maps","toric varieties","irreducible components","abelian cones","decorated dual graphs","Picard stack","quasimaps","blow-up of the projective plane"],"falsifier":"For a small toric example, enumerate all stable decorated trees, compute $d_{\\mathcal{G}}$ by $h^0$ on $\\mathbb{P}^1$, and independently compute the irreducible components of $\\overline{\\mathcal{M}}_{0,n}(X,\\beta)$ by a different method; a concrete test is the claimed decomposition of $\\overline{\\mathcal{M}}_{0,0}(\\mathrm{Bl}_{\\mathrm{pt}}\\mathbb{P}^2,2\\ell)$, where $\\overline{\\mathcal{M}}_{\\mathcal{G}_1}$ must be a genuinely non-empty second 7-dimensional component not contained in the main component.","tokens_in":29979,"feed_emoji":"🌳","tokens_out":14704,"duration_ms":130671,"temperature":0.7,"pith_summary":"The paper proves that, for every smooth projective toric variety $X$ and every effective curve class $\\beta$, the irreducible components of $\\overline{\\mathcal{M}}_{0,n}(X,\\beta)$ are exactly the closed loci $\\overline{\\mathcal{M}}_{\\mathcal{G}}$ indexed by isomorphism classes of stable decorated marked trees $\\mathcal{G}\\in\\Gamma^{\\mathrm{st}}_{0,n}(X,\\beta)$ whose loci are non-empty and whose numerical invariant $d_{\\mathcal{G}}$ dominates $d_{\\mathcal{G}'}$ for every edge contraction $\\mathcal{G}'$ of $\\mathcal{G}$. The condition is finite and combinatorial: it asks for $h^0$ of line bundles on $\\mathbb{P}^1$. The result is applied to the blow-up of the projective plane: $\\overline{\\mathcal{M}}_{0,0}(\\mathrm{Bl}_{\\mathrm{pt}}\\mathbb{P}^2,2\\ell)$ has two irreducible components, both of dimension 7, and $\\overline{\\mathcal{M}}_{0,0}(\\mathrm{Bl}_{\\mathrm{pt}}\\mathbb{P}^2,3\\ell)$ has five, one of dimension 9. This gives the first description of a smoothable locus for a target that is not a projective space.","feed_headline":"Tree combinatorics classifies stable-map moduli components","feed_subtitle":"For toric targets, component counting reduces to h^0 counts on P^1; examples include the blow-up of P^2.","key_machinery":"The load-bearing mechanism is a general theorem about abelian cones: if $\\mathcal{B}$ is a smooth Noetherian Artin stack and $\\mathcal{F}$ a coherent sheaf on it, then after stratifying $\\mathcal{B}$ by locally closed irreducible strata on which the rank of $\\mathcal{F}$ is constant, the irreducible components of $\\operatorname{Spec}_{\\mathcal{B}}\\operatorname{Sym}(\\mathcal{F})$ are the preimages of strata whose rank-minus-codimension invariant is maximal among the strata they contain. The paper applies this with $\\mathcal{B}=\\mathcal{P}\\mathrm{ic}^{\\mathrm{st}}_{0,n,X,\\beta,\\sigma}$, a product of Picard stacks over prestable genus-zero curves, and with $\\mathcal{F}$ the direct sum of $R^1$ pushforwards of dual universal line bundles. The strata are labelled by stable decorated marked trees, which record the dual graph of the curve, a curve class at each vertex, and the marked points. On each stratum the relevant invariant becomes the combinatorial number $d_{\\mathcal{G}}$, so the component-finding problem reduces to comparing these numbers under edge contractions.","core_discovery":"The central claim is Theorem A: for a smooth projective toric variety $X$ and an effective curve class $\\beta$, the irreducible components of $\\overline{\\mathcal{M}}_{0,n}(X,\\beta)$ are exactly the closures $\\overline{\\mathcal{M}}_{\\mathcal{G}}$ with $\\mathcal{G}\\in\\Gamma^{\\mathrm{st}}_{0,n}(X,\\beta)$ such that $\\mathcal{M}_{\\mathcal{G}}\\neq\\emptyset$ and $d_{\\mathcal{G}}\\ge d_{\\mathcal{G}'}$ whenever $\\mathrm{Pic}_{\\mathcal{G}}\\subseteq\\mathrm{Pic}_{\\mathcal{G}'}$. The inclusion of Picard strata holds precisely when $\\mathcal{G}'$ is obtained from $\\mathcal{G}$ by contracting edges. The number $d_{\\mathcal{G}}$ is the excess of the total $h^0$ of the relevant line bundles over the constant value on smooth genus-zero curves, minus the number of edges of the dual graph, so the theorem converts the geometric component problem into finitely many $h^0$ computations on $\\mathbb{P}^1$.","pith_inferences":["Beyond the paper: the proof of the abelian-cone component theorem does not use genus zero or toricity, so the same rank-stratification strategy may extend to higher genus or other targets whenever an analogous open embedding into an abelian cone exists; the paper does not establish such an extension.","Beyond the paper: if Theorem A is correct, the component classification is algorithmically computable, and the only non-combinatorial bottleneck in each example is the non-emptiness condition $\\mathcal{M}_{\\mathcal{G}}\\neq\\emptyset$.","Beyond the paper: the $3\\ell$ example, with an extra component of dimension 9, suggests that genus-zero Gromov-Witten invariants of toric varieties may split into contributions from individual components; deriving such a splitting would be a natural next step that the paper leaves open.","Beyond the paper: repeating the enumeration on another toric surface or on larger degrees, such as $\\beta=4\\ell$ for $\\mathrm{Bl}_{\\mathrm{pt}}\\mathbb{P}^2$, would provide a direct computational test of the criterion."],"forward_implications":["For any fixed toric $X$ and class $\\beta$, listing the irreducible components of $\\overline{\\mathcal{M}}_{0,n}(X,\\beta)$ becomes a finite enumeration of decorated trees followed by $h^0$ comparisons; only non-emptiness of the corresponding map loci remains a genuine geometric check.","For $X=\\mathrm{Bl}_{\\mathrm{pt}}\\mathbb{P}^2$ and $\\beta=2\\ell$, the space has exactly two components, both of dimension 7, and the paper describes the generic point of their intersection.","For $\\beta=3\\ell$, the space has five components, four of dimension 8 and one of dimension 9, so an extra component can be strictly larger than the main component.","The same combinatorial criterion, with a modified stability condition on trees, describes the irreducible components of genus-0 stable quasimaps to toric varieties.","In the two-marked degree $2\\ell$ blow-up example, the locus where the contraction morphism from maps to quasimaps is defined is a union of irreducible components, and the paper records where each component is sent."],"supporting_citations":[{"why":"Supplies the open embedding of the stable-map stack into the abelian cone over the Picard stack, the bridge used to transfer the component classification.","marker":"[CL12]"},{"why":"Gives the line-bundle-and-section description of maps to a toric variety, which defines the abelian cone and the decorated curve classes.","marker":"[Cox95]"},{"why":"Provides the stack-theoretic facts about irreducible components, Fitting ideals, and smooth descent used in the proof of the abelian-cone theorem.","marker":"[Sta22]"},{"why":"Frames the problem of irreducible components of abelian cones that the paper generalizes to Artin stacks.","marker":"[Sta03]"},{"why":"Supplies the toric intersection theory, including the Picard basis and relations, used to associate curve classes to vertices of decorated trees.","marker":"[CLS11]"}],"fun_headline_variants":["Toric maps: components from tree combinatorics","Stable-map components for toric targets via excess counts","Genus-0 toric moduli: irreducible components combinatorial","First non-projective smoothable locus described"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The classification rests on the previously established open embedding $\\overline{\\mathcal{M}}_{0,n}(X,\\beta)\\hookrightarrow S$ taken as a black box: if that embedding were not open in the abelian cone, the irreducible components of the moduli space would not be obtained by intersecting the cone's components, and the main theorem would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Toric maps: components from tree combinatorics","Stable-map components for toric targets via excess counts","Genus-0 toric moduli: irreducible components combinatorial","First non-projective smoothable locus described"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000372,"raw_usage":{"total_tokens":1952,"prompt_tokens":874,"completion_tokens":1078,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":490,"completion_tokens_details":{"reasoning_tokens":1016}},"tokens_in":490,"tokens_out":1078,"duration_ms":10778,"temperature":1.0,"reasoning_tokens":1016,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:28:11.125743+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a small toric example, enumerate all stable decorated trees, compute $d_{\\mathcal{G}}$ by $h^0$ on $\\mathbb{P}^1$, and independently compute the irreducible components of $\\overline{\\mathcal{M}}_{0,n}(X,\\beta)$ by a different method; a concrete test is the claimed decomposition of $\\overline{\\mathcal{M}}_{0,0}(\\mathrm{Bl}_{\\mathrm{pt}}\\mathbb{P}^2,2\\ell)$, where $\\overline{\\mathcal{M}}_{\\mathcal{G}_1}$ must be a genuinely non-empty second 7-dimensional component not contained in the main component.","supporting_citations":[],"review_version":2}