{"id":"45f69d5a-39db-4aed-9db5-63d00494bdf0","arxiv_id":"2505.07645","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For del Pezzo surfaces of degree at most 5 and for smooth cubic hypersurfaces and intersections of two quadrics over F_q(t), the paper obtains upper bounds N(q^e) = O((C q)^e) with C independent of e, so the exponent is close to Batyrev-Manin when q is large.","lead":"This paper proves upper bounds for the number of rational points of bounded height on Fano varieties over function fields, using the dimension of spaces of rational curves. The bounds come close to the predicted linear growth for del Pezzo surfaces of low degree, cubic hypersurfaces, and intersections of two quadrics when the constant field is large.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.1 may silently depend on BHLT hypotheses that fail for purely inseparable or quasi-elliptic del Pezzo surfaces in characteristic 2; the unstated characteristic condition in Theorem 3.2 is the load-bearing point.","rationale":"The reader's weakest assumption identifies the same linchpin: Theorem 1.1 for del Pezzo surfaces is only as strong as Theorem 3.2, and the paper never verifies the characteristic hypotheses of [3, Theorem 1.1]. The paper's Remark 3.3 makes a positive-characteristic smoothness assertion that is not self-evidently true, and in characteristic 2 there are smooth del Pezzo surfaces of degree 2 whose anticanonical divisors are all non-reduced. If BHLT excludes these, the proof of Theorem 1.1 has no cover for them. The false coordinate-hyperplane step in the proof of Proposition 2.1 is a real defect in the paper's self-contained proof, but Proposition 2.1 is a known lemma and can be repaired by citing [12], [18], [16], or [10]; it does not by itself threaten the truth of the counting bound. The positive-characteristic lower bound for dim Mor should also be supplied. All of this supports the reader's CONDITIONAL verdict rather than changing it, pending a precise check of [3].","tokens_in":16424,"tokens_out":39210,"duration_ms":408668,"concrete_test":"Read [3, Theorem 1.1] and Section 3 and record every hypothesis beyond 'smooth del Pezzo', especially characteristic restrictions and any requirement that a general member of |-K_X| be smooth. Then take a smooth quartic f_4 over F_2 and the degree-2 del Pezzo surface X = V(y^2 - f_4(u,v,w)) in P(2,1,1,1), and check whether X satisfies those hypotheses. If it does not, compute the dimensions of the irreducible components of Mor(P1,X,e) for e = 2 and e = 3 over an algebraic closure of F_2. If some component has dimension greater than e + 2 and is not contained in the locus of multiple covers of a rational curve C with -K_X.C <= 1, then Theorem 1.1 fails for this X; if the dimension statement of Theorem 3.2 still holds for this X, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central del Pezzo result rests entirely on Theorem 3.2, imported from [3, Theorem 1.1] for arbitrary smooth del Pezzo surfaces over F_q. The paper does not state the hypotheses of [3, Theorem 1.1], and Remark 3.3 asserts that the weak del Pezzo condition 'a general member of |-K_X| is smooth' is automatic for smooth del Pezzo surfaces, citing [3, Section 3]. This is not automatic in all characteristics: in characteristic 2, a smooth del Pezzo surface of degree 2 can be written as y^2 = f_4(u,v,w) in P(2,1,1,1) with f_4 a smooth plane quartic, giving a purely inseparable double cover of P^2 whose anticanonical divisors are all non-reduced preimages of lines. If [3] excludes such surfaces, or if its classification of non-expected components changes for them, Theorem 1.1 is unproved or false for those F_q. The lower bound dim Mor >= e + 2 is also cited to Debarre without an explicit positive-characteristic proof or reference, and every dimension conclusion in the proof uses it. The reader's conditional verdict is therefore appropriate, but the decisive question is the exact scope of [3, Theorem 1.1].","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes an elementary counting method over finite fields and applies it to moduli spaces of rational curves on Fano varieties over F_q. Its main result (Theorem 1.1) states that for a smooth del Pezzo surface X of degree 1 ≤ d ≤ 5 over F_q, with U the complement of all rational curves C with -K_X·C ≤ 1, the number of F_q(t)-points of anticanonical height q^e is O(C_d^e q^e), with explicit constants. Theorem 1.2 gives a general bound for Fano complete intersections in terms of the dimension of the relevant moduli space, and Theorem 1.3 establishes that spaces of rational curves on smooth cubic hypersurfaces (char > 3) and smooth intersections of two quadrics (dim ≥ 3) have the expected dimension, leading to Theorem 1.4 bounds. The proofs use a Bezout-type point-counting lemma (Proposition 2.1), degree bounds for affine cones of moduli spaces, and imported results: [3, Theorem 1.1] for del Pezzo surfaces and bend-and-break arguments following [22], [13], [29]. The paper is clearly written and the numerical constants are explicit.","tokens_in":16678,"tokens_out":22825,"duration_ms":217318,"significance":"If the hypotheses of the imported results are satisfied, Theorem 1.1 is a substantial step: for large fixed q it gives upper bounds arbitrarily close to linear growth, improves on known bounds over F_q(t) for large q, and gives the first F_q(t)-version of Browning's problem on cubic surfaces. The general counting lemma and the positive-characteristic expected-dimension results for complete intersections are of independent interest. However, the main del Pezzo theorem rests entirely on the unstated characteristic scope of [3], and the self-contained proof of Proposition 2.1 is flawed as written; both issues must be addressed before the results can be considered established.","major_comments":[{"comment":"The proof asserts that one can find an index i0 such that the hyperplane x_i0 = a intersects Z2 properly for every a in F_q. This is false in general: for Z2 = V(x1) union V(x2) in A^2, the hyperplane x1 = 0 contains the component V(x1) and the hyperplane x2 = 0 contains V(x2), so no coordinate hyperplane has the required property for all a. Consequently, the displayed summation over a in F_q of #(Z2 ∩ Ha)(F_q) is not justified. The proposition itself is true, as the author notes by citing [12], [18], [16], [10], so this is a flaw in the self-contained proof rather than in the statement; the proof should be repaired or replaced by a citation to one of the existing proofs.","section":"2, proof of Proposition 2.1"},{"comment":"The characteristic hypotheses of [3, Theorem 1.1] are not stated. Remark 3.3 claims that the weak del Pezzo condition 'a general member of |-K_X| is smooth' is automatic for smooth del Pezzo surfaces, citing [3, Section 3]. This is not correct in characteristic 2: a smooth del Pezzo surface of degree 2 given by y^2 = f_4(u,v,w) in P(2,1,1,1) has anticanonical members that are non-reduced double covers of lines, so the general member of |-K_X| is not smooth. Unless [3, Theorem 1.1] nevertheless applies to such surfaces, Theorem 1.1 is overclaimed in all characteristics. Please state the exact hypotheses of [3, Theorem 1.1], confirm they hold for all smooth del Pezzo surfaces over F_q, or add the necessary characteristic restriction to Theorem 1.1.","section":"3, Theorem 3.2 and Remark 3.3"}],"minor_comments":[{"comment":"The constants C_2 and C_1 are printed as '44' and '66', which is easy to misread; they should be typeset as 4^4 and 6^6.","section":"1, Theorem 1.1 display"},{"comment":"The text says 'f in F_q[u,v,w] is a binary quartic form'; since f has three variables, this should be 'ternary quartic form'.","section":"3, d = 2 paragraph"},{"comment":"Corollary 4.10 is stated without the characteristic assumption, but its proof relies on Lemma 4.9, which requires char(K) != 2,3. The corollary should either be restricted to char != 2,3 or an argument covering characteristics 2 and 3 should be supplied.","section":"4, Corollary 4.10 and Lemma 4.9"},{"comment":"In the incidence argument, the fiber whose dimension is n-3 by Lemma 4.2 is the fiber of p1 over x, not 'p2(x)^{-1}'; the notation should be corrected.","section":"4, proof of Corollary 4.10"},{"comment":"In the base case of the induction, the phrase 'since we assume n + d − 1 ≥ 2' appears to be a typo for 'n + 1 − d ≥ 2', which is the actual hypothesis used in the paper.","section":"5, Proposition 5.3"},{"comment":"The open substack of stable maps with irreducible domain is denoted by the same symbol as the full Kontsevich stack M_{0,k}(X,e); a distinct notation (for example M^0_{0,k}(X,e)) would avoid confusion. Also, in the definition of M_{0,k}(X,e), the marked points are listed as p1,...,pn but should be p1,...,pk.","section":"5, notation for Kontsevich spaces"},{"comment":"The lower bound 'dim every non-empty irreducible component of Mor(P1,X,e) is at least e + dim(X)' is cited to Debarre [14, Theorem 2.6], which is a standard characteristic-0 reference; since the paper works over possibly positive characteristic, a reference covering positive characteristic (e.g., Kollár, Rational Curves on Algebraic Varieties, or Harris-Roth-Starr) or a brief justification would be more appropriate.","section":"1, dimension lower bound"}],"recommendation":"major_revision","confidential_remarks":"The decisive point for Theorem 1.1 is the exact statement of [3, Theorem 1.1]; I recommend the editor ensure that the author verifies whether that theorem applies to all smooth del Pezzo surfaces in every characteristic, particularly characteristic 2. The flaw in the proof of Proposition 2.1 is real but easily repaired, since the result is known and cited. The paper does not exhibit circularity: the main geometric input is an independent published theorem, and prior work of the author is cited only for comparison."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [Colleague],\n\nHere's my take on Glas's paper. The core move is simple and effective: instead of Grothendieck-Lefschetz or the circle method, he counts rational points over F_q(t) by bounding the number of F_q-points on moduli spaces of rational curves via a degree-dimension estimate. That yields O(C^e q^{dim}) bounds, and when the moduli space has the expected dimension, the exponent is essentially linear in the height with a constant C that is independent of e. This is a genuinely useful observation, and Theorem 1.1 is the first function-field bound of this sharpness for low-degree del Pezzo surfaces. For large fixed q it resolves a function-field analogue of Browning's challenge for cubic surfaces. Theorem 1.3, extending expected-dimension results for cubic hypersurfaces and intersections of two quadrics to positive characteristic, is also a solid piece of work, with a careful bend-and-break argument and a nice treatment of Eckardt points.\n\nThe soft spots are real but concentrated. Proposition 2.1 is a standard bound, but the self-contained proof as written is wrong: the claim that some coordinate hyperplane family x_i = a intersects every component properly for all a in F_q fails for Y = V(x_1) union V(x_2) in A^2, since the hyperplane x_1=0 contains one component. The result itself is well-known, so this is fixable, but it should be corrected.\n\nThe more serious issue is Theorem 3.2. The del Pezzo bounds depend entirely on BHLT's theorem, and the paper doesn't state its characteristic hypotheses. Remark 3.3 claims that 'a general member of |-K_X| is smooth' is automatic for smooth del Pezzo surfaces, but that is false in characteristic 2 for degree-2 surfaces that are purely inseparable double covers of P^2: the anticanonical divisors are non-reduced preimages of lines. If BHLT excludes such surfaces--or if their expected-dimension classification changes for them--Theorem 1.1 is overclaimed. The author needs to spell out the exact scope of [3, Theorem 1.1] and either prove or rule out the char-2 cases. The lower bound dim Mor >= e + dim X is also cited to Debarre without a positive-characteristic argument; that bound likely holds, but it deserves a sentence or a proper reference.\n\nThe rest of the counting framework, including the weighted projective space treatments for d=2 and d=1, is careful and correct. The paper deserves a serious referee: the strategy is new, the results are important if the characteristic questions resolve, and the gaps are likely fixable. It should not be desk-rejected, but the referee needs to press on the BHLT hypothesis.","headline":"A clean moduli-counting trick yields near-linear function-field bounds for del Pezzo surfaces, but the del Pezzo half rests on unverified characteristic assumptions and the counting lemma has a false proof.","tokens_in":17207,"tokens_out":5366,"would_cite":false,"duration_ms":46827,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14H10","11D45","11P55","14G05","14J70"],"pacs":[],"model":"deepseek-v4-flash","headline":"For del Pezzo surfaces over $\\mathbb{F}_q(t)$, rational-point counts off the exceptional curves come arbitrarily close to linear growth when $q$ is large.","keywords":["rational curves","moduli spaces","del Pezzo surfaces","Batyrev–Manin conjecture","positive characteristic","F_q(t)-points","cubic hypersurfaces","intersections of two quadrics"],"falsifier":"Check the hypotheses of the cited theorem in the source: if the classification of non-expected-dimensional components of the del Pezzo morphism space is proved only away from characteristic $2$ or $3$, then Theorem 1.1 is overclaimed for the remaining characteristics. Alternatively, compute $\\operatorname{Mor}(\\mathbb{P}^1,X,e)$ for a smooth del Pezzo surface of degree $4$ or $5$ over an algebraically closed field of characteristic $2$ and look for an irreducible component of dimension strictly above $e+2$ whose general member is not a multiple cover of a line; finding one would break the key input.","tokens_in":16198,"feed_emoji":"","tokens_out":12593,"duration_ms":114613,"temperature":0.7,"pith_summary":"This paper proves strong upper bounds for the number of $\\mathbb{F}_q(t)$-points of bounded height on several Fano varieties over finite fields. The main result is that on a smooth del Pezzo surface of degree $1 \\le d \\le 5$, after deleting all rational curves whose anticanonical degree is at most $1$, the number of points of height $q^e$ is $O(C_d^e q^e)$, with constants $C_5=1024$, $C_4=16$, $C_3=27$, $C_2=4^4$, $C_1=6^6$ independent of $e$. Since $C_d$ is fixed, for large $q$ this is arbitrarily close to the linear growth $q^{e(1+\\varepsilon)}$ predicted by the Batyrev--Manin conjecture. The method also shows that the moduli spaces of rational curves on smooth cubic hypersurfaces and on smooth intersections of two quadrics of dimension at least $3$ have the expected dimension in positive characteristic, yielding analogous point-counting bounds.","feed_headline":"Point counts on del Pezzo surfaces approach linear growth","feed_subtitle":"A dimension bound for rational-curve moduli brings these counts arbitrarily close to linear growth for large q.","key_machinery":"The engine is an elementary finite-field point count: any locally closed subvariety $Y\\subset \\mathbb{A}^N$ over $\\mathbb{F}_q$ satisfies $\\#Y(\\mathbb{F}_q) \\le \\deg(Y)q^{\\dim Y}$, proved by slicing with hyperplanes and applying B\\'ezout's inequality. Because $\\mathbb{F}_q(t)$-points of height $q^e$ are exactly the $\\mathbb{F}_q$-points of the morphism scheme $\\operatorname{Mor}(\\mathbb{P}^1,X,e)$, this turns counting into a dimension estimate for those moduli spaces. For del Pezzo surfaces, the dimension estimate is imported: the only components of $\\operatorname{Mor}(\\mathbb{P}^1,X,e)$ of larger than expected dimension are multiple covers of rational curves $C$ with $-K_X\\cdot C \\le 1$, so deleting those curves leaves components of dimension $e+2$. For cubic hypersurfaces and intersections of two quadrics, the paper proves the expected-dimension statement directly, using a bend-and-break reduction to lines together with the geometry of Fano schemes of lines and a positive-characteristic finiteness result for Eckardt points.","core_discovery":"On the author's own terms, the central discovery is that point counting over $\\mathbb{F}_q(t)$ can be reduced to dimension bounds for moduli spaces of rational curves, and that those bounds are available in positive characteristic for del Pezzo surfaces of degree at most $5$. The precise form is $N_U(q^e) = O(C_d^e q^e)$ for $U$ the complement of all rational curves $C$ with $-K_X \\cdot C \\le 1$, with the implied constant depending only on $q$. Because $C_d$ is a fixed constant, taking $q$ large forces the exponent of $q^e$ arbitrarily close to $1$, matching the growth predicted by Batyrev--Manin. In the same framework, smooth cubic hypersurfaces of dimension at least $3$ satisfy $N_X(e)=O(27^e q^{e(n-3)})$ and smooth intersections of two quadrics satisfy $N_X(e)=O(16^e q^{e(n-4)})$.","pith_inferences":["If the imported del Pezzo classification carries an unstated characteristic restriction, Theorem 1.1 would need to be narrowed to the range where that theorem holds; checking the hypotheses of the cited source is the cheapest way to test the main result.","The same counting strategy should transfer to any Fano variety once a positive-characteristic expected-dimension statement is known; the constants in the bound would come from the degrees of the defining equations, so improving those constants is a separate combinatorial problem.","Because the argument identifies height-$q^e$ points with morphisms, it is particular to $\\mathbb{F}_q(t)$; moving to number fields would require an arithmetic analogue of the moduli-space dimension bounds, which is not provided here."],"forward_implications":["For any smooth del Pezzo surface of degree $1\\le d\\le 5$ over $\\mathbb{F}_q$, the bound $N_U(q^e)\\ll (C_d q)^e$ gives exponents $1+\\log_q C_d$, which tend to $1$ as $q\\to\\infty$; this locates the dominant contribution in the deleted low-degree curves.","For smooth cubic surfaces over $\\mathbb{F}_q$ with $q$ sufficiently large, the theorem yields $N_U(B)\\ll B^{4/3-\\theta}$ for any fixed $\\theta<1/3$, settling the $\\mathbb{F}_q(t)$-version of a problem raised by Browning.","For every $e\\ge 1$, the moduli spaces of degree-$e$ rational curves on smooth cubic hypersurfaces (in characteristic $>3$) and on smooth intersections of two quadrics of dimension at least $3$ have the expected dimension, extending characteristic-zero results.","The general counting theorem shows that any Fano variety over $\\mathbb{F}_q$ whose morphism spaces of rational curves have the expected dimension for all large $e$ automatically has near-linear growth; the bottleneck is purely the moduli-space dimension theory."],"supporting_citations":[{"why":"Supplies the classification result for the del Pezzo morphism space that makes Theorem 1.1 work.","marker":"[3]"},{"why":"Justifies the expected-dimension lower bound for components of the morphism space.","marker":"[14]"},{"why":"The bend-and-break lemma, argued here to be characteristic-free, that drives the expected-dimension proofs.","marker":"[22]"},{"why":"The earlier characteristic-zero treatment of cubic hypersurfaces that this paper extends.","marker":"[13]"},{"why":"The earlier treatment of intersections of two quadrics that this paper extends.","marker":"[29]"},{"why":"Establishes smoothness and dimension of the Fano scheme of lines on cubic hypersurfaces.","marker":"[1]"},{"why":"Supplies the dimension bound for the relevant locus of lines, used in the finiteness argument.","marker":"[27]"},{"why":"Provides the cubic-surface base case for bounding Eckardt points on a line.","marker":"[15]"},{"why":"Bounds linear-subspace dimensions on smooth hypersurfaces, excluding impossible configurations.","marker":"[7]"}],"fun_headline_variants":["Rational curve moduli bound yields near-linear point counts","Positive characteristic: point counts nearly hit Batyrev–Manin","Del Pezzo surfaces: dimension bounds force near-linear growth","Moduli space dimension predicts point counts on del Pezzo","Batyrev–Manin nearly achieved via rational curve dimensions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on two imported facts that the paper does not verify in the full characteristic range: that every component of the morphism space has dimension at least the expected one, and that on del Pezzo surfaces the only components exceeding it are multiple covers of curves with $-K_X\\cdot C\\le 1$; if either fails over $\\mathbb{F}_2$ or $\\mathbb{F}_3$, the main estimate is not established as stated.","fun_headline_variants_meta":{"raw":{"variants":["Rational curve moduli bound yields near-linear point counts","Positive characteristic: point counts nearly hit Batyrev–Manin","Del Pezzo surfaces: dimension bounds force near-linear growth","Moduli space dimension predicts point counts on del Pezzo","Batyrev–Manin nearly achieved via rational curve dimensions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000236,"raw_usage":{"total_tokens":1485,"prompt_tokens":908,"completion_tokens":577,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":524,"completion_tokens_details":{"reasoning_tokens":504}},"tokens_in":524,"tokens_out":577,"duration_ms":6077,"temperature":1.0,"reasoning_tokens":504,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T22:15:34.717278+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check the hypotheses of the cited theorem in the source: if the classification of non-expected-dimensional components of the del Pezzo morphism space is proved only away from characteristic $2$ or $3$, then Theorem 1.1 is overclaimed for the remaining characteristics. Alternatively, compute $\\operatorname{Mor}(\\mathbb{P}^1,X,e)$ for a smooth del Pezzo surface of degree $4$ or $5$ over an algebraically closed field of characteristic $2$ and look for an irreducible component of dimension strictly above $e+2$ whose general member is not a multiple cover of a line; finding one would break the key input.","supporting_citations":[{"cited_title":"Beheshti, B","cited_arxiv_id":null,"evidence_quote":"Supplies the classification result for the del Pezzo morphism space that makes Theorem 1.1 work."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Justifies the expected-dimension lower bound for components of the morphism space."},{"cited_title":"Harris, M","cited_arxiv_id":null,"evidence_quote":"The bend-and-break lemma, argued here to be characteristic-free, that drives the expected-dimension proofs."},{"cited_title":"Coskun and J","cited_arxiv_id":null,"evidence_quote":"The earlier characteristic-zero treatment of cubic hypersurfaces that this paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The earlier treatment of intersections of two quadrics that this paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes smoothness and dimension of the Fano scheme of lines on cubic hypersurfaces."},{"cited_title":"Huybrechts","cited_arxiv_id":null,"evidence_quote":"Supplies the dimension bound for the relevant locus of lines, used in the finiteness argument."},{"cited_title":"Dolgachev and G","cited_arxiv_id":null,"evidence_quote":"Provides the cubic-surface base case for bounding Eckardt points on a line."},{"cited_title":"Browning and D","cited_arxiv_id":null,"evidence_quote":"Bounds linear-subspace dimensions on smooth hypersurfaces, excluding impossible configurations."}],"review_version":1}