{"id":"f4172cf3-38c7-4661-8d02-369eafc87527","arxiv_id":"2608.12076","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Products of fixed Fano varieties of Picard number one are classified by the partition of the total dimension.","lead":"This paper proves that products built from a fixed Fano variety in each dimension remain non-isomorphic whenever the factor dimensions form different partitions. The proof uses extremal contractions of the cone of curves and yields a classification of products of smooth quadrics of dimension at least three.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.13's final step is a gap: it compares the chosen projections rather than an intrinsic invariant, and never proves that these projections exhaust the relevant extremal contractions; the simplicial cones themselves are isomorphic for all such products.","rationale":"The central Mori-theoretic computation is sound: for a Fano variety with rho=1, N_1 has dimension 1 and an ample class puts all effective curve classes on one ray, so Lemma 3.11 is essentially trivial and not a real risk. The Picard splitting and nef-cone formula are also standard. My concern is the final inference. An isomorphism of varieties induces an isomorphism of Mori cones, but the paper's argument treats the selected projections as if they were known to be all the contractions that matter. In fact, any two products with the same number of factors have abstractly isomorphic simplicial cones, so the cone alone cannot encode the partition. The missing invariant is the multiset of image dimensions of contractions of codimension-one faces. This is easily supplied from Proposition 3.12, and I expect the theorem is true, but the proof as printed is incomplete. Hence a conditional acceptance, not a rejection.","tokens_in":8875,"tokens_out":20103,"duration_ms":206265,"concrete_test":"Add a lemma: for X=F_1^{m_1}×...×F_1^{m_r}, the codimension-one faces of NE(X) are exactly the F_{pr_i^*H_i}, and the contraction of the face spanned by all rays except the i-th has image of dimension m_i. Then verify the invariant on the 4-dimensional examples P^1×P^3 and P^2×P^2: compute the two simplicial cones, their codimension-one faces, and the image dimensions of the corresponding contractions; the multisets must be {1,3} and {2,2}. If this holds, the main theorem follows after replacing the comparison of unnamed contraction sets with this multiset; if it fails, the theorem itself needs re-examination.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The classification in Theorems 3.5 and 3.13 rests on Proposition 3.12, which describes NE(X) as the simplicial cone generated by r rays and shows that the faces F_{L_i} have the projections as their contractions. But in the final paragraph the proof only compares the sets {pr_{m_i}} and {pr_{n_j}}. These sets are not shown to be isomorphism invariants: the faces F_{L_i} are defined relative to the chosen product decomposition, and an isomorphism could in principle send one of them to another extremal face of the target. Abstractly, the cone of curves of every product of r such factors is the same r-dimensional simplicial cone, so the statement 'the cone of curves themselves are different' cannot distinguish P^1×P^3 from P^2×P^2. What is needed is a proof that the codimension-one faces of NE(X) are exactly the F_{L_i}, that the contraction of the face omitting the i-th ray has image dimension m_i, and that this multiset of dimensions is preserved by isomorphism. Proposition 3.12 provides the ingredients but the paper never states this invariant. This is a repairable gap rather than a counterexample, but as written it leaves the classification unproved.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper gives a new proof that distinct partitions of n yield non-isomorphic multiprojective spaces P^{n_1} × ... × P^{n_r}, and extends this to products of Fano varieties of Picard number one. After establishing via Kodaira vanishing that Pic(X × Y) splits for Fano factors, the author computes the nef cone and the closed cone of curves of such products as simplicial cones generated by pullbacks of ample classes and by curves in the factors. The projections to the factors are then identified with the extremal contractions of the codimension-one faces of the Mori cone. The main classification theorems are Theorem 3.5 for multiprojective spaces and Theorem 3.13 for general Fano factors, with Corollary 3.14 specializing to products of smooth quadrics of dimension at least 3.","tokens_in":9107,"tokens_out":11988,"duration_ms":129180,"significance":"If fully proved, the paper provides a uniform, MMP-based classification that improves on earlier case-by-case arguments. The technical ingredients are standard and mostly well handled: the Picard-group splitting via H^1(X,O_X)=0, the Cone Theorem for the Mori cone of a Fano factor, and the construction of the projections as extremal contractions are all sound. The applications to quadrics, rational homogeneous spaces, and fixed-determinant moduli spaces make the result attractive. However, as written, the final inference in Theorems 3.5 and 3.13 compares sets of contractions that are not shown to be isomorphism invariants, and the proof needs a local but load-bearing repair before the classification is logically complete.","major_comments":[{"comment":"The proof compares the sets {pr_{m_i}} and {pr_{n_j}} and then concludes that the corresponding faces and cones of curves are different, hence the products are non-isomorphic. This is not justified as written. The contractions pr_{m_i} are defined relative to a chosen product decomposition, and an isomorphism between two products need not send the face F_{L_i} to the analogous face of the target; indeed, for any two r-factor products covered by Proposition 3.12, the closed cones of curves are abstractly the same r-dimensional simplicial cone. The missing invariant is the multiset of dimensions of the images of the contractions of the codimension-one faces of NE(X). Proposition 3.12 supplies all the needed data: the cone has exactly r codimension-one faces F_{L_i}, and the contraction of F_{L_i} has image of dimension d_i. The proof should state and use this invariant explicitly. Without it, the classification is not established, although the statement is true and the gap is local.","section":"§3, Theorem 3.13 (and Theorem 3.5), final paragraph"}],"minor_comments":[{"comment":"The projections from a product to a factor are called 'birational projective morphisms'; this is not generally true (the projection P^1 × P^1 → P^1 is not birational). The equality (pr_{n_i})_* O_X = O_{P^{n_i}} follows from properness and connectedness of the fibers via Stein factorization/Zariski's Main Theorem, so the wording should be corrected rather than the argument changed.","section":"§3, Proposition 3.4 and Proposition 3.12"},{"comment":"The phrase 'finite rational polytope' should be 'rational polyhedral cone', and the object under discussion is the closed cone of curves \\overline{NE}(X), not the non-closed NE(X). The cited Cone Theorem statement should be made precise, since the finiteness conclusion for Fano varieties is a standard but nontrivial consequence.","section":"§3, Lemma 3.11"},{"comment":"There are several typographical slips, including 'F ANO V ARIETIES' in the running title and 'the factor P is are rational' in Section 1; these should be corrected in the final version.","section":"Abstract and Section 1"},{"comment":"The name 'Lazersfeld' in the reference list should be 'Lazarsfeld'.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the main results are true and the gap in the final step of Theorems 3.5 and 3.13 is local and readily repairable by adding the standard invariant described in the report. I do not see grounds for rejection, but the current proof, as written, would not convince a careful reader. The paper is within the scope of math.AG and the citation of the author's earlier work is appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper gives a uniform Mori-theoretic proof that products of Fano varieties of Picard number one, with a fixed factor in each dimension, are classified by partitions. The multiprojective case was already known, but the Fano generalization, and the quadric corollary, are new. The setup is clean: Picard group splitting via Kodaira vanishing, the simplicial cone description, and the projection contractions all follow from standard references. I think the result is very likely true, and the paper is worth engaging with.\n\nThe soft spot is in the final inference of Theorems 3.5 and 3.13. The proof compares the sets {pr_{m_i}} and {pr_{n_j}} of projections attached to the chosen product decompositions. Those sets are not isomorphism invariants as stated. An isomorphism can move extremal faces around, and the cone of curves of any r-fold product of such factors is abstractly the same simplicial cone. What needs to be said is that the codimension-one faces of NE(X) are exactly the F_{L_i}, that the contraction of the face omitting the i-th ray has image dimension m_i, and that this multiset of dimensions is preserved under isomorphism. Proposition 3.12 contains all the ingredients; the paper just never states the invariant. Without it, the classification as written is unproved. This is a repairable gap, not a counterexample, and a referee should catch it and ask for the two added sentences.\n\nThe citation pattern is fine. Prior papers [11] and [12] are cited as earlier solutions of Question 1.1, not used as inputs. No code or data apply. The paper is written for people working on Fano classification or Mori theory; for them it is a useful structural result, though not a breakthrough.\n\nMy recommendation: send it to peer review. The gap is genuine but minor, and the main theorem is likely correct. With a small revision spelling out the dimension-of-image invariant, this would be a solid accepted paper.","headline":"A likely-correct Fano-product classification with a repairable gap in the final step: the proof compares chosen projections instead of an intrinsic invariant.","tokens_in":9670,"tokens_out":3469,"would_cite":true,"duration_ms":34589,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14E30","14J45","14M22"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that for every dimension $d$, once a smooth Fano variety of Picard number one is fixed, the products built from distinct partitions of $n$ are pairwise non-isomorphic; the same cone-of-curves argument classifies…","keywords":["Multiprojective space","Fano variety","Picard number one","cone of curves","extremal contraction","quadric hypersurface","partition classification"],"falsifier":"Compute $\\overline{NE}(Q^3\\times Q^3)$ directly: if the only extremal rays are the two curve classes coming from the factors and the only connected-fiber contractions are the two projections, the product-cone picture is confirmed; the appearance of an additional extremal contraction would force a refinement of the classification. Equivalently, exhibiting a smooth Fano variety of Picard number one for which Lemma 3.11 fails—a Mori cone not generated by a single rational curve—would collapse the proof of Theorem 3.13.","tokens_in":8639,"feed_emoji":"🧩","tokens_out":13398,"duration_ms":114073,"temperature":0.7,"pith_summary":"The paper establishes that the isomorphism type of a product of projective spaces—and more generally of a product of Fano varieties of Picard number one—is determined by the partition of the total dimension into the dimensions of the factors, provided one fixes one Fano variety in each dimension. For multiprojective spaces this gives a uniform answer to a classical question, avoiding the earlier case split by partition length. The proof reads the factors off from the extremal contractions of the closed cone of curves: each factor contributes one extremal ray, and each projection is the unique contraction of that ray, so distinct partitions yield different sets of contractions and therefore non-isomorphic varieties. A corollary classifies products of smooth quadrics of dimension at least three by partitions. The argument combines the splitting of the Picard group of Fano varieties, the Cone Theorem, and the existence of extremal contractions for $K_X$-negative faces.","feed_headline":"Products of Fano varieties are classified by partitions","feed_subtitle":"A cone-of-curves proof shows distinct partitions of n yield non-isomorphic products, including quadrics.","key_machinery":"The load-bearing object is the closed cone of curves $\\overline{NE}(X)$ together with the extremal contractions of its faces. For a product $F_1^{d_1}\\times\\cdots\\times F_1^{d_r}$ of Fano varieties of Picard number one, the paper first proves (Proposition 3.10) that the Picard group splits as $\\mathbb{Z}^{\\oplus r}$, using Kodaira vanishing to get $H^1(X,\\mathcal{O}_X)=0$. Lemma 3.11 then gives that each factor's cone of curves is a single ray $\\mathbb{R}_{\\ge0}[\\ell_i]$ spanned by a rational curve, so the product's nef cone is the positive orthant generated by the pullbacks of the ample generators and its cone of curves is the simplicial cone generated by the curves $\\tilde\\ell_i$ inside the factors. The projections are exactly the extremal contractions of the faces $F_{L_i}=\\sum_{j\\neq i}\\mathbb{R}_{\\ge0}[\\tilde\\ell_j]$, and since an extremal contraction is uniquely determined by its face, the set of projections is an isomorphism invariant; different partitions give different sets of contractions.","core_discovery":"The central discovery is Theorem 3.13: fix, for each dimension $d$, a smooth Fano variety $F_1^d$ of Picard number one; then for any two distinct partitions $(m_1,\\ldots,m_r)$ and $(n_1,\\ldots,n_s)$ of a positive integer $n$, the products $F_1^{m_1}\\times\\cdots\\times F_1^{m_r}$ and $F_1^{n_1}\\times\\cdots\\times F_1^{n_s}$ are non-isomorphic. With $F_1^d=\\mathbb{P}^d$ this recovers the classification of multiprojective spaces (Theorem 3.5), and with $F_1^d=Q^d$, the smooth quadric hypersurface of dimension $d\\ge 3$, it classifies products of quadrics (Corollary 3.14). The proof shows that the closed cone of curves of such a product is simplicial, with one extremal ray coming from a rational curve inside each factor, and that the projections are precisely the extremal contractions of the faces spanned by all but one ray; hence the partition is recoverable from the birational geometry of the product.","pith_inferences":["Our inference: the mechanism should work for any class of projective varieties whose Picard group splits across products, whose nef cone is simplicial with one extremal ray per factor, and whose $K_X$-negative faces admit contractions; the Fano hypothesis is one sufficient route, not necessarily the only one.","Our inference: because the proof recovers the set of projection contractions from the product, it suggests the product actually remembers its factors as the targets of its contractions, so a fuller classification could recover each $F_1^{m_i}$ up to isomorphism, not just the partition.","Our inference: the quadric classification stops at dimension $3$ because lower-dimensional quadrics are not of Picard number one; a natural extension would treat $\\mathbb{P}^1\\times\\mathbb{P}^1$ as an additional fixed factor and ask whether mixed products with higher-dimensional quadrics are still classified by partitions augmented by the number of $\\mathbb{P}^1\\times\\mathbb{P}^1$ factors."],"forward_implications":["The classical classification of multiprojective spaces $\\mathbb{P}^{m_1}\\times\\cdots\\times\\mathbb{P}^{m_r}$ by partitions of $n$ follows from a single uniform argument, without separating partitions of equal length from partitions of different lengths.","For any fixed choice of a smooth Fano variety $F_1^d$ of Picard number one in each dimension $d$, products attached to distinct partitions of $n$ are pairwise non-isomorphic.","Products of smooth quadrics of dimension at least $3$ are classified by partitions of the total dimension into parts of size at least $3$.","The theorem applies verbatim to products of Grassmannians and of fixed-determinant moduli spaces of stable bundles over a curve, because these are Fano varieties of Picard number one."],"supporting_citations":[{"why":"Supplies the earlier case-split solution of the multiprojective-space classification that this paper's uniform cone argument replaces and extends.","marker":"[12]"},{"why":"Gives an earlier representation-theoretic proof of the same multiprojective classification, offered as a contrast to the geometric method.","marker":"[11]"},{"why":"Provides the Cone Theorem and the duality between the nef cone and the closed cone of curves used in Lemma 3.11 and Proposition 3.12, plus the Lefschetz hyperplane theorem used for quadrics.","marker":"[9]"},{"why":"Is the source of the existence criterion for extremal contractions of $K_X$-negative faces, quoted as Theorem 2.7.","marker":"[10]"},{"why":"Gives uniqueness of an extremal contraction with a given face, the property that turns different partitions into different contraction sets.","marker":"[7]"},{"why":"Provides the Picard-group splitting for products with vanishing $H^1(X,\\mathcal{O}_X)$ and the Zariski Main Theorem used to identify the projections as contractions.","marker":"[5]"}],"fun_headline_variants":["Partitions pin down Fano product types","Fano products: partition determines isomorphism","Cone-of-curves proof classifies Fano products","Fano product isomorphism forces equal partitions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof rests on Lemma 3.11, which asserts—via the Cone Theorem—that the closed cone of curves of a smooth Fano variety of Picard number one is a single ray spanned by a rational curve; if that ray structure failed, the product cone would no longer have one extremal ray per factor and the set of projection contractions would not be pinned down by the partition.","fun_headline_variants_meta":{"raw":{"variants":["Partitions pin down Fano product types","Fano products: partition determines isomorphism","Cone-of-curves proof classifies Fano products","Fano product isomorphism forces equal partitions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000394,"raw_usage":{"total_tokens":2069,"prompt_tokens":944,"completion_tokens":1125,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":560,"completion_tokens_details":{"reasoning_tokens":1069}},"tokens_in":560,"tokens_out":1125,"duration_ms":12111,"temperature":1.0,"reasoning_tokens":1069,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:18:41.221106+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $\\overline{NE}(Q^3\\times Q^3)$ directly: if the only extremal rays are the two curve classes coming from the factors and the only connected-fiber contractions are the two projections, the product-cone picture is confirmed; the appearance of an additional extremal contraction would force a refinement of the classification. Equivalently, exhibiting a smooth Fano variety of Picard number one for which Lemma 3.11 fails—a Mori cone not generated by a single rational curve—would collapse the proof of Theorem 3.13.","supporting_citations":[{"cited_title":"A representation theoretic classification of multiprojective spaces","cited_arxiv_id":"2405.16198","evidence_quote":"Gives an earlier representation-theoretic proof of the same multiprojective classification, offered as a contrast to the geometric method."},{"cited_title":"and Mori, S","cited_arxiv_id":null,"evidence_quote":"Gives uniqueness of an extremal contraction with a given face, the property that turns different partitions into different contraction sets."}],"review_version":1}