{"id":"27f2eb0e-e1b7-4629-8fe0-3ac2a1c84be5","arxiv_id":"2506.16325","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A smooth projective weak Fano threefold of Picard rank 2 that is either a toric image or admits an int-amplified endomorphism is toric, with three explicit non-Fano exceptions.","lead":"A mathematics paper proves that certain three-dimensional spaces built from toric building blocks must themselves be toric, when they have two independent divisor classes and a mild positivity condition. It also classifies the exceptions, extending earlier results from Fano to weak Fano threefolds.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Corollary 3.2's cited table exclusions are the load-bearing step; if any row in [9], [12], or [13] is missing or mistyped, the Bott non-vanishing test misses a family and both main theorems lose their exclusion step.","rationale":"The reader's weakest_assumption correctly identifies the main load-bearing dependency: the proof of both theorems funnels through Corollary 3.2, which excludes families by substituting Chern numbers into Theorem 3.1(1). Several substitutions are not displayed, and the final step in several cases is literally 'by looking at' a classification table. This is a testable external dependency rather than an internal contradiction. My own reading found no clear arithmetic error in the displayed Bott computations, and the F-liftable/toric-image framework supplies independent evidence that the overall strategy is sound. I also noted the incomplete sentence in Step 2 of the proof of Theorem A and the brevity of Lemma 4.5, but neither is as close to the central claim's exclusion step as the table dependency. Since the concern is already captured by the reader's conditional verdict and does not point to a definite flaw, the verdict should remain UNCHANGED.","tokens_in":14838,"tokens_out":19752,"duration_ms":246495,"concrete_test":"Independently reconstruct, for every row used in Corollary 3.2, the Chern numbers c1^3, h = h^{2,1}, and d where applicable from [9, Table 1], [12, Table A.3], and [13, Tables 7.2, 7.6, 7.7]. Substitute each row into the formulae -chi = 13 + h - c1^3/2 and -chi = h - c1^3/2 + 2d + 3 obtained from Theorem 3.1(1), and verify that every entry gives a negative Euler characteristic (or zero with positive h^0 in the exceptional P^1-bundle case). Also re-derive the two rows of [5, Tables 8 and 9] using the stated relations -K_X = H - E and H = E - 2K_X to check the numerics in Corollary 3.2(4) and (5). If any row fails the inequality, the exclusion argument collapses.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Corollary 3.2 is the bottleneck: it must rule out every non-toric weak Fano threefold of Picard rank 2 before the structural steps of Theorems A and B can close. For the degree-6 del Pezzo fibration case, the paper derives only -chi = 13 + h - c1^3/2 from Theorem 3.1(1) and then says 'the result follows by looking at [9, Table 1]'. For the conic bundle case it derives -chi = h - c1^3/2 + 2d + 3 and then cites [12, Table A.3] and [13, Tables 7.2, 7.6, 7.7] without displaying the substituted Chern numbers. Case (6) says the proof is 'identical' to case (4) and gives no numerics. Whether these Euler characteristics are negative depends entirely on the c1^3 and h values hidden in those tables. If a table row is missing, misidentified, or contains a wrong Chern number, that family survives the Bott test; then the proof has no way to exclude it from the final classification, and Theorem A and Theorem B both fail. The paper's own Bott computations in 3.1(2) and 3.1(3) are explicit and give independent support, so the concern is not internal inconsistency but an unverified database dependency.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies weak Fano threefolds of Picard rank two. The main results are Theorem A: every smooth projective threefold with rho(X)=2 and -K_X nef that is the image of a projective toric variety is either toric Fano or one of P_{P1}(O_{P1} oplus O_{P1}(1)^2), P_{P1}(O^2_{P1} oplus O_{P1}(2)), P_{P2}(O_{P2} oplus O_{P2}(3)); and Theorem B: the same conclusion holds if the variety admits an int-amplified endomorphism. The proof works by deriving explicit formulas for certain Euler characteristics in Theorem 3.1, using them in Corollary 3.2 to show that all non-toric candidates in the available classification tables fail Bott vanishing, and then running a Mori-theoretic analysis that reduces the possibilities to the listed cases.","tokens_in":15069,"tokens_out":26921,"duration_ms":291395,"significance":"If correct, these results establish two new special cases of the Occhetta-Wisniewski image-of-toric conjecture and the Fakhruddin-Meng-Zhang-Zhong amplified-endomorphism conjecture, extending earlier work from Fano threefolds to weak Fano threefolds of Picard rank two. The Chern-class computations in Theorem 3.1 are explicit and checkable, and the final classification is concrete and falsifiable. The main caveat is that several exclusions are delegated to classification tables whose entries are not displayed; the proof is therefore only as strong as those tables and the reader's ability to verify the substitutions.","major_comments":[{"comment":"The proofs repeatedly invoke 'X does not have Bott vanishing, a contradiction' (e.g., Claim 3 in Step 1 of Theorem A and the conic-bundle case in Step 2), but nowhere in the proof of Theorem A is it stated or justified that a toric image has Bott vanishing. The connection is only made in Remark 5.5 via [1, Theorem 4.4.1] and [1, Theorem 3.2.4]. Similarly, the proof of Theorem B uses the same contradiction without stating the implication 'int-amplified endomorphism implies Bott vanishing', presumably from [15]. These are load-bearing premises; without them the contradictions are non-sequiturs. Please add explicit statements and references at the start of Sections 4 and 5.","section":"Sections 4 and 5 (proofs of Theorems A and B)"},{"comment":"The Bott non-vanishing test is the only mechanism that excludes the non-toric weak Fano threefolds of Picard rank two in Theorems A and B. However, in case (1) the string 'the result follows by looking at [9, Table 1]' replaces the numerical check for -chi = 13 + h - c1^3/2; in the conic-bundle part of case (3) the check for -chi = h - c1^3/2 + 2d + 3 is delegated to [12, Table A.3] and [13, Tables 7.2, 7.6, 7.7]; and case (6) is dismissed as 'identical' to case (4) with no numbers. If any table row is omitted, misidentified, or contains a wrong Chern number, the corresponding family survives the Bott test and both main theorems lose an exclusion step. Please display, for every cited row, the values of c1^3, h, d and the resulting sign of the Euler characteristic (or the required H^0 computation).","section":"Corollary 3.2, cases (1), (3), (6)"},{"comment":"The proof of Claim 1 says only: 'Replacing f by a power of f, by Lemmas 5.3 and 5.2, we have f^{-1}(D_red)=D, f^{-1}(E_red)=E.' Lemma 5.3 assumes f^{-1}D=D and f^{-1}C=C for all flopping curves, which is exactly the preservation property being asserted; Lemma 5.2 concerns the induced map on the target of a birational contraction and does not establish preservation of both E and D in X. As written, the step appears circular. The author should either prove the invariance directly from the int-amplified condition or cite a precise lemma (with the hypotheses verified) that guarantees that a power of an int-amplified endomorphism preserves these two boundary divisors. This is load-bearing because Claim 1 is used to obtain normal crossing in codimension two via Lemma 5.1 in Claim 2.","section":"Theorem B, Step 1, Claim 1"},{"comment":"Theorem 3.1(3) is applied under the hypothesis that the rank-two bundle E on P^2 satisfies E(1) ample. In the application, however, the bundles are only described as E := F(2) (in the first three cases) or E := F or F^+ (in the other three cases), with the statement that they are 'nef but not ample'. Nefness of E does not imply ampleness of E(1), and no check is provided for each of the six cases. Without verifying E(1) is ample, the formulas of Theorem 3.1(3) cannot be used. Please add the missing ampleness verification or replace the reference with a version of the theorem that does not require it.","section":"Corollary 3.2(3), Case 1"},{"comment":"The transition from the geometric hypotheses to 'X is as in [12, No. 1, Table A.5]' is another unshown classification lookup. The proof does not display the numerical invariants that identify this unique row, and it is not clear which hypotheses rule out the other rows of the table. Since Corollary 3.2(6) then applies to that row, a mistaken identification would again break both proofs. Please either display the relevant invariants or add a sentence explaining the comparison.","section":"Theorem A, Step 1, Claim 3; Theorem B, Step 1, Claim 3"}],"minor_comments":[{"comment":"Lemma 5.3 states that the induced map is 'f+ : X -> X', but it should be 'f+ : X+ -> X+'; in the proof of Theorem B, 'by Lemma 5.3 X' has int-amplified endomorphism' should clarify whether X' or X+ is meant, since both appear in the flop diagram.","section":"Lemma 5.3 and proof of Theorem B"},{"comment":"The sentence 'If -K_X is not spanned, by [12, Corollary 1.5].' is incomplete; it should state the conclusion obtained from [12, Corollary 1.5].","section":"Theorem A, Step 2"},{"comment":"The phrase 'Let and B = psi(D)' contains a typo: it should be 'Let B = psi(D)'.","section":"Theorem B, Step 1, Claim 3"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is an early arXiv version and is honest about the competing preprint [4]. The main concerns for a journal are not the core Chern-class computations, which are explicit, but the unverified database dependence and a few missing justifications; these should be fixable with a verification table and additional citations. The reliance on the author's own preprint [14] for Lemma 4.7 should also be checked against the eventual published version of [14]."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know before you read it. The genuinely new material is Theorem 3.1: explicit formulas for -chi(X, Omega_X^2(L)) for three classes of weak Fano threefolds, derived by Hirzebruch-Riemann-Roch and toric Euler-Jaczewski computations. These look solid and are reusable. The two main theorems, A and B, are special cases of Occhetta-Wisniewski and Fakhruddin-Meng-Zhang-Zhong, and the strategy is a credible extension of Totaro's Bott-vanishing program. But the proof is not self-contained at the point where it matters: Corollary 3.2 rules out many families by citing classification tables without displaying the substituted Chern numbers. That is the load-bearing step. If any row in Fukuoka's table, Jahnke-Peternell-Radloff's tables, or Cutrone-Marshburn's tables is missing or mistyped, the exclusion fails and both theorems lose their conclusion.\n\nWhat it does well: Theorem 3.1's computations are detailed; in 3.1(3) when chi = 0 it checks h^0 separately rather than assuming vanishing; the author discloses that [4] independently proves the endomorphism statement, so Theorem B's novelty is honestly bounded. The argument in Section 4 is coherent: reduce to fibrations, exclude bad families by Bott non-vanishing, then identify the survivors from tables.\n\nSoft spots: the stress-test note is on target. In Corollary 3.2(1) the proof ends with 'the result follows by looking at [9, Table 1]'; in (3) the same for [12] and [13]; (6) says 'identical' to (4) with no numerics. A referee will need to check those pages and either reproduce the substitutions or have the author include them. Minor but real: Step 2 of Theorem A has an incomplete sentence, 'If -K_X is not spanned, by [12, Corollary 1.5],' and the implication is lost. Lemma 4.7 is taken from the author's coauthored preprint [14]; it is cited, not hidden, but it is a nontrivial dependency.\n\nSummary: this paper is for people working on toric images, Bott vanishing, and endomorphisms of threefolds. The central approach seems sound and the new computations deserve to be in the literature. I would send it to a serious referee and ask for the Corollary 3.2 numerics to be shown and the typos fixed. If the table checks pass, it should be accepted; if a row is wrong, the exclusion step needs repair.","headline":"The Bott-vanishing formulas in Theorem 3.1 are the real contribution; the classification theorems are credible but rest on unshown table numerics in Corollary 3.2, so the paper needs referee checks rather than desk reject.","tokens_in":15648,"tokens_out":4898,"would_cite":true,"duration_ms":52415,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14M25","14E20"],"pacs":[],"model":"deepseek-v4-flash","headline":"A smooth projective threefold of Picard rank 2 with nef anticanonical bundle that is the image of a projective toric variety is itself toric, with exactly three non-Fano exceptions.","keywords":["toric image","weak Fano threefold","Picard rank 2","Bott vanishing","int-amplified endomorphism","toric variety","del Pezzo fibration","conic bundle"],"falsifier":"A concrete way to disprove the central claim would be to exhibit a smooth projective weak Fano threefold $X$ with $\\rho(X)=2$, not Fano and not one of the three listed projective bundles, for which Bott vanishing holds, or equivalently for which the formula of Theorem 3.1(1) gives a nonnegative value after substituting the true Chern numbers. If any entry of the cited classification tables has a wrong Chern number, re-computation would reveal it as a surviving case.","tokens_in":14560,"feed_emoji":"📐","tokens_out":14688,"duration_ms":146762,"temperature":0.7,"pith_summary":"This paper proves a dichotomy for smooth projective threefolds of Picard rank 2 whose anticanonical bundle is nef and big (the weak Fano condition): if such a variety is the image of a projective toric variety, or if it carries an int-amplified endomorphism, then it is toric, with the same three explicit exceptions in both cases. An int-amplified endomorphism is a finite surjective self-map that pulls some ample divisor back to a divisor whose difference from the original is again ample. The proof eliminates every other family in the known classification by showing it fails Bott vanishing, a cohomology vanishing property that toric images and varieties with such endomorphisms must satisfy. These results are special cases of two open conjectures previously settled for surfaces and for Fano threefolds. If correct, the paper gives the full list of rank-2 toric weak Fano threefolds that are not Fano.","feed_headline":"Toric images of weak Fano threefolds are toric, with 3 exceptions","feed_subtitle":"A Bott-vanishing argument pins down the only rank-2 weak Fano threefolds that could be non-toric toric images.","key_machinery":"The workhorse is Bott vanishing, the cohomology vanishing property that toric varieties and their images satisfy and that varieties with an int-amplified endomorphism inherit. In the form used here, it forces $\\chi(X,\\Omega_X^2\\otimes L)\\ge 0$ for every ample line bundle $L$. The paper proves three Hirzebruch–Riemann–Roch formulas (Theorem 3.1) that turn Bott vanishing into a numerical test for rank-$2$ weak Fano threefolds, for example $$-\\chi(X,\\$Omega_X^{2}$(H-K_X))=16+h-\\frac{$c_1^{3}$}{2}-\\frac54($c_1^{2}$H+$c_1H^{2}$)+\\frac34c_2H-\\$frac12H^{3}$.$$ Substituting the Chern numbers of each family from the classification tables into these formulas, Corollary 3.2 shows the test excludes every family except the three listed bundles.","core_discovery":"The paper establishes that the only smooth projective weak Fano threefolds of Picard rank $2$ that can be images of projective toric varieties, and the only ones that can carry an int-amplified endomorphism, are the same three varieties: the projective bundles $\\mathbb{P}_{\\mathbb{P}^1}(\\mathcal{O}_{\\mathbb{P}^1}\\oplus\\mathcal{O}_{\\mathbb{P}^1}(1)^2)$, $\\mathbb{P}_{\\mathbb{P}^1}(\\mathcal{O}_{\\mathbb{P}^1}^2\\oplus\\mathcal{O}_{\\mathbb{P}^1}(2))$, and $\\mathbb{P}_{\\mathbb{P}^2}(\\mathcal{O}_{\\mathbb{P}^2}\\oplus\\mathcal{O}_{\\mathbb{P}^2}(3))$. Every other rank-$2$ weak Fano threefold fails Bott vanishing, an obstruction that toric images and varieties with an int-amplified endomorphism cannot have.","pith_inferences":["The same table-driven Bott test could be repeated for weak Fano threefolds of higher Picard rank once complete numerical classification lists exist, potentially proving both conjectures beyond rank 2.","A reader could independently verify the paper's exclusions by recomputing the Chern numbers of each listed family from its defining fibration and substituting them into Theorem 3.1(1), which would also check the internal consistency of the cited tables.","The three surviving exceptions are all projective bundles over $\\mathbb{P}^1$ or $\\mathbb{P}^2$; in higher rank one might expect any surviving non-Fano toric image to be a toric projective bundle as well."],"forward_implications":["The three varieties $\\mathbb{P}_{\\mathbb{P}^1}(\\mathcal{O}_{\\mathbb{P}^1}\\oplus\\mathcal{O}_{\\mathbb{P}^1}(1)^2)$, $\\mathbb{P}_{\\mathbb{P}^1}(\\mathcal{O}_{\\mathbb{P}^1}^2\\oplus\\mathcal{O}_{\\mathbb{P}^1}(2))$, and $\\mathbb{P}_{\\mathbb{P}^2}(\\mathcal{O}_{\\mathbb{P}^2}\\oplus\\mathcal{O}_{\\mathbb{P}^2}(3))$ are the complete classification of rank-2 toric weak Fano threefolds that are not Fano.","Any rank-2 weak Fano threefold with an int-amplified endomorphism is one of those three bundles, hence toric, settling the amplified-endomorphism conjecture in this class.","The families covered by Corollary 3.2—degree 6 and degree 8 del Pezzo fibrations over $\\mathbb{P}^1$, most conic bundles over $\\mathbb{P}^2$, and the birational cases of parts (4)–(6)—are all shown not to satisfy Bott vanishing.","A Frobenius-liftable weak Fano threefold of Picard rank 2 is toric, a generalization stated in Remark 5.5 that follows by the same Bott-vanishing argument."],"supporting_citations":[{"why":"Supplies Tables 8 and 9 of weak Fano threefolds with rho=2 used in Corollary 3.2(4)-(5) to rule out the birational cases.","marker":"[5]"},{"why":"Supplies the refined list of degree-6 del Pezzo fibrations over P1 used in Corollary 3.2(1).","marker":"[9]"},{"why":"Supplies Tables 7.1 and A.3 of weak Fano threefolds used to identify the degree-8 del Pezzo fibration cases and the P1-bundle cases in Corollary 3.2 and in the main proof.","marker":"[12]"},{"why":"Supplies Tables 7.2, 7.5, 7.6, 7.7 and the No. 1 Table 7.7 case used to rule out conic bundles in Corollary 3.2(3) and in Theorem A's Step 2.","marker":"[13]"},{"why":"Develops the Bott-vanishing machinery for endomorphisms on which the paper's exclusion strategy rests.","marker":"[15]"},{"why":"Establishes the rho=1 toric image case and the Euler-Jaczewski sequence used to compute chi in Theorem 3.1(2).","marker":"[24]"},{"why":"Classifies weak Fano threefolds with del Pezzo fibration, giving the k values used in Corollary 3.2(2).","marker":"[27]"},{"why":"Proves Bott vanishing for Fano 3-folds and provides the technique this paper extends to weak Fano threefolds.","marker":"[28]"},{"why":"Provides log Bott vanishing and the lemma that an int-amplified endomorphism descends after a birational contraction, used in Lemma 5.2.","marker":"[29]"},{"why":"Classifies extremal divisorial contractions on smooth threefolds, identifying the possible exceptional divisors in the Claims 2 of both theorems.","marker":"[17]"}],"fun_headline_variants":["Weak Fano 3-folds: toric images are toric, except 3","Three exceptions: rank-2 weak Fano toric images","Toric image or amplified endomorphism forces toric","Bott vanishing reveals only three non-toric candidates","Rank-2 weak Fano threefolds: toric images classified"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument depends on the published lists of all weak Fano threefolds of Picard rank 2 being complete and numerically accurate; if a list misses a family or contains a wrong Chern number, an excluded case could survive the Bott-vanishing test and the conclusions would fail.","fun_headline_variants_meta":{"raw":{"variants":["Weak Fano 3-folds: toric images are toric, except 3","Three exceptions: rank-2 weak Fano toric images","Toric image or amplified endomorphism forces toric","Bott vanishing reveals only three non-toric candidates","Rank-2 weak Fano threefolds: toric images classified"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00019,"raw_usage":{"total_tokens":1304,"prompt_tokens":874,"completion_tokens":430,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":490,"completion_tokens_details":{"reasoning_tokens":340}},"tokens_in":490,"tokens_out":430,"duration_ms":4444,"temperature":1.0,"reasoning_tokens":340,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T23:44:40.221896+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete way to disprove the central claim would be to exhibit a smooth projective weak Fano threefold $X$ with $\\rho(X)=2$, not Fano and not one of the three listed projective bundles, for which Bott vanishing holds, or equivalently for which the formula of Theorem 3.1(1) gives a nonnegative value after substituting the true Chern numbers. If any entry of the cited classification tables has a wrong Chern number, re-computation would reveal it as a surviving case.","supporting_citations":[{"cited_title":"Towards the classification of weak Fano threefolds with ρ= 2","cited_arxiv_id":null,"evidence_quote":"Supplies Tables 8 and 9 of weak Fano threefolds with rho=2 used in Corollary 3.2(4)-(5) to rule out the birational cases."},{"cited_title":"Refinement of the classification of weak Fano threefolds with sextic del Pezzo fibrations","cited_arxiv_id":"1903.06872","evidence_quote":"Supplies the refined list of degree-6 del Pezzo fibrations over P1 used in Corollary 3.2(1)."},{"cited_title":"Threefolds with big and nef anticanonical bundles I","cited_arxiv_id":"math/0407484","evidence_quote":"Supplies Tables 7.1 and A.3 of weak Fano threefolds used to identify the degree-8 del Pezzo fibration cases and the P1-bundle cases in Corollary 3.2 and in the main proof."},{"cited_title":"Threefolds with big and nef anti- canonical bundles II","cited_arxiv_id":null,"evidence_quote":"Supplies Tables 7.2, 7.5, 7.6, 7.7 and the No. 1 Table 7.7 case used to rule out conic bundles in Corollary 3.2(3) and in Theorem A's Step 2."},{"cited_title":"Endomorphisms of varieties and Bott vanishing","cited_arxiv_id":"2302.11921","evidence_quote":"Develops the Bott-vanishing machinery for endomorphisms on which the paper's exclusion strategy rests."},{"cited_title":"On Euler–Jaczewski sequence and Remmert–van de Ven problem for toric varieties","cited_arxiv_id":null,"evidence_quote":"Establishes the rho=1 toric image case and the Euler-Jaczewski sequence used to compute chi in Theorem 3.1(2)."},{"cited_title":"Weak Fano threefolds with del Pezzo fibration","cited_arxiv_id":null,"evidence_quote":"Classifies weak Fano threefolds with del Pezzo fibration, giving the k values used in Corollary 3.2(2)."},{"cited_title":"Bott vanishing for Fano 3-folds","cited_arxiv_id":"2302.08142","evidence_quote":"Proves Bott vanishing for Fano 3-folds and provides the technique this paper extends to weak Fano threefolds."},{"cited_title":"Endomorphisms of Fano 3-folds and log Bott vanishing","cited_arxiv_id":"2305.18660","evidence_quote":"Provides log Bott vanishing and the lemma that an int-amplified endomorphism descends after a birational contraction, used in Lemma 5.2."},{"cited_title":"Extremal rays on smooth threefolds","cited_arxiv_id":null,"evidence_quote":"Classifies extremal divisorial contractions on smooth threefolds, identifying the possible exceptional divisors in the Claims 2 of both theorems."}],"review_version":1}