{"id":"1c1f2b58-c451-4b96-8423-b96808aad268","arxiv_id":"1908.08283","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Certain blow-ups of projective spaces, including the plane in up to nine points, satisfy Orlov's conjecture that the Rouquier dimension of the derived category equals the dimension.","lead":"This mathematics paper proves new cases of a conjecture about the minimal number of objects needed to generate the derived category of coherent sheaves on a variety. The new cases are certain blow-ups of projective spaces, including a plane blown up in up to nine points.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the claimed flaw in Lemma 2.4 does not land, and the main argument is internally consistent.","rationale":"The paper's central claim is Theorem 4.1, and the sole substantive objection raised by the Reader is that Lemma 2.4's proof relies on a false implication. I checked that implication carefully. In the derived category of a hereditary abelian category, every bounded complex is formal: the truncation triangle relating cohomology objects splits because the relevant obstruction lives in Ext^2, which vanishes. Thus an ADE quiver category has only finitely many indecomposable objects up to shift (the indecomposable representations and their shifts), so Lemma 2.3 applies and Rouquier dimension is zero. The reader's example of non-split two-term complexes in hereditary categories concerns short exact sequences in the abelian category, not non-formal objects in the derived category; an extension 0->A->E->B->0 is just the object E, which is already a representation. The proof of Lemma 2.4 is terse and would benefit from a citation or a one-line proof of formality, but it is not mathematically wrong. I also checked the surrounding construction: Proposition 3.1 correctly uses Orlov's semiorthogonal decomposition to obtain the one-sided morphism spaces RHom(L, S_b)=k[0], and the reverse morphisms vanish by the orthogonality of the blow-up decomposition. The subsequent mutations and the tower argument in Proposition 4.2 preserve the number of zero-dimensional components as claimed. Because the only identified concern is not valid, no adjustment to the Reader's conditional verdict is needed.","tokens_in":9419,"tokens_out":51208,"duration_ms":521222,"concrete_test":"Verify Lemma 2.4 directly: prove or cite the standard fact that in a hereditary abelian category every bounded complex is quasi-isomorphic to the direct sum of its cohomology objects, e.g., for a two-term complex A->B the class of 0->ker d->A->B->coker d->0 in Ext^2(coker d, ker d) vanishes. If this fact is accepted, the remaining argument needs no further condition.","verdict_should_be":"UNCHANGED","load_bearing_attack":"No significant objection identified. The reader's load-bearing concern is that Lemma 2.4 is invalid because homological dimension one does not imply that every object of D^b(Q-rep) is a direct sum of shifts of representations. This objection does not land: for a hereditary abelian category, in particular finite-dimensional modules over an acyclic ADE quiver, Ext^2 = 0, and the truncation triangle splits, so every bounded complex is isomorphic to the direct sum of its cohomology objects, each of which is a representation. Nonzero Ext^1 appears as morphisms between shifts in the formal model, not as an obstruction to formality. The rest of the argument, including the ADE-quiver grouping in Proposition 3.1, the mutation step, and the three-level tower for blow-ups of P^2, is consistent; reverse morphisms from the exceptional-divisor objects to the line bundles vanish by the semiorthogonality of Orlov's decomposition.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a method for proving Orlov's conjecture for certain blow-ups of projective spaces. The main technical device is to construct semiorthogonal decompositions of the blow-up whose components have Rouquier dimension zero, where the zero-dimensional components are derived categories of ADE quiver representations. The announced results are Theorem 4.1 (blow-ups of P^n in at most three disjoint linear subspaces, each a point or of codimension two), Proposition 4.2 (towers of three blow-ups of P^2 in at most three points each), and Corollary 4.3 (blow-ups of P^2 in up to nine arbitrary points), together with a P^3 analogue. The proof relies on an appendix that computes a 'dual' of Orlov's semiorthogonal decomposition for a point blow-up, and on a proposition that groups exceptional objects into quiver categories of Dynkin type.","tokens_in":9546,"tokens_out":40765,"duration_ms":391499,"significance":"If the main results were established, they would give new cases of Orlov's conjecture, including a uniform treatment of del Pezzo surfaces and some higher-dimensional blow-ups, and the use of ADE quiver categories as Rouquier-dimension-zero building blocks is a natural and potentially useful idea. The paper is concise and self-contained, and Lemma 2.4, despite a terse proof, is essentially correct via standard formality of hereditary categories. However, the central semiorthogonal decomposition used in the proof is not valid as stated, and the main theorems are not established by the argument presented.","major_comments":[{"comment":"The claimed semiorthogonal decomposition is false as stated. For a point blow-up in a surface (n=2), the claimed decomposition is <π*D^b(X), τ≥0 π*O_x>, and τ≥0 π*O_x is isomorphic to O_E, the structure sheaf of the exceptional divisor. The definition of semiorthogonal decomposition in Section 2 requires RHom(B,A)=0 for A in the left component and B in the right component. Taking X=P^2 and F=O(2), we have Hom_Y(O_E, π*O(2)) ≅ H^0(P^1, O_{P^1}(2)) ≠ 0, so the pair is not semiorthogonal. The proof of Prop. A.2 only verifies orthogonality inside the block T_k generated by the exceptional-divisor line bundles; it never verifies the required vanishing between the truncations τ≥-k π*O_x and the subcategory π*D^b(X). Moreover, the ordering contradicts the 'right dual' procedure defined in Section 2, which moves the mutated components to the left of all original components. Since Lemma 3.2, Proposition 3.1, and Theorem 4.1 all depend on this decomposition, the main proof is not valid.","section":"Appendix A, Prop. A.2"},{"comment":"The identification of the subcategory T=<L_{n-2}, {(S_b)_{n-2}}> with the derived category of a Dynkin quiver requires vanishings in the correct direction. The lemmas establish only RHom_Y(L_i,S_i)=k[0], i.e., morphisms from the earlier object to the later object. Semiorthogonality in the order <..., L_i, ..., S_i, ...> instead requires RHom_Y(S_i,L_i)=0, as well as RHom_Y(S_i,L_j)=0 for j≠i and vanishing of higher Ext groups. In the point case these reverse vanishings are not true in general: for n=2 and L_0=O(2), S_0=O_E, we have Hom_Y(O_E,O(2)) ≠ 0. Thus the grouping into a quiver category is not justified, even if one attempted to repair the preceding decomposition.","section":"Section 3, proof of Prop. 3.1"},{"comment":"The step 'Let T0 be the mutation of T through the subcategory ...' does not specify whether a left or right mutation is intended, and no orthogonality conditions needed for that mutation are verified. Because the preceding semiorthogonal decomposition is invalid, the iterative construction of T0,...,T_{n-2} is not established. This is load-bearing: the upper bound rdim(Y) ≤ n in Theorem 4.1 is obtained by applying Lemma 2.2 to the components constructed by this mutation process.","section":"Section 3, proof of Prop. 3.1, mutation step"}],"minor_comments":[{"comment":"The sentence 'the path algebra has homological dimension one, therefore any object of D^b(Q-rep) is quasiisomorphic to a direct sum of shifts of representations' is a formality claim that is true for hereditary abelian categories but not immediate from the quoted homological-dimension fact. Please add a sentence explaining that the truncation triangles split because Ext^{≥2}=0.","section":"Lemma 2.4"},{"comment":"The displayed conclusion 'RHom_Y(L_i, τ≥i π*O_x) ≅ k[0]' appears to contain a sign error: it should presumably be τ≥-iπ*O_x.","section":"Lemma 3.2, proof"},{"comment":"There are several typos: 'cardinatly' in Proposition 3.1, 'Rouquier dimenson' in the introduction, and the index set written as 'b∈b' instead of 'b∈B' in the proof of Proposition 3.1.","section":"Throughout"},{"comment":"In the first mutation step the text says 'through the two exceptional line bundles' but the displayed decomposition has three components; it would help to explicitly name which line bundle serves as L_0 when the second blow-up argument is applied.","section":"Proposition 4.2"}],"recommendation":"reject","confidential_remarks":"The reader's report and stress-test note focus on Lemma 2.4, but in my reading the more serious problem is in Appendix A and Lemma 3.2: the claimed dual semiorthogonal decomposition is false, with an explicit counterexample for n=2. The main theorems therefore rest on an invalid decomposition. I would not recommend further processing unless the authors can supply a correct blow-up decomposition or substantially revise the strategy."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid short note, and the reader's main worry—that Lemma 2.4 is false—doesn't survive contact with the paper. The lemma is true, the proof is just compressed. The paper proves Orlov's conjecture for a plane blown up in up to nine arbitrary points and for some higher-dimensional blow-ups of P^n in up to three points or codimension-two linear subspaces. The nine-point case is genuinely new; earlier del Pezzo results stopped at eight. The ADE-quiver grouping of exceptional objects is a nice new idea and gives a shorter route to the known del Pezzo cases too.\n\nWhat is actually new: Theorem 4.1 and the tower propositions. The method is to look for subcategories of Rouquier dimension zero beyond the usual exceptional-object categories: derived categories of ADE quivers. That is a neat observation and is applied cleanly through Orlov's decomposition and Bondal mutability. The paper is self-contained against standard tools (Rouquier's bound, Orlov's blow-up decomposition, Gabriel's theorem) and has no fitted parameters or circular claims.\n\nSoft spots: The proof of Lemma 2.4 is too terse. 'Homological dimension one' does imply that every bounded complex in a hereditary category is isomorphic to a direct sum of shifts of its cohomology objects—but only because Ext^2 vanishes and the truncation triangles split; the reader's non-split two-term complex is not actually a counterexample. Still, a referee should ask the author to add a sentence of justification or a citation. Proposition 3.1's mutation and regrouping steps are also compressed; I verified that the mutation of T through the lower (S_b) objects behaves as claimed, but 'repeating the argument n−2 times' hides a few checks about RHom vanishing after mutation. These are presentation gaps, not mathematical ones. Minor typos like 'cardinatly' are present but harmless.\n\nBottom line: it's a carefully reasoned note with a reusable technical point. People in derived algebraic geometry and representation theory will want it. I would send it to review; with small revisions to expand Lemma 2.4 and Proposition 3.1, it's ready.","headline":"A solid short note proving new cases of Orlov's conjecture; the reader's main worry about Lemma 2.4 is a false alarm, and the paper deserves a careful referee.","tokens_in":10055,"tokens_out":8160,"would_cite":true,"duration_ms":86280,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14F08","14E05"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that blow-ups of projective space in at most three points or codimension-two linear subspaces have Rouquier dimension equal to their ordinary dimension, confirming Orlov's conjecture for these varieties.","keywords":["Rouquier dimension","Orlov's conjecture","semiorthogonal decomposition","blow-up","derived category","ADE quiver","del Pezzo surfaces","exceptional collection"],"falsifier":"Search the derived category of representations of the $D_4$ quiver for a bounded complex that is not quasi-isomorphic to a direct sum of shifts of indecomposable representations; finding one would invalidate the proof of Lemma 2.4 and remove the certification that the decomposition's components have Rouquier dimension zero. Failing to find one, as expected from the hereditary structure, would support the paper's chain of reasoning.","tokens_in":9201,"feed_emoji":"📐","tokens_out":7231,"duration_ms":76048,"temperature":0.7,"pith_summary":"This paper proves that certain blow-ups of projective space satisfy Orlov's conjecture: the Rouquier dimension of the bounded derived category of coherent sheaves equals the usual dimension of the variety. The main theorem covers the blow-up of $\\mathbb{P}^n$ in at most three disjoint linear subspaces, each either a point or of codimension two. It also covers towers of small blow-ups, including a blow-up of $\\mathbb{P}^2$ in up to nine arbitrary distinct points. The proof works by decomposing the derived category into pieces of Rouquier dimension zero, rather than only into exceptional line bundles, so the sharp upper bound survives even though the exceptional collection grows longer.","feed_headline":"Blow-ups in at most three centers satisfy Orlov's conjecture","feed_subtitle":"Derived categories split into zero-dimensional quiver pieces, so the dimension bound survives repeated blow-ups.","key_machinery":"The load-bearing construction is a semiorthogonal decomposition of the blow-up's derived category obtained by mutating Orlov's blow-up decomposition. The key objects are exceptional sheaves supported on the exceptional divisors that have exactly one nontrivial morphism space from a chosen line bundle; together with that line bundle they generate a subcategory equivalent to the derived category of an ADE quiver, in particular the $D_4$ quiver, which has Rouquier dimension zero. Proposition 3.1 groups the exceptional objects coming from up to three blow-up centers into such zero-dimensional quiver categories, preserving the sharp bound from the glueing estimate.","core_discovery":"The central claim is Theorem 4.1: if $\\{Z_b\\}_{b\\in B}$ is a set of at most three disjoint linear subspaces of $\\mathbb{P}^n$, each a point or of codimension two, and $Y$ is the blow-up of $\\mathbb{P}^n$ in their union, then $\\operatorname{rdim} Y = n$. The proof constructs a semiorthogonal decomposition of $D^b_{\\mathrm{coh}}(Y)$ into two exceptional line bundles and $n-1$ subcategories each equivalent to the derived category of representations of a $D_4$ quiver, hence of Rouquier dimension zero. The glueing estimate for semiorthogonal decompositions then gives the upper bound $\\operatorname{rdim} Y \\le n$, while smoothness gives the lower bound $n \\le \\operatorname{rdim} Y$. Low-dimensional corollaries include $\\operatorname{rdim} = 2$ for a blow-up of $\\mathbb{P}^2$ in up to nine arbitrary points, which covers all del Pezzo surfaces, and $\\operatorname{rdim} = 3$ for a two-level tower of point and line blow-ups starting from $\\mathbb{P}^3$.","pith_inferences":["The same decomposition strategy should apply to any smooth $n$-fold carrying a full exceptional collection of length $n+1$ whose last $n-1$ bundles restrict to full exceptional collections on the blow-up centers; the projective-space assumption is likely convenient rather than essential.","Because the zero-dimensional components are ADE quiver categories, the groupability into $D_4$ quivers is the main structural obstruction; if a different ADE grouping could be arranged, the bound might extend to configurations with more centers.","The general pattern suggests a sufficient criterion: any blow-up whose new exceptional objects can be organized into finitely many ADE quiver categories inherits the sharp Rouquier dimension bound from its base variety."],"forward_implications":["Blowing up $\\mathbb{P}^n$ in at most three disjoint points or codimension-two linear subspaces produces a variety with Rouquier dimension exactly $n$.","A blow-up of $\\mathbb{P}^2$ in up to nine arbitrary distinct points has Rouquier dimension $2$, so all del Pezzo surfaces satisfy Orlov's conjecture.","A three-level tower of blow-ups of $\\mathbb{P}^2$ in at most three points per level satisfies the conjecture, as does a two-level tower of point and line blow-ups of $\\mathbb{P}^3$.","The decomposition strategy preserves two exceptional line bundles from the full exceptional collection on projective space, which is what allows the sharp bound to be iterated over several blow-ups in low dimension."],"supporting_citations":[{"why":"Defines Rouquier dimension and supplies the glueing estimate for semiorthogonal decompositions used to bound the dimension above.","marker":"[Rou08]"},{"why":"Provides Orlov's semiorthogonal decomposition for blow-ups, which is mutated to build the zero-dimensional components.","marker":"[Orl93]"},{"why":"Supplies the mutation machinery and the theorem identifying the subcategory generated by one source and up to three sinks with the derived category of a Dynkin quiver.","marker":"[Bon89]"},{"why":"Gives the full exceptional collection on projective space that anchors the induction.","marker":"[Bei78]"},{"why":"Established the conjecture for del Pezzo surfaces previously; the paper's proof is an alternative route to the same cases.","marker":"[BF12]"},{"why":"States the conjecture being verified and gives background on generators and dimensions of triangulated categories.","marker":"[Orl09]"},{"why":"Provides the standard property of dual exceptional collections used to compute the morphism spaces in the blow-up decomposition.","marker":"[Kap88]"},{"why":"Sources the equivalence between Rouquier dimension zero and having finitely many indecomposable objects up to shift, used in Lemma 2.3.","marker":"[CYZ08]"}],"fun_headline_variants":["Orlov's conjecture verified for blow-ups in up to three centers","Triple blow-ups: Rouquier dimension equals space dimension","Quiver pieces prove Orlov for blow-ups of projective spaces","Rouquier dimension = dimension for three-center blow-ups","Del Pezzo blow-ups confirm Orlov's conjecture via quivers"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument's load-bearing premise is that the derived category of finite-dimensional representations of an ADE quiver, in particular the $D_4$ quiver, has Rouquier dimension zero, with the proof as written relying on the step that every bounded complex over such a quiver splits into shifts of representations.","fun_headline_variants_meta":{"raw":{"variants":["Orlov's conjecture verified for blow-ups in up to three centers","Triple blow-ups: Rouquier dimension equals space dimension","Quiver pieces prove Orlov for blow-ups of projective spaces","Rouquier dimension = dimension for three-center blow-ups","Del Pezzo blow-ups confirm Orlov's conjecture via quivers"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000774,"raw_usage":{"total_tokens":3411,"prompt_tokens":919,"completion_tokens":2492,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":535,"completion_tokens_details":{"reasoning_tokens":2406}},"tokens_in":535,"tokens_out":2492,"duration_ms":18223,"temperature":1.0,"reasoning_tokens":2406,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:50:12.641532+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search the derived category of representations of the $D_4$ quiver for a bounded complex that is not quasi-isomorphic to a direct sum of shifts of indecomposable representations; finding one would invalidate the proof of Lemma 2.4 and remove the certification that the decomposition's components have Rouquier dimension zero. Failing to find one, as expected from the hereditary structure, would support the paper's chain of reasoning.","supporting_citations":[],"review_version":1}