{"id":"645ca133-2b9b-4dfe-9ae0-3b55c6fa9571","arxiv_id":"2607.14055","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Triangle surfaces are exactly the Markov-type cubics xyz=x^2+y^2+z^2+ax+by+cz+d, and their automorphism groups are Gσ⋊Γ with Γ one of five finite groups from Table 1.","lead":"The authors classify algebraic surfaces that can be completed by adding a triangle of rational curves at infinity: they are exactly the Markov-type cubic surfaces, and their full symmetry groups are computed. The result links a geometric boundary condition to the classical Markov equation, the Farey tessellation, and the modular group.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Triangle-complex construction rests on unproved imported theorem [26, Prop 3.3]; if it fails or does not apply to the stated NC-pair generality, the PGL2(Z) action and automorphism-group computation collapse.","rationale":"The reader's weakest_assumption identifies Proposition 3.3 as the most load-bearing premise, and I agree. The triangle complex construction and hence the kernel computation all depend on it. However, I do not think this concern by itself warrants changing the ACCEPT verdict: [26] is accepted for publication, and for the triangle surfaces that are the focus of the paper, Y is normal affine, which forces every NC-completion to be normal; thus the original hypotheses of [26] are likely satisfied. The gap is one of self-containedness and verification, not an observed counterexample. I also considered the unproved anticanonical embedding in Corollary 5.4, but this is a standard fact for Gorenstein del Pezzo surfaces of degree 3 and is less central than Proposition 3.3. Proposition 6.7 from El-Huti is another imported theorem, but it concerns the cubic side directly and is explicitly stated to cover singular surfaces. The rest of the proof is internally consistent: Lemma 6.2 establishes normality of the cubics, the marking arguments in Section 3 are careful, and the table in Theorem 6.12 is checked against stabilizers in S4. Therefore the central claim is well-supported conditional on the imported structural theorem; the reader's accept-with-moderate-confidence remains appropriate.","tokens_in":23885,"tokens_out":49634,"duration_ms":423184,"concrete_test":"Independently re-derive [26, Thm 4.17] in the setting of normal NC-pairs and verify Proposition 3.3. Concretely, for the smooth Markov surface Y, take the standard (0,0)-completion and apply the two distinct elementary transformations in the Farey tessellation to obtain two (0,0)-completions separated by two moves; compute the relatively minimal resolution of the identity map between them and check that the dual graph of the total transform is a chain of exactly two triangles. If the dual graph branches or contains a cycle, Proposition 3.3(iv) is false and the Farey identification fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The entire combinatorial machinery — Lemma 3.11, Lemma 3.12, the marking bijection of Lemma 3.14, and the homomorphism Aut(Y)→PGL2(Z) in Theorem 3.21 — depends on Proposition 3.3, imported from the authors' earlier paper [26, Thm 4.17, Prop. A.4]. That proposition guarantees that any birational map between (0,0)-completions of a fixed surface decomposes into inner blowups/blowdowns and that a relatively minimal resolution has a dual graph which is a triangulable chain of triangles. If (ii) or (iv) of Prop. 3.3 were false for some triangle surface, the adjacency graph of triangles in T(Y) would not be an infinite 3-regular tree, the identification with the Farey tessellation would fail, and the computation of the kernel of the action would be invalid. The paper does not reproduce the proof; it only notes that [26] assumed normal X whereas here we allow isolated singularities. For the main case Y normal affine, X is normal (as the paper observes in Section 4), so the original hypothesis is probably satisfied, but this is not shown in the text. This is a genuine proof gap rather than a demonstrated mathematical error: the central claim may still be true, but the paper's proof delegates a load-bearing structural theorem to an unpublished reference.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces triangle surfaces: normal affine surfaces admitting a completion whose boundary is a triangle of contractible (-1)-curves. It constructs a triangle complex T(Y) whose vertices are inner components, edges are (0,0)-completions, and triangles are triangle completions, and proves that T(Y) is identified with the Farey tessellation, giving a homomorphism Aut(Y) -> PGL2(Z) (Theorem 3.21). The central classification result is that every triangle surface is isomorphic to a Markov-type cubic xyz = x^2+y^2+z^2+ax+by+cz+d (Corollary 5.4), and conversely every such cubic is a triangle surface (Proposition 6.3). The automorphism group is then computed as G_sigma ⋊ Gamma, where G_sigma is the free product of the three Vieta involutions and Gamma is a finite group determined by a,b,c via Table 1 (Theorem 6.12). Applications to the Markov surface, double Fricke surface, and generalized Markov numbers are given.","tokens_in":24204,"tokens_out":19671,"duration_ms":187452,"significance":"If the two flagged gaps are filled, this is a valuable contribution: it gives a purely geometric characterization of Markov-type cubics, unifies several known families, and attaches to each surface a beautiful combinatorial model (the Farey tessellation) that controls completions and automorphisms. The explicit automorphism groups, the semidirect product structure, and the recovery of prior results on Markov-like surfaces are significant. The paper is generally well written, with many checkable details, and the main conclusion is plausible and does not, on inspection, depend on assuming its own conclusion.","major_comments":[{"comment":"Proposition 3.3 is imported without proof from [26, Thm 4.17, Prop. A.4], but [26] assumes normal surfaces, whereas the present paper's NC-pair definition (§2) allows isolated singularities. The proposition is then used in Lemmas 3.10–3.12 and Theorem 3.21 for arbitrary surfaces admitting a (0,0)-completion. The triangle complex, the marking bijection, and the homomorphism Aut(Y) -> PGL2(Z) all rest on this result. Since the main classification concerns normal affine surfaces, the authors should either prove the needed variant with the present hypotheses or restrict §3 to normal Y and explicitly verify that every NC-completion of a normal Y is normal (the observation that this holds appears only in §4). This is a genuine proof gap, not a demonstrated mathematical error.","section":"§3, Prop. 3.3"},{"comment":"The corollary asserts that the anticanonical map of the possibly singular degree-3 del Pezzo completion X is an embedding into P^3. Proposition 5.3 proves only that D is ample, Cartier, and -K_X ~ D; it does not show that |-K_X| is base-point-free and very ample. For singular Gorenstein del Pezzo surfaces this is a nontrivial statement and needs a proof or a precise reference covering the singular case (the cited [11, Thm 8.3.2] appears to address the smooth case). This step is load-bearing: it produces the cubic equation (4) on which the normal-form and automorphism analysis depends. Please supply the missing argument or an exact reference.","section":"§5, Cor. 5.4"}],"minor_comments":[{"comment":"The sentence 'Rescaling all three coordinates by 1/3' appears to state the inverse of the intended substitution: the next sentence shows that a Markov triple (m,n,k) becomes (3m,3n,3k). The rescaling should be 'by 3'.","section":"§7.1"},{"comment":"The condition 'a=b≠0, ±c≠a' is ambiguous; it should be written 'c≠±a' or 'c^2≠a^2'.","section":"§6, Table 1"},{"comment":"The Gröbner basis computation is not shown. Since the argument is a computational assertion, please include the polynomial P or provide a reproducible link to the computation.","section":"§6, Lemma 6.2"},{"comment":"The reduction to the smooth case says 'We may assume X, and hence Y, to be smooth' and refers forward to the proof of Proposition 5.3. This is fine, but the minimal-resolution claim should be stated explicitly, since it is used to justify the identification of T(Y) and the fibrations.","section":"§4, Prop. 4.11"},{"comment":"There is a typographical issue: 'nonzerotranslationwouldreintroducemixedquadratic' should have spaces. Also, 'stibilizers' in Theorem 6.12 is a typo.","section":"§6, proof of Lemma 6.10"}],"recommendation":"major_revision","confidential_remarks":"The main concern is the dependence on [26], a paper co-authored by one of the current authors that is accepted but not yet published. To make the present paper self-contained, it would be advisable to provide a proof of Proposition 3.3 in an appendix or to restate it with full hypotheses and verify them. The Gröbner basis claim in Lemma 6.2 is a finite computation; the authors should make it traceable. The two major comments are both addressable within the manuscript's scope, so major revision is appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a real paper and I'd send it out. The main novelty is the triangle complex T(Y) and the clean statement that affine triangle surfaces are exactly the Markov-type cubics xyz = x^2 + y^2 + z^2 + ax + by + cz + d, with a five-row table for Aut(Y). The Farey tessellation picture is genuinely illuminating, and the automorphism table for all a,b,c is new. The special cases (Markov surface, double Fricke, generalized Markov) all land correctly.\n\nWhat's good: the proofs are mostly detailed and checkable. Proposition 5.3 (the completion is a degree-3 del Pezzo with anticanonical boundary) is solid, and Lemma 6.2's Gröbner basis argument for normality is acceptable. The use of El-Huti's free product theorem is legitimate and independent.\n\nThe one real soft spot is Proposition 3.3. It is the foundation for the whole triangle complex: without it you don't get the decomposition into inner blowups, the triangulation of the dual graph, or the PGL(2,Z) action. The proof is not given; it's quoted from [26], an accepted paper but by one of the authors. The stress-test worry that this is a gap is fair. It's not a demonstrated error, and for the affine case Y is normal so the original normal-surface hypothesis of [26] should apply, but the paper doesn't show that. I'd want the referee to check the generality of Prop. 3.3 for the (0,0)-completions they use, or ask the authors to include the proof.\n\nSecond minor thing: Corollary 5.4 asserts the anticanonical map embeds a possibly singular degree-3 del Pezzo into P3. That's standard but unstated; a one-line reference to a textbook would settle it. Also, the normalization issue in Section 4 is handled implicitly.\n\nBottom line: if [26] holds up, the main theorem is true and the paper is a significant contribution. It deserves a serious referee, not a desk reject. I'd cite it. Take it to the reading group.","headline":"Solid, novel classification of affine triangle surfaces with a useful automorphism table, but the main combinatorial engine is an unproved import from a companion paper.","tokens_in":24690,"tokens_out":3441,"would_cite":true,"duration_ms":33550,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14J50","14R20","14L30","05C60"],"pacs":[],"model":"deepseek-v4-flash","headline":"A normal affine surface whose boundary at infinity is a triangle of contractible (−1)-curves is necessarily a Markov-type cubic xyz = x² + y² + z² + ax + by + cz + d, and its full automorphism group is the free product of the three Vieta in","keywords":["triangle surface","Markov-type cubic","automorphism group","Farey tessellation","PGL(2,Z)","Vieta involution","affine surface","del Pezzo surface"],"falsifier":"Exhibit a normal affine surface admitting a completion by a triangle of smooth rational (−1)-curves whose anticanonical completion is not a cubic hypersurface in P³ (for instance, a degree-3 del Pezzo whose anticanonical linear system fails to be an embedding); this would contradict Corollary 5.4 and the classification. Alternatively, produce two (0,0)-completions whose relatively minimal connecting resolution has a circular dual graph that admits no triangulation, falsifying Proposition 3.3.","tokens_in":23777,"feed_emoji":"🔺","tokens_out":6592,"duration_ms":61953,"temperature":0.7,"pith_summary":"The paper sets out to show that the purely geometric condition of admitting a completion by a triangle of contractible (−1)-curves completely pins down an affine surface's algebra: such surfaces are exactly the cubic surfaces of Markov type, given by one normal form equation. A companion combinatorial result describes all completions of such a surface by the triangle complex, a simplicial complex identical to the Farey tessellation, whose vertices correspond to the surfaces' fibrations over the affine line. The payoff is a complete, explicit computation of the automorphism group: the three Vieta involutions generate a free product, and the full group is that free product extended by a finite group of signed coordinate permutations read off from the coefficients. If correct, this gives a uniform explanation of why Markov's classical tree and many Diophantine generalizations all share the same and modular-group-shaped symmetry.","feed_headline":"Every affine triangle surface is a Markov cubic","feed_subtitle":"A boundary triangle pins the surface to a Markov-type cubic and determines all its symmetries.","key_machinery":"The triangle complex T(Y) is a 2-dimensional simplicial complex whose vertices are the inner components of Y (divisorial valuations arising from repeated blowups at boundary nodes), whose edges are the (0,0)-completions (boundary a cycle of two 0-curves), and whose triangles are the triangle completions. Marking its vertices by primitive vectors of Z² identifies T(Y) with the Farey tessellation, so automorphisms of Y act through PGL(2, Z); the three Vieta involutions correspond to the three elementary matrices. This complex is the load-bearing combinatorial skeleton: it converts every birational self-map into an automorphism of a canonical infinite tessellated tree, from which the kernel com","core_discovery":"The central claim is that 'triangle surface' is not a new class but a new characterization of a known one: every normal affine surface completed by a triangle of contractible (−1)-curves is isomorphic to a cubic surface of Markov type, xyz = x² + y² + z² + ax + by + cz + d, and conversely every such cubic admits such a triangle completion (Corollary 5.4, Proposition 6.3). The proof runs through the geometry of the completion: the boundary triangle is an anticanonical divisor on a (possibly singular) del Pezzo surface of degree 3, the anticanonical embedding realizes it as a plane triangle of lines, and the equation follows. On the automorphism side, the paper constructs a homomorphism from t","pith_inferences":["The Farey-tessellation identification suggests a modular-dynamics reading: the action of Aut(Y) on T(Y) factors through PGL(2, Z) acting on rational cusps, so one could measure the complexity of a triangle completion by the continued-fraction length of the corresponding rational point; an extension would associate a continued-fraction expansion to each Vieta-word decomposition of an automorphism.","Because the classification holds over any algebraically closed field of characteristic zero, the same single equation likely governs triangle surfaces over number fields; if so, integral points of each triangle surface would be orbit points under the Vieta-like involutions, giving explicit parametrizations that generalize Markov triples and cluster mutations.","The five-row table suggests a stratification of the coefficient space of Markov-type cubics by stabilizer type; quotienting (a, b, c) by the S₄ of monomial symmetries yields exactly five strata, so the automorphism group is locally constant on strata and jumps only at special equalities — a useful skeleton for studying degenerations of the moduli space.","The finiteness of the kernel (order ≤ 4) is a rigidity statement: an open surface with a triangle at infinity has no nontrivial automorphism fixing the boundary even setwise? (actually componentwise) — a natural testable extension is whether this rigidity persists for completions by a polygon of contractible curves."],"forward_implications":["Every affine triangle surface is presented by a single cubic equation xyz = x² + y² + z² + ax + by + cz + d; conversely every such cubic is a triangle surface, so the class is closed and recognizable by one normal form.","The automorphism group of any triangle surface is explicitly computable from the coefficients: Aut(Y) = Gσ ⋊ Γ, with the five cases in Table 1. In particular, the Markov surface itself has the maximal group K₄ ⋊ PGL₂(Z).","The classical Markov tree, the double Fricke surface, and the generalized Markov equations are all the same phenomenon: their integral-point mutations are generated by the Vieta involutions, and the symmetries of the solution tree form a quotient of PGL₂(Z) (or its kernel Γ(2)).","The vertices of the triangle complex give an intrinsic classification of fibrations of such surfaces over A¹ with general fiber A¹∖{0} (Proposition 4.11), so the combinatorics of completions has a geometric meaning independent of the chosen completion.","As a corollary of Theorem 3.21, any automorphism fixing every component and node of the boundary triangle acts as the identity on all inner components and is a sign change of order at most 4; in particular, there are no nontrivial automorphisms fixing the boundary divisor pointwise."],"fun_headline_variants":["Triangle at infinity marks a Markov cubic","Affine triangle surfaces equal Markov-type cubics","Triangle surface boundary pins down its cubic type","Open triangle surfaces are all Markov cubics","Triangle at infinity forces Markov cubic shape"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"Everything rests on the imported theorem that birational maps between two-curve completions decompose into local blow-ups and blow-downs with a triangulable dual graph, together with the unproved assertion that the anticanonical map of the possibly singular del Pezzo completion embeds into P³; if either fails, the classification into Markov-type cubics collapses.","fun_headline_variants_meta":{"raw":{"variants":["Triangle at infinity marks a Markov cubic","Affine triangle surfaces equal Markov-type cubics","Triangle surface boundary pins down its cubic type","Open triangle surfaces are all Markov cubics","Triangle at infinity forces Markov cubic shape"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000297,"raw_usage":{"total_tokens":1467,"prompt_tokens":562,"completion_tokens":905,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":306,"completion_tokens_details":{"reasoning_tokens":850}},"tokens_in":306,"tokens_out":905,"duration_ms":9079,"temperature":1.0,"reasoning_tokens":850,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T02:55:18.000290+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a normal affine surface admitting a completion by a triangle of smooth rational (−1)-curves whose anticanonical completion is not a cubic hypersurface in P³ (for instance, a degree-3 del Pezzo whose anticanonical linear system fails to be an embedding); this would contradict Corollary 5.4 and the classification. Alternatively, produce two (0,0)-completions whose relatively minimal connecting resolution has a circular dual graph that admits no triangulation, falsifying Proposition 3.3.","supporting_citations":[],"review_version":1}