{"id":"b44155a2-cb00-4d9f-b46f-d607e61437aa","arxiv_id":"2505.08211","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Open sets in braid varieties defined by transversality to a coordinate flag are isomorphic to products of two simpler braid varieties, and in the double Bott-Samelson case this splicing respects cluster structures.","lead":"This mathematics paper shows that certain spaces built from braids, called braid varieties, can be cut into open pieces that are products of two smaller braid varieties. It conjectures that these cuts preserve a hidden algebraic structure called a cluster structure, and proves this in an important special case.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.1's splicing isomorphism is asserted to be independent of the reduced expression for w0 (Remark 5.4), but the paper only says 'one can check'; if different reduced expressions give different maps, the canonical statement of the theorem fails.","rationale":"The reader's weakest assumption identifies exactly the same gap: the independence of the splicing map from the reduced expression for w0, plus the incomplete verification of the inverse map. This is the most load-bearing concern because Theorem 1.1 is the paper's main geometric result, and the statement 'we have an isomorphism' requires a well-defined map, not a family of maps depending on an auxiliary reduced expression. The paper explicitly says 'one can check' in Remark 5.4 and 'We leave details to the reader' in Step 3 of Theorem 5.2, so these are admitted omissions rather than hidden errors. The concern is concrete and testable: a small explicit computation with two reduced expressions for w0 in S4 would settle whether the maps agree. If they agree, the theorem stands as stated with a routine gap to be filled; if they differ, the theorem would need a modified statement or an additional canonicality argument. Since the issue is an unverified but plausibly routine detail, the appropriate verdict is the same CONDITIONAL given by the reader: the central claim is likely correct but not yet fully justified. No grounds for rejection are apparent, and no alternative concern is more load-bearing.","tokens_in":36772,"tokens_out":11168,"duration_ms":114577,"concrete_test":"Perform an explicit computation for k=4, β = σ1σ2σ1σ3σ2σ1σ1 (so δ(β)=w0), with r1=1, β1=σ1, β2=σ2σ1σ3σ2σ1σ1, and w=s1. Use two different reduced expressions for w0 that both start with s1, e.g. A = s1s2s1s3s2s1 and B = s1s3s2s1s3s2. Choose a numerical point z ∈ X(β) satisfying the equations of Corollary 4.5 (w0Bβ(z) upper triangular). For each of A and B, compute the image under Φ_{r1,w} using the explicit matrix construction in Steps 1–2 of Theorem 5.2, using Lemma 4.9 to determine the y-functions. Then identify the resulting points in X(β2w) via their flag chains, and check whether they coincide under the canonical identification of braid variety realizations. If the two outputs differ, Remark 5.4 fails and Theorem 1.1 is choice-dependent; if they agree, the independence claim is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central isomorphism Ψ_{r1,w} in Theorem 1.1 is not a priori canonical: its construction in Steps 1–2 of Theorem 5.2 fixes a reduced expression (5.2) for w0 that begins with a reduced word for w, and uses the functions y from Lemma 4.9 and the intermediate coordinate flags eF_i of diagram (5.5) determined by that expression. Remark 5.4 asserts that a different choice of reduced expression gives 'essentially the same map', but no proof is given. The target X(β2w)×X(w^{-1}w0β1) is canonically independent of the choice, so the issue is whether the constructed point in the target is unchanged (up to the canonical identifications) when the continuation of the reduced expression after w is varied. Without this verification, Theorem 1.1 defines not one map but a family of maps depending on an auxiliary choice, and the statement 'we have an isomorphism' is only as strong as the unproved independence claim. The inverse construction in Step 3 of Theorem 5.2 also ends with 'We leave details to the reader', so even the existence of a well-defined two-sided inverse is not fully documented. These are addressable gaps, but they are load-bearing because they concern the well-definedness of the main theorem, not merely a peripheral remark.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies braid varieties X(β) for positive braids β in type A. For a decomposition β = β1β2 and a permutation w, it defines an open subset U_{r1,w} of X(β) by the condition that the r1-th flag is transverse to the coordinate flag F(w0w). The central result, Theorem 1.1 (restated as Theorem 5.2), claims an isomorphism Ψ_{r1,w} : X((w^{-1}w0)β1) × X(β2w) → U_{r1,w}, with an explicit coordinate construction of the map and its inverse. For fixed r1, the sets U_{r1,w} cover X(β), and a Demazure-product criterion for nonemptiness is given. The paper then conjectures that this splicing map is compatible with the cluster structures on braid varieties and proves the conjecture in the special cases w = e, w = w0, and for double Bott–Samelson varieties. Applications to open Richardson varieties include an open embedding R(u,v) × R(v,w) → R(u,w) and a formula for the number of frozen variables.","tokens_in":37006,"tokens_out":16463,"duration_ms":163045,"significance":"If the central isomorphism is fully established, it gives a systematic splicing decomposition of braid varieties that unifies earlier positroid and Richardson constructions, yields a cover of X(β) by products of braid varieties, and implies the frozen-variable inequality f1 + f2 ≥ f. The explicit coordinate form of the map is a strength, as is the complete cluster quasi-isomorphism proof in the double Bott–Samelson case. The paper is honest about what is conjectural, and the special cases proved are nontrivial. However, the main theorem currently rests on an unproved independence claim and an inverse construction that is partly deferred, so the central statement is not yet fully documented.","major_comments":[{"comment":"The claim that Ψ_{r1,w} is independent of the reduced expression (5.2) is asserted with the phrase 'one can check' and no verification is given. The construction in Steps 1–2 of Theorem 5.2 uses the specific suffix a_{ℓ(w)+1},…,a_{ℓ(w0)} of the reduced expression for w0 to define the functions y_L via Lemma 4.9 and the intermediate coordinate flags eF_i of diagram (5.5). If different reduced expressions for w0, or different reduced words for w, produce different maps, then Theorem 1.1 does not define a single map Ψ_{r1,w} but rather a family of maps depending on an auxiliary choice. This is load-bearing for the well-definedness of the main isomorphism. Please provide a complete proof of the claimed independence, or reformulate Theorem 1.1 and Theorem 5.2 so that the chosen reduced expression is part of the data.","section":"Remark 5.4"},{"comment":"The construction of the inverse map is incomplete. After specifying matrices M1 and M2 and the decomposition M2^{-1}M1 = U' w0 V', the proof stops with 'We leave details to the reader.' A complete proof must show that the resulting element g is independent of the auxiliary choices, that translating the flags in diagram (5.14) by g produces coordinate flags of the required form, and that the two resulting tuples lie in X(β2w) and X((w^{-1}w0)β1) respectively. Without this, the well-definedness of Φ_{r1,w}^{-1} is not established, and the two-sided inverse claim in Theorem 5.2 is only partially verified.","section":"Theorem 5.2, Step 3"}],"minor_comments":[{"comment":"The assertion that Theorem 1.1 holds in arbitrary type is not proved and is not immediate from the given type-A proof, which uses type-A-specific facts such as Lemma 2.4, Lemma 4.8, and the explicit coordinate flags F(w). Please either add a proof or clearly label this as a conjecture.","section":"Remark 1.2"},{"comment":"The target of Φ1 is written X(w^{-1}w0β1) in the introduction and X((w^{-1}w0)β1) in Theorem 5.2; please make the parenthesization and order of factors consistent throughout.","section":"Section 5.1, equations (5.3)–(5.4)"},{"comment":"The translation from the Demazure-product condition δ((w^{-1}w0)β1)=w0 to the inequality δ(β1) ≥ w0ww0 is not shown. Since this is a useful nonemptiness criterion, a one-line derivation from Lemma 2.1 would improve readability.","section":"Corollary 5.5"},{"comment":"The remark refers to unpublished work [1] for a weighted-flag realization; if this is not available, please either include the needed definitions or state the remark as a sketch.","section":"Remark 5.3"},{"comment":"The notation S^{◦,1}_{λ/μ} is used without definition before it is introduced; please define or give a precise reference at first use.","section":"Section 6.3"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is plausible and the gaps identified are omissions of proof rather than evident contradictions. The manuscript would be strengthened by making the reduced-expression independence an actual proof, since it is part of the main theorem's well-definedness. The cluster-theoretic parts rely heavily on previous work of overlapping authorship; this is prior published support rather than circularity, but the novelty relative to [17,18] should be stated more sharply."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"I'd want you to know two things about this paper. First, Theorem 1.1 is real: for every positive braid beta and every cut position r1, the open sets U_{r1,w} are isomorphic to products of two braid varieties, and for fixed r1 they cover X(beta). That is new beyond the positroid and skew-shaped positroid cases, and it is the main event. Second, the proof has a gap that is load-bearing but almost certainly repairable. The construction of Psi_{r1,w} in Theorem 5.2 fixes a reduced expression for w0 starting with a reduced word for w. Remark 5.4 says different choices give 'essentially the same map' and gives no proof. If that is not true, Theorem 1.1 is not a single isomorphism but a family depending on auxiliary data. A referee should make the authors prove Remark 5.4, not just assert it.\n\nWhat the paper does well: the coordinate proof of the isomorphism is explicit and mostly convincing. The cover statement and the nonemptiness criterion (Corollary 5.5) are clean. The specialization to open Richardson varieties (Theorem 1.5) is a nice application, and the double Bott-Samelson results (Theorem 1.7) actually prove the cluster-compatibility conjecture in that case, with quiver computations that I checked in the examples. The paper is honest about what is conjectural: Conjecture 5.6 is labeled a conjecture throughout.\n\nThe other soft spots are proportionally smaller. Step 3 of Theorem 5.2 ends with 'We leave details to the reader' for the inverse map; that needs to be written out. Remark 1.2 claims arbitrary type without proof, but it is a remark. The self-citation load in the cluster part is real, but the cited cluster structure is already published, so it is not circular.\n\nWho is this for: anyone working on braid varieties, cluster structures on Richardson varieties, or the link-homology side of splicing maps. It deserves a serious referee. My recommendation: send it to review, and have the referee insist on a proof of Remark 5.4 and a fuller Step 3. With those fixed, the paper will be in good shape.","headline":"Theorem 1.1 is a genuinely new structural result, but the canonicality of the splicing map (Remark 5.4) is asserted rather than proved and needs to be fixed before publication.","tokens_in":37600,"tokens_out":2833,"would_cite":true,"duration_ms":28480,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14M15","13F60","20F36"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every positive braid variety has an open cover by products of two braid varieties.","keywords":["braid varieties","splicing maps","open Richardson varieties","double Bott-Samelson varieties","cluster algebras","Demazure product","flag varieties","quasi-cluster isomorphisms"],"falsifier":"Compute the two splicing maps for a small braid in $S_4$ with $w=s_1s_2$ using two different reduced expressions for $w_0$ with the same prefix, and check whether the resulting isomorphisms to $X(w^{-1}w_0\\beta_1)\\times X(\\beta_2 w)$ coincide; a mismatch, or a point where the inverse is not defined, would break the canonical statement of Theorem 1.1.","tokens_in":36541,"feed_emoji":"🧬","tokens_out":8325,"duration_ms":72026,"temperature":0.7,"pith_summary":"This paper studies braid varieties $X(\\beta)$, affine varieties built from chains of flags whose successive relative positions are prescribed by a positive braid $\\beta$. For any factorization $\\beta=\\beta_1\\beta_2$ and any permutation $w$, it defines an open set $U_{r_1,w}\\subseteq X(\\beta)$ by requiring the $r_1$-th flag to be transverse to a certain coordinate flag, and proves that $U_{r_1,w}$ is isomorphic as an algebraic variety to the product $X(w^{-1}w_0\\beta_1)\\times X(\\beta_2 w)$. For a fixed cutting point $r_1$ these open sets cover $X(\\beta)$, so every braid variety can be assembled from simpler pieces. The paper conjectures that the splicing map is compatible with the cluster algebra structure on $X(\\beta)$ up to a quasi-cluster isomorphism, and proves the conjecture for double Bott-Samelson varieties. A direct corollary is the frozen-variable inequality $f_1+f_2\\geq f$.","feed_headline":"Braid varieties split into product pieces","feed_subtitle":"New open sets cover every braid variety, each isomorphic to a product of two simpler braid varieties.","key_machinery":"The load-bearing object is the splicing map $\\Psi_{r_1,w}$ and its inverse $\\Phi_{r_1,w}$, built from a matrix decomposition: a flag transverse to $F(w_0w)$ admits a unique decomposition $M=w_0w L U$ with $L$ lower-unitriangular and $U$ upper-triangular. Sliding $U$ past the remaining braid factors via the braid-matrix commutation rewrites the two halves of the chain in the coordinates of $X(w^{-1}w_0\\beta_1)$ and $X(\\beta_2 w)$. The Demazure product $\\delta$, the permutation obtained by greedily taking longest reduced subwords, controls nonemptiness. The left-to-right inductive Deodhar torus supplies the cluster variables whose freezing defines the spliced open charts in the double Bott-Samelson case.","core_discovery":"The central result, Theorem 1.1, asserts that for every positive braid $\\beta$ with a decomposition $\\beta=\\beta_1\\beta_2$ and every $w\\in S_k$, the open set $U_{r_1,w}(\\beta)\\subseteq X(\\beta)$ is isomorphic as an algebraic variety to $X(w^{-1}w_0\\beta_1)\\times X(\\beta_2 w)$. The isomorphism is realized by an explicit splicing map: split the flag chain at position $r_1$, translate the two halves by elements that straighten the transverse pair, and insert coordinate flags along a reduced expression for $w_0$ that begins with a reduced expression for $w$. The nonemptiness of $U_{r_1,w}$ is characterized by the Demazure product conditions $\\delta(w^{-1}w_0\\beta_1)=w_0=\\delta(\\beta_2 w)$. Specializing to open Richardson varieties yields $U_{u,v,w}\\cong R(u,v)\\times R(v,w)$, and for double Bott-Samelson varieties the paper proves the full cluster-theoretic version of the conjecture.","pith_inferences":["A direct extension: if the independence claim in Remark 5.4 is verified, the splicing construction becomes canonical on the braid monoid rather than on a chosen word, and the same flag-position proof should carry Theorem 1.1 to other Weyl groups.","An unstated consequence: because the braid factors on the two sides of (1.1) splice to a braid conjugate to $\\beta_1\\beta_2$, the isomorphism suggests a geometric realization of the multiplication maps on Khovanov-Rozansky homology, a connection the paper raises as motivation but does not prove.","A testable computational extension: in examples such as the braid of Example 5.11, check Conjecture 5.6(1) by factoring the minors that define $U_{r_1,w}$ and testing whether their irreducible factors form a cluster; the example already shows the quasi-isomorphism need not preserve the mutable variables, so the conjecture leaves room for multiple maps."],"forward_implications":["Every braid variety $X(\\beta)$ is covered by the open sets $U_{r_1,w}$, each isomorphic to a product of two braid varieties from shorter braids; this yields a recursive way to compute invariants such as the number of frozen cluster variables.","The frozen-variable inequality $f_1+f_2\\geq f$ follows from Theorem 1.1 together with the fact that invertible functions on a cluster variety are monomials in frozen variables.","In the open Richardson case, splicing gives $U_{u,v,w}\\cong R(u,v)\\times R(v,w)$, and iterating along a maximal chain embeds $(\\mathbb{C}^\\times)^\\ell$ into $R(u,w)$.","For double Bott-Samelson varieties $BS(\\beta_1\\beta_2)$, the paper proves the spliced and product cluster structures are quasi-cluster equivalent, making the splicing map a bona fide cluster quasi-isomorphism in this case."],"supporting_citations":[{"why":"Defines braid varieties and constructs the cluster structure on $\\mathbb{C}[X(\\beta)]$ that the splicing conjecture refers to.","marker":"[2]"},{"why":"Supplies Deodhar-torus coordinates and the cluster structure for general braid varieties.","marker":"[12]"},{"why":"Supplies the 3D plabic graph construction of braid variety cluster structures.","marker":"[13]"},{"why":"Gives the cluster structure on double Bott-Samelson cells that Theorem 6.9 compares with the spliced structure.","marker":"[26]"},{"why":"Constructs open embeddings of Richardson varieties of the form $R(u,v)\\times R(v,w)\\to R(u,w)$, the comparison point for Theorem 1.5.","marker":"[6]"},{"why":"Provides the result that global invertible functions on a cluster variety are monomials in frozen variables, used for Corollary 1.3.","marker":"[14]"},{"why":"Earlier splicing construction for skew shaped positroids that the paper generalizes.","marker":"[17]"},{"why":"Earlier splicing construction for positroid varieties that the paper generalizes.","marker":"[18]"}],"fun_headline_variants":["Braid varieties splice into product pairs","Open sets in braid varieties are products","Splicing braid varieties into two factors","Braid varieties: a product decomposition via splicing","Every braid variety open set factors as a product"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The theorem's canonical form depends on the assertion, checked only in principle in Remark 5.4, that the splicing isomorphism is independent of the chosen reduced expression for $w_0$ that starts with a reduced word for $w$; the well-definedness of the inverse is likewise left partly to the reader in Step 3 of Theorem 5.2.","fun_headline_variants_meta":{"raw":{"variants":["Braid varieties splice into product pairs","Open sets in braid varieties are products","Splicing braid varieties into two factors","Braid varieties: a product decomposition via splicing","Every braid variety open set factors as a product"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000303,"raw_usage":{"total_tokens":1744,"prompt_tokens":944,"completion_tokens":800,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":560,"completion_tokens_details":{"reasoning_tokens":731}},"tokens_in":560,"tokens_out":800,"duration_ms":7276,"temperature":1.0,"reasoning_tokens":731,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T22:00:06.668386+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the two splicing maps for a small braid in $S_4$ with $w=s_1s_2$ using two different reduced expressions for $w_0$ with the same prefix, and check whether the resulting isomorphisms to $X(w^{-1}w_0\\beta_1)\\times X(\\beta_2 w)$ coincide; a mismatch, or a point where the inverse is not defined, would break the canonical statement of Theorem 1.1.","supporting_citations":[{"cited_title":"Cluster structures on double Bott-Samelson cells","cited_arxiv_id":null,"evidence_quote":"Gives the cluster structure on double Bott-Samelson cells that Theorem 6.9 compares with the spliced structure."},{"cited_title":"Factorial cluster algebras","cited_arxiv_id":null,"evidence_quote":"Provides the result that global invertible functions on a cluster variety are monomials in frozen variables, used for Corollary 1.3."},{"cited_title":"Splicing skew shaped positroids","cited_arxiv_id":"2503.04923","evidence_quote":"Earlier splicing construction for skew shaped positroids that the paper generalizes."}],"review_version":1}