{"id":"cdb984f5-3413-4d64-9412-b9dba21c7796","arxiv_id":"2411.09920","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"New bijective proofs show that one-leg and two-leg skew plane partitions and reverse plane partitions have generating functions differing exactly by MacMahon's function.","lead":"The authors find a direct, reversible way to connect two families of number-stacking objects whose counting formulas were known to match. Their technique uses simple local moves called toggles and gives the first such connection for two important special cases.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 5.3's stabilization proof has a reversed inequality in part 1 and a diagram-dependent induction; the two-leg bijection needs a corrected proof.","rationale":"The reader's weakest assumption is exactly Proposition 5.3, so we agree. Our stress-test sharpens this: part 1 of the proposition as written contains a logically reversed inequality; the proof of part 2 is sketched and depends on an omitted figure; part 3's weight argument is compressed. Because Theorem 5.4 cannot be defined without Proposition 5.3, the paper should be accepted only if the stabilization proof is corrected and made fully explicit. No error was found in the overall strategy, and the one-leg results (Section 4) are independent of this issue, so the paper remains fundamentally sound.","tokens_in":24777,"tokens_out":26733,"duration_ms":237929,"concrete_test":"Implement the toggle algorithm for two-leg SPPs exactly as in the paper. For a fixed small shape (e.g., λ=(2,2), μ=(3,1)) and all SPPs of weight ≤ 5, compute σ_n for increasing n and check: (i) an N exists with all later pops zero, and (ii) the diagonal equalities of Proposition 5.3 parts 2 and 3 hold. Also re-run the proof's part 1 with the inequality reversed. Any failure is a counterexample to the theorem; success supports the corrected proof.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim (Theorem 5.4) is well-defined only if Proposition 5.3 holds: after finitely many toggles, future pops are zero and the SPP stabilizes on diagonals. The proof of this proposition is the paper's weakest point. In part 1, the text states: 'if σ_n(i,j) ≠ σ(i,j), then |j−i| must be at least n'. But the diagonals toggled to produce σ_n are exactly those with |j−i| ≤ n−1, so the correct implication is the opposite ('less than n'). As written, the subsequent conclusion ('for |j−i| ≥ n, σ_n = max{λ_j, μ_i}') does not follow; the proof needs this inequality reversed. Part 2's induction is summarized with reference to Figure 5.4 and an unstated toggle formula, and part 3's weight comparison is terse. If any step fails, the infinite toggle process is not finite-time computable and the bijection collapses. This is a concrete, fixable gap, but it is load-bearing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents bijective proofs of two special cases of the Pandharipande–Thomas/Donaldson–Thomas correspondence, which states that the generating function for skew plane partitions equals MacMahon's function times the generating function for reverse plane partitions. The authors express the relevant generating functions as vacuum expectation values of vertex operators Γ±, and reinterpret the algebraic commutation relations Γ−(b)Γ+(a)=Γ+(a)Γ−(b)/(1−ab) as Pak-style toggles of diagonals. For the one-leg case (shape (∅,∅,λ)) they prove a bijection between SPPs and pairs (RPP, plane partition) using edge sign/power sequences, hook-length formulas, and an n-quotient map φλ. For the two-leg case (shape (λ,μ,∅)) they state a stabilization result (Proposition 5.3) and use it to define an infinite toggle process, yielding the bijection of Theorem 5.4. The paper closes with a discussion of obstacles in the three-leg case.","tokens_in":24969,"tokens_out":11106,"duration_ms":98940,"significance":"If correct, the paper provides the first genuinely combinatorial (bijective) proof of the PT–DT correspondence for the one-leg and two-leg families, complementing the recursive double-dimer proof of [JWY22]. The vertex-operator/toggle framework is elegant and likely to be reusable; the use of n-quotients to convert hook-length tableaux of different shapes is a nice idea. The paper is also commendably explicit about the limitation that the three-leg case requires new vertex-operator machinery. However, the main two-leg theorem depends crucially on Proposition 5.3, whose proof currently contains a reversed inequality and an under-specified induction; the one-leg theorem is also proved by analogy to Theorem 3.7 rather than by a self-contained argument. These points make the paper's central claims plausible but not yet fully rigorous.","major_comments":[{"comment":"The displayed implication 'if σ_n(i,j)≠σ(i,j), then |j−i| must be at least n' is reversed: the diagonals toggled to pass from σ to σ_n are exactly those with |j−i| < n, so a cell that changed must satisfy |j−i| < n. The subsequent conclusion that σ_n(i,j)=max{λ_j,μ_i} for |j−i|≥n is correct once the implication is reversed, but as written the proof contains a false intermediate assertion. Since part 1 supplies the base case for the induction in part 2, this needs to be corrected before Theorem 5.4 is supported.","section":"Section 5, Proposition 5.3(1)"},{"comment":"The induction step uses the formula 'each a_k toggles to min{a_{k−1}, b_{k−1}} + max{a_k, b_k} − a_k' without deriving it from Definitions 3.1/3.2, and the 'diagonal immediately above a' is asserted to be unchanged without proof. The diagram in Figure 5.4 cannot substitute for an index-level derivation. Without a rigorous induction establishing α(n+1,i)=α(n,i) and β(n+1,i)=β(n,i), the stabilization property that makes the infinite toggle process finite-time computable is not established.","section":"Section 5, Proposition 5.3(2)"},{"comment":"The proof of the one-leg bijection is by analogy to Theorem 3.7, relying on 'the same logic' for the hook-length weight preservation and for the well-definedness of the composition of toggles. Theorem 4.4 gives the relevant hook-length identity, but the text does not explicitly verify that the iterative commutation of Γ operators for an arbitrary edge sign sequence eλ produces toggles that terminate with the claimed weight bookkeeping. Given that Theorem 4.9 is one of the paper's two central results, a self-contained proof or a precise reduction to Theorem 3.7 is needed.","section":"Section 4, Theorem 4.9"}],"minor_comments":[{"comment":"The definition of z2 should explicitly account for the factor (1−q^{p_m+p_{m+1}})^{-1} when the swapped operators have opposite signs; as written, 'equal ... except swapped' is not literally true, and the objects counted by S2 for opposite signs include a nonnegative integer in addition to the partition-like object.","section":"Section 3, Lemma 3.6"},{"comment":"The relationship between the integer edge labels and the half-integer powers pλ(n)=±|n+1/2| is not stated clearly; please specify that n indexes edges by integers while the power sequence takes half-integer values, and reconcile this with the labels in Figure 4.1.","section":"Section 4, Definition 4.2"},{"comment":"The domain 'Z×N ∪ N×Z' has overlapping parts, and the weight sum 'over the entire diagram' should be made precise by writing the summation over the appropriate copies of the cells with explicit indicator functions.","section":"Section 5, Definition 5.2"},{"comment":"The assertion that 'the value of N guaranteed by Proposition 5.3 is N = 3' is not verified against the conditions of the proposition; a short check of part 1 for this example would increase confidence in the illustration.","section":"Section 5, Example 5.5"},{"comment":"There are several small textual errors, including 'a partition ν with with' in Definition 3.2 and 'the value of1' in Example 3.3; these should be corrected in a final revision.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The reversed inequality in Proposition 5.3(1) appears to be a typo rather than a conceptual error, and the missing toggle formula in part (2) can likely be supplied from Definition 3.2. If the authors provide a corrected and slightly expanded proof of Proposition 5.3, and flesh out the 'same logic' step in Theorem 4.9, the paper would be suitable for publication. The paper is within the scope of math.CO and the exposition is generally very clear. No concerns about novelty or attribution; the authors appropriately credit Pak, Sulzgruber, Hopkins, and the vertex-operator literature."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper does something new and useful. It gives the first combinatorial bijections for the one-leg and two-leg cases of the PT-DT correspondence, and it is refreshingly clear about what is borrowed (the toggle bijection tau is functionally Pak–Sulzgruber) and what is original (the vertex-operator framing plus the n-quotient map phi_lambda that makes the correspondence a reversible map). The one-leg Theorem 4.9 looks correct; the examples are detailed, and the hook-length bookkeeping checks out. The algebraic derivation of V = M*W from the vertex operators is also done carefully and does not assume the target identity.\n\nThe soft spot is in Section 5, Proposition 5.3, which is load-bearing for the two-leg bijection. In part 1, the text says that if sigma_n(i,j) != sigma(i,j), then |j-i| must be at least n. That is backwards: the diagonals toggled to produce sigma_n are exactly those with |j-i| <= n-1, so the implication should be |j-i| < n. As written, the subsequent conclusion that sigma_n equals the minimal configuration for |j-i| >= n does not logically follow. The good news is that the intended conclusion is true by a simpler argument: toggles only affect cells on toggled diagonals, and N was chosen so that sigma already equals the minimal configuration outside [1,N]^2. So the proof can be repaired by reversing the inequality and rewriting a few sentences. Part 2 of the same proposition is also sketched with a reference to Figure 5.4 and an implicit toggle formula; it is terse but plausible. Part 3's weight comparison is quick but I do not see a problem there.\n\nMinor issues: Lemma 3.6 proves only the opposite-sign case and waves at the others; Theorem 4.9 says the SPP toggle bijection works by \"the same logic\" as Theorem 3.7 without all the details. These are standard compression in a combinatorics paper, not fatal.\n\nWho this is for: people working on plane partitions, vertex operators, and the combinatorics of PT/DT invariants. The paper advances a subfield and opens a plausible route toward the three-leg case, while being explicit that the three-leg case is still open. It deserves a serious referee. I would send it out and ask for a corrected proof of Proposition 5.3 before accepting; with that fixed, it is a solid contribution.","headline":"Genuinely new bijections for two special cases of PT-DT; the one-leg proof is solid, and the two-leg proof has a concrete but fixable mistake in Proposition 5.3.","tokens_in":25506,"tokens_out":2755,"would_cite":true,"duration_ms":29310,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05A17","05A19","05E05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper constructs explicit weight-preserving bijections realizing the PT–DT correspondence for the one-leg and two-leg cases, using vertex operators and toggles.","keywords":["plane partitions","reverse plane partitions","skew plane partitions","PT-DT correspondence","vertex operators","toggles","hook lengths","n-quotients"],"falsifier":"Take a two-leg skew plane partition $\\sigma$ of some shape $(\\lambda,\\mu,\\emptyset)$ and run the toggle order used in the proof of Proposition 5.3, recording the popped-off values and the main-diagonal partition after each round; if for any $\\sigma$ infinitely many nonzero values are popped, or if the main-diagonal partition never becomes constant, then Proposition 5.3 is false and the two-leg bijection collapses.","tokens_in":24565,"feed_emoji":"🧮","tokens_out":11036,"duration_ms":96564,"temperature":0.7,"pith_summary":"The paper sets out to show that the PT–DT correspondence, an enumerative-geometry identity whose two sides count plane-partition-like objects, is not merely an equality of generating functions but a reversible combinatorial map, at least for the one-leg and two-leg cases. The main theorems produce, for any Young diagrams, a weight-preserving bijection between a skew plane partition and a pair consisting of a reverse plane partition and an ordinary plane partition, with weights adding. A sympathetic reader would care because the general identity was previously established by an involved recursive double-dimer argument, and a bijective proof explains the counting directly and may extend to the full three-leg setting. The method is to write each generating function as a product of vertex operators and to interpret each operator commutation as a local toggle of a diagonal.","feed_headline":"Toggles turn the PT–DT identity into an explicit bijection","feed_subtitle":"One- and two-leg skew plane partitions split into a reverse plane partition plus a plane partition.","key_machinery":"The load-bearing device is the toggle: a local involution, introduced in the form used here by a cited reference, that rewrites a diagonal of a plane-partition-like array relative to its two neighbouring diagonals and records a popped-off nonnegative integer. Toggling a diagonal is exactly a bijective realization of the commutation relation $\\Gamma_-(b)\\Gamma_+(a)=\\frac{1}{1-ab}\\Gamma_+(a)\\Gamma_-(b)$ of the vertex operators $\\Gamma_\\pm$, so an algebraic proof by successive commutations can be converted term-by-term into a bijection. The paper augments this with edge sign and edge power sequences encoding a shape and its hook lengths, with $n$-quotients used in the one-leg case to move hooks between the asymptotic Young diagram and the Young diagram itself, and with a stabilization theorem asserting that for two-leg skew plane partitions all but finitely many toggles pop off zeros and the diagonal partitions eventually become constant.","core_discovery":"The paper's central claim is that the PT–DT correspondence can be bijectivized for one- and two-leg shapes. For any Young diagrams $\\lambda$ and $\\mu$, Theorem 5.4 constructs a weight-preserving bijection between skew plane partitions of shape $(\\lambda,\\mu,\\emptyset)$ and pairs $(\\rho,\\pi)$ where $\\rho$ is a reverse plane partition of shape $(\\lambda,\\mu,\\emptyset)$ and $\\pi$ is a plane partition, with $|\\sigma|=|\\rho|+|\\pi|$. Theorem 4.9 gives the one-leg case of shape $(\\emptyset,\\emptyset,\\lambda)$. The proof realizes each generating function as a product of vertex operators $\\Gamma_\\pm$, then interprets every commutation of adjacent operators as a toggle of the corresponding diagonal; in the two-leg case a stabilization result makes the infinite toggle process finite and well-defined.","pith_inferences":["A natural next test, not pursued in the paper, would be to compare the output of the two-leg bijection with the double-dimer condensation construction on explicit small shapes; agreement would strengthen the case that the two approaches are computing the same correspondence.","If three-leg reverse plane partitions ever acquire a vertex-operator description, the toggling and stabilization apparatus here suggests a route to a fully general bijective PT–DT proof that bypasses the recursive double-dimer argument.","The role of $n$-quotients in the one-leg proof suggests a finer statement not stated in the paper: the bijection may be organized by hook-length residues, so that the distribution of hook lengths in the output tableau is governed by the $n$-quotients of the shape $\\lambda$; this could be tested on explicit examples."],"forward_implications":["For one-leg shapes, Theorem 4.9 gives a weight-preserving bijection between skew plane partitions of shape $(\\emptyset,\\emptyset,\\lambda)$ and pairs of a reverse plane partition of the same shape and a plane partition.","For two-leg shapes, Theorem 5.4 gives the analogous bijection for shape $(\\lambda,\\mu,\\emptyset)$, making $V_{(\\lambda,\\mu,\\emptyset)}(q)=M(q)W_{(\\lambda,\\mu,\\emptyset)}(q)$ a reversible combinatorial statement.","The bijections are algorithmic: toggling diagonals in any order compatible with corners yields the same hook-length-weighted tableau, so the resulting map is independent of the chosen toggle order.","The proof of the two-leg theorem depends on the stabilization property of Proposition 5.3, which asserts that every two-leg skew plane partition reaches a state after finitely many toggles from which all future toggles pop off zeros.","For the full three-leg case, the paper shows that the same methods face a substantial obstacle: three-leg reverse plane partitions are double-dimer objects with global labeling conditions and no known vertex-operator description, so a naive toggle generalization is not available."],"supporting_citations":[{"why":"Proved the general PT–DT generating-function identity via double-dimer condensation; this is the identity the paper bijectivizes in two special cases.","marker":"[JWY22]"},{"why":"Introduced the 3-fold vertex and stated the generating-function equality whose one- and two-leg cases are bijectivized here.","marker":"[PT09]"},{"why":"Supplies the vertex-operator expressions for skew plane partitions, reverse plane partitions, and MacMahon's function that set up the toggle arguments.","marker":"[ORV06]"},{"why":"Defines the $\\Gamma_\\pm$ vertex operators and their commutation relations, which the paper interprets bijectively via toggles.","marker":"[OR07]"},{"why":"Introduced the toggle operation in the exact form used here, together with the hook-length bijection that the paper re-derives and generalizes.","marker":"[Pak01]"},{"why":"Gave an independent toggle-based bijection for reverse plane partitions, cited as equivalent to the one-leg map used in the paper.","marker":"[Sul17]"},{"why":"Provided the earlier Hillman–Grassl bijection between reverse plane partitions and hook tableaux, the classical link that the toggle approach extends.","marker":"[HG76]"}],"fun_headline_variants":["Toggles make PT-DT bijective","Toggles give explicit PT-DT bijections","PT-DT identity becomes toggle bijection","Weight-preserving toggles biject PT-DT","Toggle bijectivizes one- and two-leg PT-DT"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that every two-leg skew plane partition stabilizes under repeated toggling: after finitely many steps all later toggles pop off only zeros and the diagonal partitions become constant, and without this the infinite toggle process would not be well-defined.","fun_headline_variants_meta":{"raw":{"variants":["Toggles make PT-DT bijective","Toggles give explicit PT-DT bijections","PT-DT identity becomes toggle bijection","Weight-preserving toggles biject PT-DT","Toggle bijectivizes one- and two-leg PT-DT"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001906,"raw_usage":{"total_tokens":7395,"prompt_tokens":802,"completion_tokens":6593,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":418,"completion_tokens_details":{"reasoning_tokens":6517}},"tokens_in":418,"tokens_out":6593,"duration_ms":48158,"temperature":1.0,"reasoning_tokens":6517,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T20:11:07.220716+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a two-leg skew plane partition $\\sigma$ of some shape $(\\lambda,\\mu,\\emptyset)$ and run the toggle order used in the proof of Proposition 5.3, recording the popped-off values and the main-diagonal partition after each round; if for any $\\sigma$ infinitely many nonzero values are popped, or if the main-diagonal partition never becomes constant, then Proposition 5.3 is false and the two-leg bijection collapses.","supporting_citations":[],"review_version":1}