{"id":"c0be7651-cdee-4815-b61a-12e75174f98b","arxiv_id":"2501.14882","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The authors prove the saturation and log-concavity conjectures for Fibonacci and Pell families of Markov polynomials and propose new conjectures about their coefficients and entropy.","lead":"This paper studies the coefficients of Markov polynomials, the Laurent polynomial solutions of a generalized Markov equation, by encoding them as weights on Newton polygons. It proves that all lattice points in the polygon appear as monomials for the Fibonacci and Pell families, and proposes new conjectures on log-concavity, divisibility, and an entropy function.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Saturation and all geometric statements rest on the unproved Newton-polygon description (Theorem 3.2); if Eq. (11) fails for some coprime a/b, the central conjecture and the special-case arguments lose their geometric basis.","rationale":"I read the paper as a conjecture-driven study whose main structural input is the explicit Newton polygon of Theorem 3.2. The reader's weakest assumption identifies exactly this theorem, and I agree that it is the most load-bearing point: the Saturation Conjecture 3.3, the coefficient formulas of Theorem 4.2, the critical triangle, the Markov sails, and the entropy domain all depend on Eq. (11). The paper's own note that the proof is omitted and the appeal to [16] is informal makes this the place where the general claim is least secure. I weighed whether the absent proof of Corollary 5.5 should be the main concern, but that corollary concerns a special family and is less foundational than the polygon theorem. The Fibonacci-family saturation proof is explicit and checkable, and the recurrence-based Pell argument is plausible, so the paper's special cases still give real evidence. Giving credit where due: the explicit coefficient formula (16), the Fibonacci log-concavity proof, and the sail coefficient theorems for Pell are concrete and verifiable. The correct verdict is CONDITIONAL: the general framework should be accepted only if Theorem 3.2 is either proved in full or replaced by a precise citation to a theorem in [16] that immediately implies Eq. (11). Since the reader already assigned CONDITIONAL, my assessment does not change the verdict.","tokens_in":86,"tokens_out":12608,"duration_ms":227844,"concrete_test":"Supply the missing induction for Theorem 3.2 from recurrence (8), proving for every Farey step that the Newton polytope of P_{(a+2c)/(b+2d)} is the convex hull of the Minkowski sum of the Newton polytopes of the two product factors plus the triangle from (u+v+w), with the correction monomial handled explicitly, and that the unique solution is Eq. (11). As a first numerical checkpoint, compute P_{3/5} and P_{4/7} in a CAS from the base cases (5) via (8) and verify that (i) the support contains exactly the lattice points of Eq. (11) and (ii) the convex hull of the support equals Eq. (11); any mismatch falsifies the polygon description.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that the support of P_{a/b} is exactly the lattice points of the region Delta_{a/b} = {i,j >= 0 : i/a + j/b >= 1, i+j <= a+b-1}. Every use of that claim, including the two special-case saturation proofs, the coefficient formulas in Theorem 4.2, and the sail/entropy constructions, presupposes Theorem 3.2. But Theorem 3.2 is stated with the proof omitted ('straightforward but a bit technical') and is delegated to an informal citation to [16] for Newton polytopes of rank 3 cluster algebras. The omitted induction is not purely formal: recurrence (8) contains a subtraction term, and positivity of coefficients does not by itself prevent the support or its convex hull from deviating from the predicted polygon. If for any coprime a/b the convex hull of supp(P_{a/b}) differs from Eq. (11), then the Saturation Conjecture 3.3 as stated, Theorem 4.2, and the sail constructions all require revision. The proof of the Pell-family saturation (Corollary 5.5) is also asserted with only 'From equations (20) we can deduce', so the two special cases do not independently verify the general polygon description.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies Markov polynomials, the Laurent-polynomial solutions of the generalized Markov equation obtained by cluster mutations from an initial triple (x,y,z), parametrized by rationals via the Conway topograph. The authors propose a description of the Newton polygon of each Markov polynomial (Theorem 3.2, Eq. (11)) and use it to formulate the Saturation Conjecture 3.3, explicit boundary and near-boundary coefficient formulas (Theorems 4.1–4.3), a log-concavity conjecture and partial results (Section 6), an entropy function in a continuum limit (Section 7), and a Markov-sail duality (Section 8). They prove the saturation conjecture and additional coefficient structure for the Fibonacci family M_{1/n} (Theorem 5.2, Corollary 5.3) and for the Pell family M_{n/(n+1)} (Corollary 5.5, Theorems 9.1–9.2), and they prove strict concavity of the entropy function for Fibonacci polynomials (Theorem 7.4).","tokens_in":18783,"tokens_out":27938,"duration_ms":200716,"significance":"If the main conjectures hold, the paper would provide a clean geometric/combinatorial model for Markov polynomial coefficients, connecting cluster algebras, continued fractions, and tropical geometry, with concrete arithmetic predictions such as the Factor-4 Conjecture and the Markov Sail Duality. The paper has clear strengths: the Fibonacci coefficient formula is derived from independent results of Caldero–Zelevinsky and Zelevinsky and is machine-checkable; the entropy function for Fibonacci polynomials is rigorously shown to be strictly concave; the Binet-type formula for Pell polynomials is explicit; and several falsifiable conjectures are stated. However, the central Newton-polygon description is not proved here, and one of the stated coefficient formulas is internally inconsistent with another theorem, so the paper currently requires major revision before the central claims can be regarded as established.","major_comments":[{"comment":"The explicit description of the Newton polygon is stated without proof: the text says \"This is straightforward but a bit technical, so we omit the details\" and cites [16] informally. This is load-bearing, since Eq. (11) is used to define the coefficient formulas in Theorem 4.2, the Saturation Conjecture 3.3, the entropy function in Section 7, and the sail constructions in Section 8. The induction is not purely formal because recurrence (8) contains a negative term; positivity of coefficients does not by itself prevent the convex hull of the support from deviating from Eq. (11). Please provide a complete proof of Theorem 3.2, or a precise statement and theorem number from Lee–Li–Schiffler [16] that implies it, and explain why subtraction in (8) cannot change the convex hull.","section":"Section 3.1, Theorem 3.2 and Eq. (11)"},{"comment":"Theorem 4.1 proves only the formula for T0 and declares the remaining cases \"similar\". Theorem 4.2, which is used in Sections 6 and 9, depends on all of the formulas in Theorem 4.1, so the missing proofs are consequential. Moreover, there is a concrete internal inconsistency: Theorem 4.1 gives R1(u,w) = u^a(3a-1)(u+w)^{b-2} + u^{a+1}(b-2a)(u+w)^{b-3}, while Theorem 4.2 states A_{i,1} = (3a-1) binom(b-2, i-a) + (b-2) binom(b-3, i-a-1). For M_{2/3}, the coefficient at (3,1) is 4, but the latter formula gives 6; the former formula gives 4. This error matters for Theorem 6.6, because Lemma 6.5 is proved only for positive A and B, while (b-2a) is negative for a/b > 1/2. Please correct the formula and supply complete proofs for all cases of Theorem 4.1.","section":"Section 4, Theorems 4.1 and 4.2"},{"comment":"The saturation claim for Markov-Pell polynomials is asserted with only \"From equations (20) we can deduce\" and no argument. Since this is one of the two special cases announced in the abstract as proved, a complete proof is needed. Recurrence (22) is a plausible starting point, but the paper does not show that the support of the coefficients is exactly the set of lattice points of the Newton polygon Δ_{n/(n+1)}. In addition, the claim that (20)-(21) is \"precisely the recurrence for the numerators of the Markov polynomials M_{k/(k+1)}\" is not immediate from (8) and needs a derivation.","section":"Section 5.2, Corollary 5.5"}],"minor_comments":[{"comment":"In the sentence introducing Theorem 3.2, \"the Newton polygon is the area on the ij-plane\" should be phrased as \"the convex hull in the ij-plane\" or \"the region\", since Newton polygons are convex hulls of supports.","section":"Section 3.1"},{"comment":"The recurrence for continued fractions is misstated: \"q_k = a_k q_{k-1} + p_{k-2}\" should be \"q_k = a_k q_{k-1} + q_{k-2}\" (and similarly for the earlier display).","section":"Section 8"},{"comment":"The displayed computation for A^{(2k+1)}_{n-1,2} omits the terms A^{(2k-1)}_{n-3,2} and 2A^{(2k-1)}_{n-2,1}; they vanish because they lie outside the relevant Newton polygon, but this should be stated explicitly.","section":"Section 9, proof of Theorem 9.2"},{"comment":"The proof of Proposition 7.1 is extremely terse: it says the result follows from the estimate A_{i,j}(ρ_n) < m_{ρ_n} and results of Fock. Since this is stated as a proposition, please either give a complete argument or clearly label the statement as a conjecture with supporting evidence.","section":"Section 7, Proposition 7.1"},{"comment":"The notation in Theorem 9.1 and surrounding text switches between n and k (e.g., A^{(2n+1)}_{1,n} vs. Eq. (22) written with k). Please harmonize the indexing so the recurrences and the final formulas are unambiguous.","section":"Section 5.2"}],"recommendation":"major_revision","confidential_remarks":"The paper contains interesting conjectures and two solid special-case computations (Fibonacci and Pell), but the unproved Newton-polygon description (Theorem 3.2) underlies most of the manuscript, and the inconsistency between Theorem 4.1's R1 and Theorem 4.2's A_{i,1} is a concrete error that affects later proofs. If the authors can supply a proof or a precise reference for Theorem 3.2, correct the coefficient formula, and give a real proof of Corollary 5.5, the paper would be suitable for publication in a number theory or combinatorics journal. I would not recommend acceptance in the current form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this for the Fibonacci and Pell coefficient formulas, not for the general theory. The paper makes a genuine contribution in the special cases: the explicit Fibonacci coefficient formula (Theorem 5.2) is a clean translation of the Caldero-Zelevinsky formulas, saturation for M_{1/n} and M_{n/(n+1)} is established, the log-concavity proof for the Fibonacci family is correct, and the Pell sail coefficients 4m (Theorem 9.2) are a nice concrete result. The authors also do good citation hygiene: the Newton polygon description is attributed to Ustinov and to Lee-Li-Schiffler, and the conjectures are clearly separated from theorems.\n\nThe soft spot is structural. Theorem 3.2, which identifies Delta_{a/b} with the region (11), has no proof ('straightforward but a bit technical'), and it is the base for everything else: Saturation Conjecture 3.3, Theorem 4.2, and the sail construction. The omitted induction is not purely formal because recurrence (8) has a subtraction term, so positive coefficients alone do not force the support to fill the predicted polygon. If Eq. (11) fails for some coprime a/b, the central conjecture and the sail duality need revision. I have not found a counterexample, and the authors' computations support it, but a proof or a precise pointer to [16] is needed before the general claims carry weight.\n\nTwo smaller gaps: Theorem 4.1 proves only T0 and says the rest are 'similar'—that's acceptable for a first version but not for the theorem as stated. Corollary 5.5 (Pell saturation) is asserted with 'From equations (20) we can deduce' and no details; it is believable but under-supported. The entropy section is interesting; the concavity proof for Fibonacci is explicit and correct, and the conjecture for general alpha is clearly labeled as a conjecture.\n\nWho is this for? People working on Markov numbers, cluster algebras, or log-concavity of combinatorial coefficients. It should not be desk-rejected: the special-case theorems are genuinely useful and the conjectures are well-motivated. A serious referee should ask for the proof of Theorem 3.2 (or exact citations) and for more detail on Theorem 4.1 and Corollary 5.5, but the paper will very likely survive those revisions.","headline":"The special-case results are real and worth publishing, but the general Newton-polygon theorem is unproved and load-bearing, so the conjectures are conditional.","tokens_in":19314,"tokens_out":3044,"would_cite":false,"duration_ms":26856,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11B39","11J70","13F60","52B20","05A20"],"pacs":[],"model":"deepseek-v4-flash","headline":"For Fibonacci and Pell Markov polynomials, the monomials that appear are exactly the lattice points of the Newton polygon; the paper conjectures this saturation for all cases and adds log-concavity, entropy, and sail-duality structure.","keywords":["Markov polynomials","Newton polygon","Saturation Conjecture","Fibonacci numbers","Pell numbers","log-concavity","entropy function","continued fractions"],"falsifier":"For a rational $a/b$ not of the two proved families, e.g. $2/5$, compute $P_{a/b}$ explicitly and check whether every lattice point of the claimed Newton polygon has a nonzero coefficient; a single zero coefficient would disprove the Saturation Conjecture, while a single missing lattice point or nonzero coefficient outside the polygon would falsify Theorem 3.2.","tokens_in":18309,"feed_emoji":"🧮","tokens_out":12116,"duration_ms":96638,"temperature":0.7,"pith_summary":"This paper studies the Laurent-polynomial solutions of the generalised Markov equation, called Markov polynomials, by treating the exponent pairs of their numerator monomials as points in a convex polygon. Its central proposal is the Saturation Conjecture: for the rational parameter $a/b$, every integer lattice point inside the Newton polygon (the region $i/a+j/b \\ge 1$, $i+j \\le a+b-1$) actually occurs as a monomial, with positive coefficient. The conjecture is proved for the Fibonacci family $a/b = 1/n$ and the Pell family $a/b = n/(n+1)$, using explicit coefficient formulas and recurrences. Along the way the paper gives exact binomial-type formulas for boundary and near-boundary coefficients, proves weak log-concavity for the Fibonacci family, computes a strictly concave entropy function in the continuum limit, and formulates two structural conjectures: the Factor 4 conjecture and a Markov sail duality that organise interior coefficients by continued-fraction geometry.","feed_headline":"Every lattice point is real: saturation proved for two Markov families","feed_subtitle":"The Saturation Conjecture holds for Fibonacci and Pell Markov polynomials, with explicit binomial coefficients.","key_machinery":"The central object is the Newton polygon $\\Delta_{a/b}$, the convex hull in the $(i,j)$-plane of the exponent pairs occurring in the homogeneous numerator $P_{a/b}(u,v,w)$; Theorem 3.2 identifies it as the region between the lines $i/a+j/b=1$ and $i+j=a+b-1$. The argument runs along the mutation formula $ZZ'=X^2+Y^2$ in the form of the numerator recurrence $P_{a+2c,b+2d}=(u+v+w)P_{c/d}P_{(a+c)/(b+d)}-u^c v^d w^{c+d}P_{a/b}$, which lets the paper induct from the base numerators $1$, $u+v$, $(u+v)^2+uw$ over the rational topograph. For the two special families the induction closes into explicit recurrences: the $1/n$ numerators satisfy the two-variable cluster recurrence $f_{m+1}=(f_m^2+1)/f_{m-1}$ and produce the closed coefficient formula $\\binom{n-j}{n+1-i-j}\\binom{i+j}{j}$, while the $n/(n+1)$ numerators satisfy $R_{2k+1}=(x^2+y^2)(x^2+y^2+z^2)R_{2k-1}-x^2y^2z^4R_{2k-3}$, whose coefficient-level recurrence propagates values across the polygon. A separate tool is the binary entropy function $H(p)=-p\\ln p-(1-p)\\ln(1-p)$, which computes the continuum limit of coefficients and yields the proved strict concavity for the Fibonacci family.","core_discovery":"The paper's starting description is Theorem 3.2: for coprime $a,b>0$, the Newton polygon $\\Delta_{a/b}$ of the numerator $P_{a/b}(u,v,w)$ is exactly the set $\\{i,j\\ge 0 : i/a+j/b \\ge 1,\\ i+j \\le a+b-1\\}$. Because Markov polynomials have positive coefficients, every lattice point in this polygon might in principle have a zero coefficient; the Saturation Conjecture asserts that the support of $P_{a/b}$ is precisely $\\Delta_{a/b} \\cap \\mathbb{Z}^2$. The paper proves this for $\\rho=1/n$ by deriving the closed coefficient formula $A_{ij}=\\binom{n-j}{n+1-i-j}\\binom{i+j}{j}$, and for $\\rho=n/(n+1)$ by a Pell-type recurrence that forces the same support. It also proves explicit binomial formulas for boundary and near-boundary coefficients, log-concavity along the principal directions for the Fibonacci family, strict concavity of the associated entropy function in the continuum limit, and carries out the first verification of the proposed sail-duality and location-of-4 conjectures on the Pell family, where the interior sail coefficients are shown to be $4m$.","pith_inferences":["If saturation holds for all rationals, the coefficient support of every numerator is a full lattice polygon; this would make Markov polynomials a test case for general saturation phenomena in cluster algebras, where only cluster variables of rigid representations are known to saturate.","The Factor 4 conjecture would follow from a free action on the perfect matchings that compute these coefficients; such an action, if it exists, would also explain the location-of-4 statement as a fixed-point contribution.","The sail-duality propagation rule resembles a discrete integrable system: starting from the seed value 4 it determines almost all interior sail coefficients by alternating differences, which suggests a direct continued-fraction proof of positivity along the sail might be available even without a full saturation proof.","The concavity of the entropy function for arbitrary rationals could be tested numerically: for large $n$ and a fixed scaled point $(\\xi,\\eta)$, ratios of coefficients along nearby rays should approach ratios of exponentials of the conjectured entropy, and the Hessian should stay negative definite."],"forward_implications":["For every Markov-Fibonacci polynomial $M_{1/n}$, the coefficient at $(i,j)$ is $\\binom{n-j}{n+1-i-j}\\binom{i+j}{j}$, so every lattice point of the Newton polygon has a positive coefficient and the Saturation Conjecture holds in this family.","For every Markov-Pell polynomial $M_{n/(n+1)}$, saturation holds as well; the interior sail coefficients are exactly $4,8,\\ldots,4(n-1)$, the coefficient at the penultimate convergent position is $4$, and the remaining boundary value is $7n-10$.","The boundary and near-boundary coefficients of every Markov polynomial are explicit sums of binomial coefficients: the top diagonal is $\\binom{a+b-1}{i}$, the vertical and horizontal edges are $\\binom{b-1}{i-a}$ and $\\binom{a-1}{j-b}$, and the next two diagonals have three-term binomial formulas.","The entropy function of the Fibonacci family is strictly concave, invariant under $(\\xi,\\eta)\\mapsto(\\xi,1-\\xi-\\eta)$, and attains its maximum $2\\ln((1+\\sqrt{5})/2)$ at a single interior point, so the growth of coefficients in the continuum limit is concentrated along one direction."],"supporting_citations":[{"why":"General description of Newton polytopes of rank-3 cluster variables from which the polygon formula of Theorem 3.2 is taken.","marker":"[16]"},{"why":"Shows Markov polynomials have non-negative integer coefficients and are all distinct, the positivity fact that makes saturation a meaningful question.","marker":"[22]"},{"why":"Introduces Markov polynomials as cluster-mutation solutions of the generalised Markov equation, the objects studied throughout.","marker":"[13]"},{"why":"Provides the explicit Laurent formulas for the type $A_1^{(1)}$ cluster variables that yield the $M_{1/n}$ coefficient formula.","marker":"[4]"},{"why":"Companion explicit formulas for the same cluster variables used in deriving Theorem 5.2.","marker":"[27]"},{"why":"Defines the continued-fraction sails and the edge-angle duality that the Markov Sail Duality conjecture extends.","marker":"[14]"},{"why":"Introduces $q$-deformed Fibonacci polynomials whose specialisations coincide with the numerators $P_{1/n}(q,1,q^2)$.","marker":"[18]"},{"why":"Supplies the real-zeros criterion used to prove log-concavity on the third diagonal for $a/b \\le 3/5$.","marker":"[25]"},{"why":"Provides the growth estimate on Markov numbers used to prove existence of the continuum entropy limit.","marker":"[10]"}],"fun_headline_variants":["Saturation proved for Fibonacci and Pell Markov polynomials","Explicit binomial coefficients settle Saturation Conjecture","Markov polynomials: support equals lattice points for Fibonacci, Pell","Proof of Saturation for Fibonacci and Pell families","Newton polygons: every lattice point realized for Fibonacci, Pell"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole structure rests on the unproved description of the Newton polygon as the region $i/a+j/b \\ge 1$, $i+j \\le a+b-1$, borrowed from a general rank-3 cluster-algebra result; if that description failed for some denominator, the saturation, log-concavity, and sail statements would need revision.","fun_headline_variants_meta":{"raw":{"variants":["Saturation proved for Fibonacci and Pell Markov polynomials","Explicit binomial coefficients settle Saturation Conjecture","Markov polynomials: support equals lattice points for Fibonacci, Pell","Proof of Saturation for Fibonacci and Pell families","Newton polygons: every lattice point realized for Fibonacci, Pell"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00065,"raw_usage":{"total_tokens":2967,"prompt_tokens":914,"completion_tokens":2053,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":530,"completion_tokens_details":{"reasoning_tokens":1976}},"tokens_in":530,"tokens_out":2053,"duration_ms":27915,"temperature":1.0,"reasoning_tokens":1976,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T14:49:55.016450+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a rational $a/b$ not of the two proved families, e.g. $2/5$, compute $P_{a/b}$ explicitly and check whether every lattice point of the claimed Newton polygon has a nonzero coefficient; a single zero coefficient would disprove the Saturation Conjecture, while a single missing lattice point or nonzero coefficient outside the polygon would falsify Theorem 3.2.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"General description of Newton polytopes of rank-3 cluster variables from which the polygon formula of Theorem 3.2 is taken."},{"cited_title":"Propp The combinatorics of frieze patterns and Markoff numbers","cited_arxiv_id":null,"evidence_quote":"Shows Markov polynomials have non-negative integer coefficients and are all distinct, the positivity fact that makes saturation a meaningful question."},{"cited_title":"Itsara, G","cited_arxiv_id":null,"evidence_quote":"Introduces Markov polynomials as cluster-mutation solutions of the generalised Markov equation, the objects studied throughout."},{"cited_title":"Caldero, A.V","cited_arxiv_id":null,"evidence_quote":"Provides the explicit Laurent formulas for the type $A_1^{(1)}$ cluster variables that yield the $M_{1/n}$ coefficient formula."},{"cited_title":"Zelevinsky Semicanonical basis generators of the cluster algebra of type A(1) 1","cited_arxiv_id":null,"evidence_quote":"Companion explicit formulas for the same cluster variables used in deriving Theorem 5.2."},{"cited_title":"Karpenkov Geometry of Continued Fractions","cited_arxiv_id":null,"evidence_quote":"Defines the continued-fraction sails and the edge-angle duality that the Markov Sail Duality conjecture extends."},{"cited_title":"Morier-Genoud and V","cited_arxiv_id":null,"evidence_quote":"Introduces $q$-deformed Fibonacci polynomials whose specialisations coincide with the numerators $P_{1/n}(q,1,q^2)$."},{"cited_title":"Stanley Log-concave and unimodal sequences in algebra, combinatorics, and geometry","cited_arxiv_id":null,"evidence_quote":"Supplies the real-zeros criterion used to prove log-concavity on the third diagonal for $a/b \\le 3/5$."}],"review_version":1}