{"id":"e86e794e-7ca6-4cdd-bcd9-9f5a54d61309","arxiv_id":"2506.01879","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Free Askey-Wilson functionals yield generating function representations for geometric LPP on a strip, giving the full phase diagram of the stationary measure and a Poisson approximation.","lead":"The authors find the exact large-scale behavior of a random growth model on a narrow grid, the geometric last passage percolation model on a strip. Their formulas determine the full phase diagram of the stationary measure and show that in one scaling limit the growth increments converge to a Poisson process.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 2.15(iii), equation (2.53), is internally inconsistent with its proof and with the later use in (3.21): the double-pole asymptotics is missing a factor n v and appears to have v^n in its place.","rationale":"The paper's central claim is the multipoint representation Theorem 1.4, from which the phase diagram (Theorem 1.3) and Poisson convergence (Theorem 1.5) are derived. The weakest spot in the chain is the asymptotic analysis in Proposition 2.15, exactly as the Reader identified. My stress-test found a concrete, checkable inconsistency in Proposition 2.15(iii): the formula as printed contradicts its own proof and the later normalization (3.21). The theorem statements appear to remain valid once the typographical correction is made, since the proof of Theorem 1.3(iv) uses the corrected version implicitly. No deeper flaw in Theorem 1.4 was identified: the main recursion (3.1) and (3.3) are supported by detailed combinatorial and functional arguments, and the omitted boundary case in Theorem 1.3(i) is a gap that can likely be filled by a limiting argument, though the paper should acknowledge it explicitly. The reader's CONDITIONAL verdict remains appropriate: the central results are credible, but the manuscript must be revised to fix (2.53) and to supply or reference the omitted details flagged in the proofs of (2.54) and Theorem 1.3(i). My recommendation is UNCHANGED because the concern does not move the verdict away from CONDITIONAL; it strengthens the case that conditional acceptance with corrections is the right outcome.","tokens_in":41258,"tokens_out":8199,"duration_ms":79987,"concrete_test":"Recompute the double-pole coefficient directly from (2.58)-(2.62): evaluate the contour integral (2.59) for H(z) for a concrete case, e.g. a=b=2, c=0, v=1/2, n=10, by high-precision numerical contour integration or by exact residue calculation, and compare the value with (i) the printed (2.53), (ii) the corrected form C(a,c) n v [a/((a-v)(1-av))]^{n+1}, and (iii) the normalization (3.21) at a=b=c, v=a. If the printed form fails and the corrected form matches, the proposition needs a correction; if the printed form somehow matches for the test case, the residue derivation in Section 2.6 and the application in (3.21) must be re-examined.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The phase-diagram result Theorem 1.3(iv) rests on the double-pole asymptotic (2.53), so the correctness of that formula is load-bearing. As printed, (2.53) states L^{(a,b,c)}[h_v^{-n}] ~ C(a,c) v^n [a/((a-v)(1-av))]^{n+1} with C(a,c) = (a^2-1)^2(1-ac)/(a^2(a-c)). The proof in Section 2.6, however, derives H(z) = B(z)/v (1-a^2 c B(z))/((1-aB(z))^2(1-cB(z))) and computes, for the double pole at z_a = (a-v)(1-av)/a, the residue contribution L ~ n z_a^{-(n+1)} Psi(z_a), where Psi(z_a) = (a^2-1)^2 v (1-ac)/(a^2(a-c)). Since z_a^{-1} = a/((a-v)(1-av)), the correct asymptotic is C(a,c) n v [a/((a-v)(1-av))]^{n+1}, with a factor n v, not v^n and not n alone. The discrepancy is not cosmetic: setting v=a, a=b=c, c=0 to obtain the normalization Z_N in Theorem 1.3(iv), the printed form yields an extra a^N and loses the linear factor N that appears in (3.21); the printed (2.53) would not produce the random mixture in (1.10). The proof of Theorem 1.3(iv) and the stated (3.21) are consistent with the corrected n v form, so this appears to be a typographical error in a central supporting proposition rather than a failure of the main result, but as written the proposition is false and must be corrected before the phase diagram proof is complete.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces free Askey-Wilson functionals L^{(a,b,c)} and uses them to give a multipoint generating function for geometric last passage percolation on a strip, stated as Theorem 1.4. The proof is an induction based on recurrences (3.1) and (3.3), with detailed appendix computations, and does not rely on Barraquand's contour integral formulas. From this representation the authors derive the phase diagram for the large-N limit of L_1(N)/N, Theorem 1.3, including the random uniform mixture on the diagonal, and a Poisson approximation under varying parameters, Theorem 1.5. The Laplace-transform asymptotics in Section 2.6 are the mechanism that turns the generating-function identity into the phase diagram.","tokens_in":41713,"tokens_out":21707,"duration_ms":193472,"significance":"If the asymptotic results are correct, the paper settles the phase diagram for the stationary measure of geometric LPP on a strip and adds a Poisson approximation, giving a substantial extension of the accessible boundary-parameter range. The free Askey-Wilson functional framework is a new tool in this context. A notable strength is that Theorem 1.4 is proved self-contained from recurrences, with the phase-diagram constants derived rather than fitted. The main weakness is the current state of Proposition 2.15(iii), whose printed asymptotic is inconsistent with its own proof; since Theorem 1.3(iv) rests on that asymptotic, the proof is not complete as written.","major_comments":[{"comment":"The double-pole asymptotic in (2.53) is inconsistent with its proof. In the proof of (2.53), the residue calculation gives L^{(a,b,c)}[h_v^{-n}] ~ n z_a^{-(n+1)} Psi(z_a), with Psi(z_a) = (a^2-1)^2 v(1-ac)/(a^2(a-c)) and z_a^{-1} = a/((a-v)(1-av)). This yields C(a,c) n v [a/((a-v)(1-av))]^{n+1}, not C(a,c) v^n [a/((a-v)(1-av))]^{n+1}. The printed formula has v^n where the proof gives n v. This is load-bearing: Theorem 1.3(iv) uses (2.53) to obtain the normalization Z_N in (3.21), and (3.23) uses that normalization to produce the uniform mixture in (1.10). As printed, (2.53) is false; for example, at v=a and c=0 it has a^N where the corrected asymptotic has N. The formula and the resulting constants in (3.21) and (3.23) must be corrected and rechecked before the phase-diagram proof is complete.","section":"2.6, Eq. (2.53)"},{"comment":"Theorem 1.3(i) is stated for c1,c2 <= 1, but the proof in Section 3.3.1 explicitly assumes c1,c2 < 1 and says 'the argument need to be modified and is omitted' for the boundary. Since c1 = 1 or c2 = 1 belongs to the statement of the theorem, this is a gap in the proof of the full phase diagram. Please either supply the boundary argument, for example using (2.54) after completing its proof, or restrict the statement of Theorem 1.3(i) to the open region and state the boundary cases separately.","section":"3.3.1, Theorem 1.3(i)"},{"comment":"The proof of (2.54) contains the sentence '(Here we omitted some details)' immediately after a dominated-convergence step, and the interchange of limit and integral is not fully justified. This asymptotic is needed for the boundary cases c = 1 in Theorem 1.3(i), which are currently omitted. As written, the proof of (2.54) is a sketch; please complete the justification, for instance by supplying a uniform integrability bound for the integrand after the change of variables u = (2-y)n.","section":"2.6, proof of (2.54)"}],"minor_comments":[{"comment":"The notation in Proposition 2.15 uses a for a functional parameter that in Theorem 1.3 is also called a, leading to confusion when the proposition is applied with a = c and v = a. Consider using different letters in the proposition or explicitly stating the substitution in Section 3.3.4.","section":"2.6, Eq. (2.53)"},{"comment":"After correcting (2.53), the displayed normalization in (3.21) still appears to have a misprint: substituting the corrected (2.53) with functional parameters c,c,0 and v=a gives a factor a in the numerator of the prefactor, not in the denominator. The subsequent line (3.23) appears consistent with the corrected normalization, so this is likely a typographical error, but it should be fixed.","section":"3.3.4, Eq. (3.21)"},{"comment":"The text says the ellipse gamma_rho in (2.22) is oriented clockwise, while later integrals are taken counter-clockwise. This is presumably intentional and resolved by the 'standard convention' sentence, but the figure caption and (2.22) could be clarified to avoid confusion.","section":"2.3.1, Fig. 2"},{"comment":"The use of [Mukherjea et al., 2006, Theorem 2] to pass from Laplace-transform convergence to weak convergence is terse; please state the exact theorem or give a self-contained argument, especially because the limiting formulas in Theorem 1.3 involve a random mixture rather than a deterministic limit.","section":"3.3, final paragraph"}],"recommendation":"major_revision","confidential_remarks":"The main theorem of the paper is the self-contained representation Theorem 1.4; the phase diagram and Poisson approximation are consequences of asymptotic analysis. The current manuscript has a false printed asymptotic in a central proposition (2.53) and an omitted boundary case in Theorem 1.3(i), so the proof is not yet complete. The issues appear fixable within the scope of the paper, so I recommend major revision rather than rejection. I would suggest that the authors, when revising, explicitly re-derive (3.21) from the corrected (2.53) and check all constants in the chain (3.21)-(3.23), since the final Laplace transform in Theorem 1.3(iv) depends on those prefactors."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does what it claims: it introduces free Askey-Wilson functionals, proves Theorem 1.4 (the multipoint generating function representation) by a self-contained recurrence induction, and derives the phase diagram (Theorem 1.3) plus a Poisson limit (Theorem 1.5). The novelty is solid: the formulas work for a wider boundary-parameter range than Barraquand's contour integrals, and the phase diagram was predicted by universality but not previously derived for this model. The authors give credit where due and the proof strategy is independent of Barraquand's contour formulas, which is good.\n\nThe main soft spot is a string of typos in the asymptotic expansions that support the diagonal case c1=c2=c>1. Proposition 2.15(iii) as printed is false: (2.53) states L ~ C(a,c) v^n [a/((a-v)(1-av))]^{n+1}, but the proof in Section 2.6 derives a double-pole residue that gives C(a,c) * n * v * [a/(...)]^{n+1}. The printed version lacks the factor n and has v^n instead of v. This is not cosmetic: Theorem 1.3(iv) and the normalization Z_N in (3.21) depend on this. I also found (3.21) itself is off by a factor a^2: it should be (c^2-1)^2 a N c^{N-2}/((c-a)(1-ac))^{N+1}, not with 1/a in the denominator. Similarly, the second atom term in (3.22) should have (c-a t_N^2)^N, not (c-a t_N)^N. Recomputing the Laplace transform ratio with these corrections gives exactly the mixture in (1.10), so the theorem is right; the printed equations are not. A referee should ask for these to be fixed, and also for the omitted boundary cases c1=1 or c2=1 in Theorem 1.3(i).\n\nThe rest of the paper is in better shape. The recurrence-based proof of Theorem 1.4 is detailed and the appendix computations are thorough. The asymptotic machinery in Proposition 2.15(i),(ii),(iv) also checks out. The Poisson theorem is technically involved and I did not find an error there, though I did not re-derive every step.\n\nWho is this for: people working in integrable probability and exact solvability, especially those interested in stationary measures of polymer models. It deserves a serious referee because the main results are significant and the proofs are largely sound once the typos are repaired. I would not desk-reject.\n\nIn short: engage with it, send it to referees, but tell the authors to fix the asymptotic formulas and complete the boundary-case proof.","headline":"The phase diagram result is real and the main proofs hold up, but Proposition 2.15(iii) and equations (3.21)-(3.22) contain typos that need correction before the paper is publishable.","tokens_in":42229,"tokens_out":34196,"would_cite":true,"duration_ms":258388,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60K35","33D45","82B23"],"pacs":[],"model":"deepseek-v4-flash","headline":"A single functional identity gives the full phase diagram of geometric LPP on a strip.","keywords":["geometric last passage percolation","stationary measure","free Askey-Wilson functionals","Askey-Wilson polynomials","phase diagram","Poisson approximation","strip","q=0"],"falsifier":"Take fixed $c_1=c_2=c\\in(1,1/a)$ and compute the normalization $Z_N=G_N(1)$ for large $N$ from the exact sum (1.3). The claim (3.21) predicts $Z_N\\sim \\frac{(c^2-1)^2}{a c^2(c-a)(1-ac)}\\,N\\,\\left(\\frac{c}{(c-a)(1-ac)}\\right)^N$. A numerical check of the ratio $Z_N N^{-1}\\left(\\frac{(c-a)(1-ac)}{c}\\right)^N$ against this constant, or a direct comparison of the two-atom Laplace transform (3.22) with the uniform-mixture formula, would settle whether the central phase-diagram claim is right.","tokens_in":41100,"feed_emoji":"🎲","tokens_out":7231,"duration_ms":63149,"temperature":0.7,"pith_summary":"This paper establishes a representation for the multipoint generating function of geometric last passage percolation (LPP) on a finite-width strip: the generating function equals an iterated free Askey-Wilson functional, a $q=0$ relative of the classical Askey-Wilson moment functional. The representation holds for boundary parameters over a wider range than previous contour-integral formulas, and from it the authors read off the full phase diagram for the large-scale limit of $L_1(N)/N$, the height of the last passage line at the left end of the strip. In the bulk regime the limit is $a/(1-a)$; in two boundary-dominated regimes the limits are $a/(c_1-a)$ and $ac_2/(1-ac_2)$; and on the critical diagonal $c_1=c_2=c>1$ the limit is a random uniform mixture of the two boundary values. The paper also proves a Poisson approximation for the line profile when the boundary parameters vary with strip width. If the main identity holds, it provides a single analytic object governing the stationary measure's large-scale behaviour.","feed_headline":"One identity yields the full phase diagram of LPP on a strip","feed_subtitle":"Boundary parameters set the limiting speed; on the critical diagonal it is a uniform random mixture.","key_machinery":"The key object is the free Askey-Wilson functional $L^{(a,b,c)}$, the $q=0$ specialization of the Askey-Wilson moment functional: a linear functional on analytic functions defined by contour integration on ellipses $z=w+1/w$ with kernel $(1-w^2)(1-abcw)/(w(1-aw)(1-bw)(1-cw))$, which reduces to integration against signed measures when the parameters are small enough. Composition is handled by letting two parameters depend on a variable through $u(x)$, the Joukowsky inverse, producing functionals $\\pi_{t;c_2}=L^{(c_2t,c_1/t)}$ and $P^{s,t;c_2}_x=L^{(c_2t,su(x)/t,s/(tu(x)))}$. The reduction formula $L^{(a,b,c)}[h_c f]=(1-ac)(1-bc)L^{(a,b)}[f]$ and its variant Corollary 2.14 replace the parameter $c_2$ by $a$ up to a multiplicative factor, and this change-of-parameter identity drives the induction proving Theorem 1.4. The accompanying asymptotic expansions of $L^{(a,b,c)}[h_v^{-n}]$ as $n\\to\\infty$, with simple-pole, double-pole and boundary cases, supply the exponential rates and prefactors whose ratios give the four limiting densities in Theorem 1.3.","core_discovery":"The paper's central claim is Theorem 1.4: for $0<a<1$, $ac_1<1$, $0<t_1\\le\\cdots\\le t_N$ with $at_N<1$ and $a t_N^2 c_2<1$, the $N$-point generating function of the stationary measure satisfies $G_N^{(a,c_1,c_2)}(t)=\\pi_{t_1,\\ldots,t_N;c_2}[\\bigotimes_{j=1}^N 1/h_{a t_j}]$, where $\\pi$ is the composition of free Askey-Wilson functionals defined through the inverse Joukowsky map. The authors prove this identity by induction using a recurrence for the generating function and a matching recurrence for the functionals, without invoking the earlier contour-integral derivation. From the identity they obtain the Laplace transform of $L_1(N)/N$ as a ratio of two evaluations of the same functional, and asymptotic expansions of those evaluations yield the phase diagram of Theorem 1.3 and the Poisson convergence of Theorem 1.5.","pith_inferences":["The same functional calculus should apply to other $q=0$ integrable models; for the six-vertex model on a strip, where a similar phase diagram has been reported, the free Askey-Wilson composition could yield an equally explicit proof of the limiting regimes.","The signed-measure representations in Appendix B suggest that fluctuation exponents in the boundary-dominated regimes can be extracted from the atomic parts of the measures, in analogy with ASEP shock fluctuations; the paper does not compute such fluctuations.","A testable extension is to replace the Laplace-transform convergence criterion by a direct steepest-descent analysis of the contour integrals in Proposition 2.12, which would sharpen error terms in Theorem 1.3 and possibly reveal the $N^{1/3}$ scale.","The Poisson theorem suggests that when $a$ and $c_2$ are tuned as $N\\to\\infty$, the strip becomes effectively one-dimensional at the left boundary; this may be the first member of a family of Poisson-type limits for multi-layer polymers on strips."],"forward_implications":["If $c_1,c_2\\le 1$, then $L_1(N)/N\\to a/(1-a)$ in probability, independent of the boundary parameters.","If $c_1>1$ and $c_2<c_1<1/a$, the limit becomes $a/(c_1-a)$, set by the left boundary.","If $c_2>1$ and $c_1<c_2<1/a$, the limit becomes $ac_2/(1-ac_2)$, set by the right boundary.","On the diagonal $c_1=c_2=c\\in(1,1/a)$, the limit law is the uniform mixture $(a/(c-a))U+(ac/(1-ac))(1-U)$ with $U\\sim\\mathrm{Unif}[0,1]$.","Under two scalings of the parameters with strip width $N$, the process $(L_1(\\lfloor Nx\\rfloor))_{0\\le x\\le 1}$ converges in finite-dimensional distributions to a Poisson process of rate $\\lambda$."],"supporting_citations":[{"why":"Defines geometric LPP on a strip, proves existence and uniqueness of the stationary measure, and supplies the weight function (1.2) that the present paper analyzes.","marker":"[Barraquand et al., 2024]"},{"why":"Derived the contour-integral formulas for the multipoint generating function that the free Askey-Wilson representation extends and makes more explicit.","marker":"[Barraquand, 2024]"},{"why":"Introduces the Askey-Wilson polynomials; the $q=0$ specialization of their moment functional is the free Askey-Wilson functional used throughout.","marker":"[Askey and Wilson, 1985]"},{"why":"Links stationary measures of ASEP to Askey-Wilson-type integral representations, motivating the phase-diagram analogy used in the introduction.","marker":"[Bryc and Weso lowski, 2017]"},{"why":"Provides the signed Askey-Wilson measure framework used in Proposition 2.8 and Appendix B for integral representations of the functionals.","marker":"[Wang et al., 2024]"},{"why":"Supplies the moment-generating-function convergence criterion that turns the Laplace-transform limits into weak convergence in Theorem 1.3.","marker":"[Mukherjea et al., 2006]"},{"why":"Gives the complex-analysis lemma (Lemma 2.11) that makes the composition of functionals analytic and well defined.","marker":"[Markushevich, 1977]"},{"why":"Used for Mergelyan's theorem in Proposition 2.4, extending the functionals from polynomials to analytic functions.","marker":"[Rudin, 1987]"}],"fun_headline_variants":["Free Askey-Wilson functionals reveal LPP strip phase diagram","New functionals yield full phase diagram for LPP on a strip","One functional identity maps LPP strip phase diagram","Free functionals pin down LPP strip boundary phases"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The phase diagram is read off from asymptotic expansions of $L^{(a,b,c)}[h_v^{-n}]$ (formulas (2.51)-(2.54)); if any of those expansions has a wrong constant or pole order — the double-pole case (2.53) is stated with details omitted — the limiting densities and the uniform mixture on the diagonal would be different.","fun_headline_variants_meta":{"raw":{"variants":["Free Askey-Wilson functionals reveal LPP strip phase diagram","New functionals yield full phase diagram for LPP on a strip","One functional identity maps LPP strip phase diagram","Free functionals pin down LPP strip boundary phases"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00084,"raw_usage":{"total_tokens":3641,"prompt_tokens":908,"completion_tokens":2733,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":524,"completion_tokens_details":{"reasoning_tokens":2664}},"tokens_in":524,"tokens_out":2733,"duration_ms":19292,"temperature":1.0,"reasoning_tokens":2664,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:32:31.337536+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take fixed $c_1=c_2=c\\in(1,1/a)$ and compute the normalization $Z_N=G_N(1)$ for large $N$ from the exact sum (1.3). The claim (3.21) predicts $Z_N\\sim \\frac{(c^2-1)^2}{a c^2(c-a)(1-ac)}\\,N\\,\\left(\\frac{c}{(c-a)(1-ac)}\\right)^N$. A numerical check of the ratio $Z_N N^{-1}\\left(\\frac{(c-a)(1-ac)}{c}\\right)^N$ against this constant, or a direct comparison of the two-atom Laplace transform (3.22) with the uniform-mixture formula, would settle whether the central phase-diagram claim is right.","supporting_citations":[{"cited_title":"and Wilson, J","cited_arxiv_id":null,"evidence_quote":"Introduces the Askey-Wilson polynomials; the $q=0$ specialization of their moment functional is the free Askey-Wilson functional used throughout."},{"cited_title":"Asymmetric Simple Exclusion Process with open boundaries and Quadratic Harnesses","cited_arxiv_id":"1511.01163","evidence_quote":"Links stationary measures of ASEP to Askey-Wilson-type integral representations, motivating the phase-diagram analogy used in the introduction."},{"cited_title":"Askey-Wilson signed measures and open ASEP in the shock region","cited_arxiv_id":"2307.06574","evidence_quote":"Provides the signed Askey-Wilson measure framework used in Proposition 2.8 and Appendix B for integral representations of the functionals."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the moment-generating-function convergence criterion that turns the Laplace-transform limits into weak convergence in Theorem 1.3."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the complex-analysis lemma (Lemma 2.11) that makes the composition of functionals analytic and well defined."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Used for Mergelyan's theorem in Proposition 2.4, extending the functionals from polynomials to analytic functions."}],"review_version":1}