{"id":"e36e24fe-d3a5-4108-8e76-65d5ff7aa8de","arxiv_id":"1908.07950","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A quaternion-real power-law banded random matrix model for the symplectic class is shown to obey existing heuristic multifractal and spectral relations at criticality in limited parameter ranges.","lead":"A random matrix model for the symplectic, spin-orbit coupled symmetry class is introduced and tested at the metal-insulator transition. The paper completes the power-law banded random matrix picture across all three Wigner-Dyson symmetry classes.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed agreement for β=4 is not established in the intermediate-b regime; Sec. 3.2 admits deviations for 0.04<b<2, yet the abstract claims good agreement, and the fitted αq and χ checks lean on this relation.","rationale":"The reader's weakest assumption concerns the universality class of the β=4 PBRM model. That is a legitimate concern for the physical relevance of the model, but the central claim as stated in the paper is about heuristic relations holding within the newly introduced model. The more immediate and internally verifiable weakness is that the key heuristic relation, Eq. (6), is explicitly reported to fail for β=4 in the intermediate-b region, while the abstract claims unqualified agreement. The fitted coefficient αq is the input to the subsequent verification of Eq. (8), so if the intermediate-b failure is real, the quantitative support for the β=4 extension is thinner than the abstract suggests. The paper does include the qualifier \"for some ranges of the model parameters\" in the conclusions, which partially protects the narrow claim, but the abstract overstates the result and the numerical evidence in the intermediate regime is under-reported: no statistical error bars on Dq, no fit ranges, no data or code. This is an addressable scope/reproducibility problem rather than a reason to reject the model outright, so the conditional verdict stands. A dedicated recomputation with larger system sizes, more realizations, and proper error propagation would settle whether the observed deviations are genuine or finite-size effects. If genuine, the paper should be revised to state explicitly that Eq. (6) is valid only for b≪1 and b≫1 for β=4, and the abstract should be softened accordingly. The reader's universality-class concern is noted but is not the single most load-bearing issue for the paper's own central claim.","tokens_in":17778,"tokens_out":15195,"duration_ms":154088,"concrete_test":"Recompute Dq for β=4 at b = 0.04, 0.1, 0.4, 1, and 2 with N = 2^12 to 2^14 and at least 200 realizations; use jackknife or bootstrap errors and linear fits of ln ⟨Σ|Ψ_i|^2q⟩ versus ln N. Then fit Eq. (6) only to the asymptotic ranges b<0.04 and b>2, and test whether the intermediate-b points fall outside the 2σ band. If they do, the abstract must be weakened and the αq values re-extracted with explicit fit ranges; if they do not, the reported deviation is a finite-size artifact.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the transfer of Eq. (6), Dq ≈ [1+(αq b)^{-1}]^{-1}, from the β=1,2 PBRM cases to the new β=4 model. Sec. 3.2 states that for β=4 the numerical Dq deviate from Eq. (6) in the intermediate range 0.04<b<2, with Dq growing faster than expected; agreement is claimed only for b≪1 and b≫1. Despite this, the αq values in Fig. 1(i) are obtained by fitting Dq(b) to Eq. (6), the abstract says \"good agreement\" without the range restriction, and the subsequent verification of Eq. (8) in Fig. 2 uses these αq values. The reported error bars are reduced rms fit residuals rather than statistical uncertainties, and no fit ranges, raw data, or code are given. Thus the paper's headline claim is not quantitatively supported in the intermediate-b regime, which is a large part of the b axis and the regime where deviations are visually evident. The conclusion's \"for some ranges\" is accurate, but the abstract's unqualified agreement is not. This is an internal correctness and scope problem, independent of whether the β=4 model belongs to the physical symplectic universality class.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a power-law banded random matrix (PBRM) model for the symplectic symmetry class (β=4), corresponding to time-reversal-symmetric systems with strong spin-orbit coupling. The authors study the multifractal dimensions of eigenvectors and the spectral statistics of this model at the putative critical point µ=1, and they compare the results with heuristic relations previously proposed for β=1 and β=2. In particular, they test the relation D_q ≈ [1+(α_q b)^{-1}]^{-1}, the relation between level compressibility and multifractal dimensions, the generalized dimension relations for q<1/2 and q>1/2, and the Nishigaki expressions for the level-spacing distribution. They also present a multifractal analysis and finite-size scaling (MFA-FSS) study in Appendix A to support the claim that the critical point of the β=4 model is µ=1. The paper concludes that the existing heuristic relations also describe the β=4 model 'for some ranges of the model parameters,' a statement that is more cautious than the abstract's unqualified 'good agreement.'","tokens_in":18085,"tokens_out":5739,"duration_ms":51788,"significance":"If the claims are established, the paper provides a numerically accessible random-matrix model for the symplectic Anderson transition and completes the PBRM picture for the three Wigner-Dyson symmetry classes. The main strengths are the introduction of the β=4 PBRM model, the MFA-FSS evidence in Appendix A that µ=1 is critical for β=4, and the systematic review of spectral statistics including ratio distributions and Nishigaki comparisons. The paper also reproduces earlier β=1,2 results, which is useful for completeness. However, the central verification of the heuristic relations is based on fitting free parameters, and the abstract overstates the agreement relative to the deviations reported in the text. These issues make the current version suitable for major revision rather than acceptance.","major_comments":[{"comment":"The abstract claims 'good agreement with heuristic relations for the eigenstate and eigenenergy statistics at criticality' without qualification, but Sec. 3.2 and Fig. 1(f) explicitly state that for β=4 the numerical D_q deviate from Eq. (6) in the intermediate range 0.04 < b < 2, with D_q growing faster than expected; agreement is claimed only for b ≪ 1 and b ≫ 1. This is a load-bearing scope problem because the central claim of the paper is precisely the validity of these relations for β=4. The abstract and conclusion must be aligned: the conclusion's 'for some ranges of the model parameters' is accurate, the abstract is not.","section":"Abstract and Sec. 3.2"},{"comment":"The validation of Eq. (6) for β=4 is underdetermined because α_q is a free parameter fitted to the same D_q(b) data that are then used to claim agreement; the reported error bars are reduced rms fit residuals rather than statistical uncertainties, and no fit ranges, raw data, or code are provided. As a result, the 'agreement' is at best a goodness-of-fit statement, not a predictive test of the relation. The authors should report fit ranges, parameter uncertainties, and a quantitative measure of how much of the b-axis is actually described by Eq. (6) within the stated error bars, especially in the intermediate-b regime where the deviations are visible.","section":"Sec. 3.2, Eq. (6), Fig. 1(d)-(f)"},{"comment":"The verification of Eq. (8) for β=4 is indirect because Eq. (7) provides analytical expressions for χ only for β=1 and β=2; the text states the dots are compared with 'the level compressibility χ given by the analytical expression of equation (7)', but for β=4 there is no such expression. In Fig. 2(c) the β=4 data are shown together with the β=1 and β=2 curves, which does not constitute a quantitative check of Eq. (8) for β=4. The authors should either compute χ directly from the number variance for β=4 or explicitly weaken the claim that Eq. (8) is verified for the symplectic case.","section":"Sec. 4.3, Fig. 2(c)"},{"comment":"The Nishigaki comparison for β=4 is limited to s ≪ 1 and uses the fitted parameter a; the agreement is not demonstrated over the full range of s, as the text acknowledges ('the correspondence between both models in the symplectic case is guaranteed only for s ≪ 1'). Since the abstract claims 'good agreement' for 'eigenenergy statistics', this limitation should be stated in the abstract and conclusion as well, not only in the detailed numerical section.","section":"Sec. 4.3, Fig. 3(l)"}],"minor_comments":[{"comment":"For β=2 and β=4 the asymptotic large-b behavior of the exponent α is fitted as α=2.25−a2/b and α=2.7−a4/b, respectively, which differs from the β=1 form α=2−a/b in Eq. (14); the text notes this but should present it more prominently as an extension of Eq. (14) rather than a confirmation of it.","section":"Sec. 4.3, Eq. (14), Fig. 3(e)-(f)"},{"comment":"The definition of the quaternion-real structure uses 'H = H^R' with H^R called the dual of H, but the dual is not explicitly defined; for readers unfamiliar with quaternion matrices, the relation between H^† and H^R and the Kramers degeneracy would benefit from a brief explanation.","section":"Sec. 2, Eq. (2)"},{"comment":"The sentence 'The reported error bars are the reduced rms error of the fittings between the numerical data and the corresponding analytical expression' conflates the error-bar definition with the goodness-of-fit; a standard statistical uncertainty would be more informative and should be stated explicitly.","section":"Sec. 3.2"},{"comment":"The paper would be strengthened by a data-availability statement and a note on whether the numerical codes are available; as a numerical study, reproducibility is a key part of the contribution.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the scope of Physica A and contains useful numerical results, especially the new β=4 PBRM model and the MFA-FSS evidence for µ=1. The main concern is that the abstract's unqualified 'good agreement' is inconsistent with the deviations reported in Sec. 3.2 and Sec. 4.3; this can be fixed by qualifying the claims and by reporting fit ranges and uncertainties. The Nishigaki comparison for β=4 is also limited, so the conclusions should be correspondingly cautious. I recommend major revision rather than rejection, as the underlying numerical work appears sound and the claims are likely correct when properly scoped."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a workmanlike numerical completion of the PBRM program. The genuinely new ingredient is a β=4 (symplectic) power-law banded random matrix model, and the paper shows that a set of heuristic multifractal and spectral relations, previously tested for β=1 and β=2, also hold for β=4 in the asymptotic limits b≪1 and b≫1. The numerics are carefully done: direct diagonalization over a range of sizes, fixed statistics, and a sensible MFA-FSS analysis supporting μ=1 as the critical point for all three classes.\n\nThe paper does what it sets out to do. It brings the symplectic class into the same framework and reproduces the known results for the other classes, which is useful for anyone wanting a single reference for the three Wigner-Dyson cases. The spectral analysis is thorough: level spacing distributions, ratio statistics, and comparison with Nishigaki's analytic model for small s. I don't see a methodological howler.\n\nThe soft spots are real but not fatal. The abstract says \"good agreement\" without qualification, but Sec. 3.2 openly acknowledges deviations from Eq. (6) in the intermediate range 0.04<b<2 for β=4, where D_q grows faster than expected. The conclusion correctly says \"for some ranges,\" so the abstract is overclaiming. The confirmation of Eq. (6) also relies on fitting α_q to the same data used to verify the relation, which is a mild circularity; the paper would be stronger if it reported fit ranges and gave statistical error bars rather than just reduced rms residuals. No data or code is provided, which makes it hard to check the claimed fits. The comparison with Nishigaki is limited to s≪1, and the paper does not test whether the β=4 PBRM model actually lies in the 2d symplectic Anderson universality class—so the physical relevance is asserted rather than demonstrated. These are addressable points, not fundamental flaws.\n\nWho is this for? People working on critical random matrix ensembles and multifractality who want the β=4 case in the literature. It deserves a serious referee. My recommendation: send it to review, and ask the authors to fix the abstract, include fit ranges and error bars, and make the data available.","headline":"Workmanlike completion of the PBRM program: the new β=4 symplectic ensemble is numerically characterized, but the abstract overstates the agreement in the intermediate-b regime.","tokens_in":18638,"tokens_out":2444,"would_cite":true,"duration_ms":25114,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper introduces a symplectic power-law banded random matrix model and shows that the heuristic multifractal and spectral-statistics relations previously checked for the orthogonal and unitary Wigner-Dyson classes also describe this…","keywords":["power-law banded random matrix","symplectic ensemble","Wigner-Dyson symmetry classes","multifractal dimensions","level compressibility","metal-insulator transition","finite-size scaling","disordered systems"],"falsifier":"Run the $\\beta = 4$ PBRM model at $b = 1$ with system sizes beyond $N = 2^{13}$ and check whether $D_q$ converges to the value predicted by Eq. (6); if the deviation seen in the paper's intermediate range $0.04 < b < 2$ persists or grows, the heuristic relation fails for the symplectic class.","tokens_in":17573,"feed_emoji":"🧲","tokens_out":11987,"duration_ms":107822,"temperature":0.7,"pith_summary":"This paper introduces a power-law banded random matrix model for the symplectic Wigner-Dyson class, the symmetry class of time-reversal-invariant systems with strong spin-orbit coupling. It claims that, at the metal-insulator transition point $\\mu = 1$, the multifractal dimensions $D_q$, the level compressibility $\\chi$, and the nearest-level spacing statistics of this new ensemble obey the same heuristic relations previously verified for the orthogonal and unitary power-law banded random matrix models. The claim matters because it extends a single numerically tractable critical ensemble to all three classical Wigner-Dyson symmetry classes, so the same toolkit can describe systems with and without time-reversal symmetry and with strong spin-orbit coupling. The paper also reproduces the $\\beta = 1$ and $\\beta = 2$ results and reports where the symplectic case deviates, notably for intermediate bandwidths around $0.04 < b < 2$.","feed_headline":"Symplectic random matrices match the same critical statistics","feed_subtitle":"Spin-orbit systems fit the same multifractal and spectral relations at criticality.","key_machinery":"The central object is the periodic power-law banded random matrix (PBRM) ensemble of Eq. (1): random matrices with independent Gaussian entries whose variance decays as a power law in the chord distance on a ring, with exponent $2\\mu$ and effective bandwidth $b$. For $\\beta = 4$ the Hamiltonian is written in quaternion units, making it Hermitian self-dual and preserving the two-fold degeneracy of every eigenvalue required by time-reversal symmetry. The load-bearing identities are the heuristic relations: $D_q \\approx [1 + (\\alpha_q b)^{-1}]^{-1}$, $\\chi \\approx (1-D_q)/(1+(q-1)D_q)$, and the two-branch relation between $D_q$ and $D_1$ given by Eqs. (9) and (10). These relations connect the spatial multifractality of eigenstates to spectral statistics, and the paper's numerical work tests them against the new $\\beta = 4$ ensemble.","core_discovery":"The paper's central claim is that the power-law banded random matrix ensemble, extended to the symplectic class by taking $2N \\times 2N$ Hermitian self-dual quaternion-real matrices with the power-law decaying variance of Eq. (1), has its metal-insulator transition at $\\mu = 1$, and that at this critical point its eigenstate and eigenenergy statistics follow the same heuristic relations as the $\\beta = 1$ and $\\beta = 2$ cases. Numerically, the multifractal dimensions extracted from the scaling of inverse participation numbers agree with $D_q \\approx [1 + (\\alpha_q b)^{-1}]^{-1}$ in the limits $b \\ll 1$ and $b \\gg 1$, and the relations connecting $D_q$ to the information dimension $D_1$ and to the level compressibility $\\chi$ hold across the studied range of $q$. The level spacing distribution shows the expected Poisson-to-Wigner-Dyson crossover in $b$, and in the small-$s$ regime it is consistent with the analytical critical estimates of Refs. [58,59]. The paper notes a genuine limitation: for the symplectic case, $D_q$ grows faster than Eq. (6) predicts in the intermediate bandwidth range $0.04 < b < 2$.","pith_inferences":["If the $\\beta = 4$ PBRM model lies in the universality class of the two-dimensional symplectic metal-insulator transition, then the same heuristic relations should appear in tight-binding models with spin-orbit coupling; testing that directly would either confirm the transfer or expose a non-universal feature of the PBRM construction.","The intermediate-$b$ deviation could be a sign that the symplectic class has a wider crossover region rather than a failure of the one-parameter heuristic; a finite-size study of $D_q$ at fixed $b$ would show whether the deviation shrinks with system size.","The fitted exponent $\\alpha = 2.7 - a_4/b$ for large $b$ is presented as a fitting parameter; if it can be connected to the correlation-length exponent of the symplectic transition, it would turn an empirical curve into a scaling prediction."],"forward_implications":["The $\\beta = 4$ PBRM ensemble gives the third Wigner-Dyson class a numerically tractable critical random-matrix model, so spin-orbit disordered systems can be studied with the same tools as orthogonal and unitary ones.","The validity of Eqs. (8)-(10) across all three symmetry classes indicates that the connection between eigenstate multifractality and spectral compressibility at criticality does not depend on the Dyson symmetry index.","The deviations from Eq. (6) at $0.04 < b < 2$ mark a concrete parameter window where the symplectic class needs a modified heuristic, not just a re-fit of $\\alpha_q$.","The finite-size scaling result that $\\mu = 1$ is critical for $\\beta = 4$ means the localization-delocalization transition occurs at the same power-law exponent as in the orthogonal and unitary PBRM models.","The large-$b$ level spacing exponent for $\\beta = 4$ fits $\\alpha = 2.7 - a_4/b$ with $a_4 = 1.05$, extending the stretched-exponential form of the critical spacing distribution to the symplectic class."],"supporting_citations":[{"why":"Supplies the original PBRM construction for $\\beta = 1$ that the symplectic version extends.","marker":"[22]"},{"why":"Proposes and verifies the heuristic $D_q$ relation of Eq. (6) for the $\\beta = 1$ PBRM model.","marker":"[25]"},{"why":"Proposes and verifies the relations between $D_q$, $D_1$, and level compressibility for the $\\beta = 2$ PBRM model.","marker":"[26]"},{"why":"Establishes the PBRM model as a critical ensemble for metal-insulator transitions, including the $\\mu = 1$ critical point and the role of bandwidth.","marker":"[10]"},{"why":"Provides the numerically verified asymptotic level spacing form $P_c(s) \\sim \\exp(-A s^\\alpha)$ for $\\beta = 1$ that the paper extends to $\\beta = 2$ and 4.","marker":"[57]"},{"why":"Gives the analytical critical level spacing estimates used for the small-$s$ comparison.","marker":"[58]"},{"why":"Validates those analytical estimates against three symmetry classes of tight-binding models.","marker":"[59]"},{"why":"Provides the closed-form ratio distribution $P_{WD}(r)$ used for the spacing-ratio statistics.","marker":"[61]"},{"why":"Supplies the quaternion-real matrix structure and the two-fold degeneracy used in the $\\beta = 4$ construction.","marker":"[15]"}],"fun_headline_variants":["Symplectic class completes critical random matrix picture","Spin-orbit criticality joins universal Wigner-Dyson stats","All three Wigner-Dyson classes share critical multifractality","Power-law banded model now covers symplectic ensemble","Critical stats hold for spin-orbit, but with a caveat"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the new quaternion-based random matrix ensemble belongs to the same universality class as the two-dimensional symplectic metal-insulator transition, since the paper locates its own critical point by finite-size scaling rather than by matching a lattice spin-orbit model.","fun_headline_variants_meta":{"raw":{"variants":["Symplectic class completes critical random matrix picture","Spin-orbit criticality joins universal Wigner-Dyson stats","All three Wigner-Dyson classes share critical multifractality","Power-law banded model now covers symplectic ensemble","Critical stats hold for spin-orbit, but with a caveat"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000696,"raw_usage":{"total_tokens":3153,"prompt_tokens":959,"completion_tokens":2194,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":575,"completion_tokens_details":{"reasoning_tokens":2111}},"tokens_in":575,"tokens_out":2194,"duration_ms":15190,"temperature":1.0,"reasoning_tokens":2111,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:42:19.049247+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the $\\beta = 4$ PBRM model at $b = 1$ with system sizes beyond $N = 2^{13}$ and check whether $D_q$ converges to the value predicted by Eq. (6); if the deviation seen in the paper's intermediate range $0.04 < b < 2$ persists or grows, the heuristic relation fails for the symplectic class.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proposes and verifies the heuristic $D_q$ relation of Eq. (6) for the $\\beta = 1$ PBRM model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proposes and verifies the relations between $D_q$, $D_1$, and level compressibility for the $\\beta = 2$ PBRM model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the analytical critical level spacing estimates used for the small-$s$ comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Validates those analytical estimates against three symmetry classes of tight-binding models."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the quaternion-real matrix structure and the two-fold degeneracy used in the $\\beta = 4$ construction."}],"review_version":1}