{"id":"96fc4fae-8374-4ccc-bad3-01d6b8802ad0","arxiv_id":"2607.00587","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Variance of second-order statistics for both hyperuniform determinantal and non-hyperuniform Gibbs point processes grows proportionally to ball volume, showing generic non-hyperuniformity.","lead":"The paper proves that for determinantal point processes and Gibbs processes with superstable interactions, the variance of second-order statistics inside a growing ball grows linearly with the ball's volume. This implies that these processes exhibit non-hyperuniform behavior in their second-order statistics, with an application to the inverse Henderson problem in statistical mechanics.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.3","headline":"Simple zero at origin in determinantal Bartlett measure should yield o(vol) variance, yet claim asserts ~vol scaling for both classes","rationale":"Reader's weakest assumption concerns membership in the stated classes; the load-bearing issue is an apparent mismatch between the reported spectral zero and the claimed volume scaling, which is internal to the argument rather than a boundary-condition question.","tokens_in":1685,"tokens_out":337,"duration_ms":23923,"concrete_test":"Extract the precise definition of the second-order statistic and its spectral measure from §3–4; recompute the small-k integral of S(k)|χ̂_R(k)|^2 using the stated simple-zero behavior and check whether the leading term remains Θ(vol(B_R)) or drops to o(vol(B_R)).","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The abstract states that the Bartlett spectral measure for determinantal projection-kernel processes has a simple zero at the origin while remaining positive elsewhere, yet asserts that the variance of the second-order statistic inside B_R is asymptotically proportional to vol(B_R) for both determinantal and Gibbs cases. In the standard Fourier representation of such variances, Var = ∫ S(k) |χ̂_R(k)|^2 dk, a zero of order 1 at k=0 produces an extra factor R^{-1} (or higher) after rescaling, yielding growth strictly slower than volume. The paper's claim therefore requires either that the second-order statistic is defined so the zero does not affect the leading term, or that the generic conditions override the zero; neither is visible from the stated spectral properties.","agreement_with_reader":"disagree"},"referee_report":{"model":"grok-4.3","summary":"The paper claims that for stationary determinantal point processes with projection kernels and for Gibbs point processes with superstable pair interactions, the variance of the second-order statistics (within a ball B_R centered at the origin) grows asymptotically proportionally to vol(B_R) as R → ∞, implying non-hyperuniform behavior of these second-order statistics in both classes. It further asserts that the structure factor (Bartlett spectral measure) is strictly positive for the Gibbs case and positive except for a simple zero at the origin for the determinantal case, with an application to the inverse Henderson problem.","tokens_in":1852,"tokens_out":378,"duration_ms":25481,"significance":"If the central claims on variance asymptotics hold, the work would provide a clear separation between hyperuniformity of first-order versus second-order statistics and furnish a concrete application to statistical mechanics via the inverse Henderson problem. The positivity results for the spectral measures would also be of independent interest in the theory of point processes.","major_comments":[{"comment":"Abstract: the claim that the variance of the second-order statistics inside B_R is asymptotically proportional to vol(B_R) for determinantal projection-kernel processes is in tension with the stated property that the Bartlett spectral measure has a simple zero at the origin. In the Fourier representation Var = ∫ S(k) |χ̂_R(k)|^2 dk, a simple zero S(k) ∼ |k| near k=0 produces, after the change of variables k = u/R, an extra factor R^{-1} and therefore growth of order vol(B_R)/R rather than vol(B_R). This scaling issue is load-bearing for the central non-hyperuniformity claim that is asserted for both classes.","section":"Abstract"}],"minor_comments":[],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for their detailed reading and for highlighting the scaling tension in the abstract. We address the comment below and will make the necessary revisions.","responses":[{"response":"We agree with the referee's Fourier analysis. The simple zero S(k) ∼ |k| at the origin does indeed produce the extra R^{-1} factor under the change of variables, yielding asymptotic growth of order vol(B_R)/R for the determinantal processes rather than vol(B_R). The abstract statement that the variance is asymptotically proportional to vol(B_R) for both classes is therefore imprecise for the determinantal case. We will revise the abstract, the introduction, and the relevant theorem statements to distinguish the two classes: for Gibbs processes with superstable interactions the variance grows proportionally to vol(B_R), while for determinantal projection-kernel processes it grows as vol(B_R)/R. We will also adjust the non-hyperuniformity phrasing to reflect this distinction accurately while preserving the application to the inverse Henderson problem for the Gibbs case.","revision_made":"yes","referee_comment":"[Abstract] Abstract: the claim that the variance of the second-order statistics inside B_R is asymptotically proportional to vol(B_R) for determinantal projection-kernel processes is in tension with the stated property that the Bartlett spectral measure has a simple zero at the origin. In the Fourier representation Var = ∫ S(k) |χ̂_R(k)|^2 dk, a simple zero S(k) ∼ |k| near k=0 produces, after the change of variables k = u/R, an extra factor R^{-1} and therefore growth of order vol(B_R)/R rather than vol(B_R). This scaling issue is load-bearing for the central non-hyperuniformity claim that is asserted for both classes."}],"tokens_in":1323,"tokens_out":385,"duration_ms":24457,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main point is that the variance of second-order statistics inside a growing ball scales like the volume for both families. This holds even for the determinantal processes that are hyperuniform at the first-order level.\n\nThe paper works out the asymptotic variance result under the stated conditions on kernels and pair interactions. It also establishes that the structure factor or Bartlett measure is positive (strictly for Gibbs, except for a simple zero at the origin for determinantal). The Gibbs side is tied to the inverse Henderson problem in statistical mechanics.\n\nThis pins down scaling behavior that was not previously recorded for these exact classes. The arguments rest on standard properties of determinantal and Gibbs processes rather than on fitted quantities.\n\nThe soft spot is the simple zero at the origin in the determinantal Bartlett measure. In the usual Fourier representation of the variance, a zero of order one typically produces sub-volume growth after integration against the squared window transform. The abstract asserts volume proportionality anyway, so the proofs must show why the zero does not control the leading term or why the generic conditions override it. That step needs to be explicit.\n\nThe work is for people already working on hyperuniformity, fluctuation scaling, and stationary point processes. A reader focused on the inverse Henderson problem might use the Gibbs part.\n\nThe paper engages honestly with the literature on these processes. It deserves a serious referee to check the derivations and the handling of the spectral zero.","headline":"The paper claims that second-order statistics have volume-proportional variance for both determinantal projection-kernel processes and superstable Gibbs processes, making them non-hyperuniform in both cases, with a side application to the inverse Henderson problem.","tokens_in":2318,"tokens_out":379,"would_cite":false,"duration_ms":19872,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Variance of second-order statistics in both hyperuniform determinantal and non-hyperuniform Gibbs point processes grows proportionally to ball volume.","keywords":["point processes","hyperuniformity","determinantal processes","Gibbs point processes","variance asymptotics","second order statistics","structure factor","inverse Henderson problem"],"falsifier":"For the Ginibre determinantal process in the plane, compute or simulate the variance of the number of pairs inside a large disk and test whether the growth is strictly linear in area or slower.","tokens_in":2574,"feed_emoji":"📊","tokens_out":635,"duration_ms":15041,"temperature":0.7,"pith_summary":"The paper studies the asymptotic variance, as ball radius tends to infinity, of second-order statistics for two families of stationary point processes: determinantal processes whose kernels are projections, and Gibbs processes with superstable pair interactions. It proves that under generic conditions on the kernel or the potential, this variance is asymptotically a positive constant times the volume of the ball. A sympathetic reader cares because first-order hyperuniformity (suppressed fluctuations in point counts) does not automatically extend to second-order measures; the result supplies a concrete counter-example class and an application to the inverse Henderson problem of recovering interactions from pair correlations.","feed_headline":"Second-order stats variance scales with ball volume in point processes","feed_subtitle":"Both determinantal and Gibbs examples show fluctuations proportional to volume rather than slower, even when first-order counts are hyperuni","key_machinery":"Asymptotic analysis of the variance of second-order statistics, showing volume proportionality under generic conditions on the projection kernel or the superstable interaction potential.","core_discovery":"Generically, for determinantal point processes with projection kernels and for Gibbs point processes with superstable pair interactions, the variance of the second-order statistics inside a ball centered at the origin is asymptotically proportional to the volume of that ball; the second-order statistics therefore behave non-hyperuniform. The structure factor (or Bartlett spectral measure) of the Gibbs processes is strictly positive, while for the determinantal processes it is positive except for a simple zero at the origin.","pith_inferences":["Hyperuniformity of point counts does not force hyperuniformity of pair-count fluctuations inside the same families.","The volume scaling may serve as a diagnostic to distinguish first-order from second-order hyperuniformity in other stationary point processes."],"forward_implications":["The structure factor of Gibbs processes with superstable interactions is strictly positive everywhere.","For determinantal processes the structure factor vanishes only at the origin and is positive elsewhere.","The non-hyperuniform scaling supplies a concrete obstruction for the inverse Henderson problem when the target pair correlation comes from such a process."],"fun_headline_variants":["Second-order variance scales with ball volume","Non-hyperuniform second-order stats variance","Second-order stats variance linear in ball volume","Point processes show volume-proportional second-order variance"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The processes are determinantal with projection kernels or Gibbs with superstable pair interactions, and the kernel or potential satisfies the generic conditions stated in the theorems.","fun_headline_variants_meta":{"raw":{"variants":["Second-order variance scales with ball volume","Non-hyperuniform second-order stats variance","Second-order stats variance linear in ball volume","Point processes show volume-proportional second-order variance"]},"model":"grok-4.3","cost_usd":0.00448,"raw_usage":{"total_tokens":2215,"prompt_tokens":631,"num_sources_used":0,"completion_tokens":53,"cost_in_usd_ticks":44799500,"prompt_tokens_details":{"text_tokens":631,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":1531,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":631,"tokens_out":53,"duration_ms":11508,"temperature":1.0,"reasoning_tokens":1531,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-02T07:26:13.256291+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"For the Ginibre determinantal process in the plane, compute or simulate the variance of the number of pairs inside a large disk and test whether the growth is strictly linear in area or slower.","supporting_citations":[],"review_version":1}