{"id":"eeafdd62-f36e-49b8-be51-9cbaef526a50","arxiv_id":"2508.20378","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"The distribution of lattice QCD nuclear correlation function samples is well described by O(N) model distributions with fitted N approximately 2/B at large time separations.","lead":"This paper compares the random-sample distributions of nuclear correlation functions in lattice QCD to a simple model distribution from the O(N) field theory. It finds the model describes proton, deuteron, triton, and helium-4 data well, with a fitted parameter N that scales roughly as 2 divided by the baryon number.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Alpha fit table (Table V) duplicates deuteron table (Table III), so the B=4 anchor of the central N~2/B scaling has no verifiable tabulated support; robustness of fitted N to histogram truncation is also unquantified.","rationale":"The paper's qualitative finding that O(N) distributions capture LQCD histograms is plausible and useful, and the use of pre-existing NPLQCD data is real evidence. However, the headline inverse-N scaling is the novel claim, and it is extracted from the same fitted parameters that are claimed to reproduce the data. The most concrete defect is that the alpha fit table in the appendix is identical to the deuteron table; this is exactly the sort of missing support the review should flag. Since Table I's alpha value is obtained from fits to alpha N_B(t), and the tabulated alpha N_B(t) is not available (or is wrong), the B=4 point is unverifiable. The histogram construction issue (10% truncation, 1600 bins, free alpha normalization) is a second-order concern because it could bias N in a B-dependent way, but the paper's footnote is qualitative. A numerical scan would settle it. I do not see grounds to reject the entire paper: the distribution comparison has visual support and the 2/B trend is also present for B=1-3. But acceptance should be conditional on correcting the alpha table and demonstrating that N_B(infinity) is stable under reasonable binning choices.","tokens_in":25672,"tokens_out":7191,"duration_ms":89634,"concrete_test":"Using the original NPLQCD samples and the authors' fitting code (or an independent reimplementation), recompute the B=4 (alpha) N_B(t) table with a (Delta_cut, N_bins) grid, e.g., Delta_cut in {1%,5%,10%,20%} and N_bins in {400,1600,6400}; then fit N_alpha(t->infinity) by the same Sec. III C procedure and compare with Table I (0.433) and Table V. In parallel, verify whether the published Table V entries are genuinely alpha results rather than deuteron duplicates. If the corrected N_alpha(infinity) departs from 0.5 by more than the quoted uncertainty, or if Table V is confirmed to be a copy of Table III, the central N~2/B scaling needs to be re-evaluated at B=4.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim N_B(t->infinity) ~ 2/B rests on four asymptotic values, with B=4 (alpha) being the point that makes the inverse-linear relation well resolved. Table V, labelled 'Fit parameters for the alpha', is numerically identical to Table III for the deuteron (e.g., t=1: N=2.49, t=48: N=1.053). This internal inconsistency removes the paper's tabulated support for the alpha fits. Either the table was copied in error or the alpha N_B(t) values shown in Fig. 5 are not the ones tabulated; in both cases the reader cannot check N_alpha(infinity)=0.433. The problem is compounded by Sec. III B: histograms are built after removing the lowest and highest 10% of samples, and the fit includes a free normalization alpha in Eq. (8). Footnote 1 gives only a qualitative claim of robustness with no systematic scan. Since the asymptotic N is extracted from these fitted N_B(t) values, any binning/truncation sensitivity would propagate directly into the 2/B relation. The paper's qualitative residual criterion (<0.3 in most bins) is too loose to discriminate whether the fitted N is physically meaningful or absorbs shape mismatch. Therefore the B=4 anchor of the scaling claim is not currently supported by the manuscript as written.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes the empirical probability distributions of zero-momentum two-point correlation functions computed in lattice QCD for systems with baryon number B = 1, 2, 3, 4 (proton, deuteron, triton, alpha). Using a single NPLQCD ensemble, the authors build histograms of Monte-Carlo samples and fit them to the O(N)-model distribution of Eq. (2), with parameters N, omega_+, omega_-. They report that the O(N) form describes the QCD histograms well across all temporal separations, and that the extracted parameter N_B(t) saturates at large t to values approximately scaling as 2/B. The paper also analyzes phase distributions via von Mises fits and speculates on a connection to O(N -> 0) self-avoiding walk behavior.","tokens_in":26179,"tokens_out":2538,"duration_ms":34049,"significance":"If the central empirical claim is robust, this is an intriguing and potentially useful observation: it would show that the noise distributions of QCD correlation functions in the multi-baryon sector are quantitatively captured by a one-parameter family of O(N) distributions, and it would suggest a specific, testable scaling N ~ 2/B at large B. The paper uses genuine lattice data and presents a new comparison that goes beyond previous signal-to-noise analyses. The strengths include the explicit presentation of the fitted distributions in extensive appendix figures, the use of phase distributions as a secondary check, and a clearly stated, falsifiable scaling law. However, the current manuscript contains an internal inconsistency in the tabulated alpha fits, and the statistical acceptance criteria are not stringent enough to fully support the 'accurate description' claim.","major_comments":[{"comment":"Table V, labelled 'Fit parameters for the alpha,' is numerically identical to Table III for the deuteron for every entry (e.g., t=1: N=2.49, omega_+=1.43; t=48: N=1.053). This makes the tabulated support for the alpha fits unverifiable. Since the B=4 column in Fig. 5 and the N_alpha(infinity)=0.433 value in Table I are the anchor that makes the N^{-1} = c0 + c1 B relation well resolved, this duplication is load-bearing. The authors must provide the correct alpha fit tables, or at minimum state explicitly that the alpha results shown in Fig. 5 are not tabulated and why.","section":"Appendix A, Tables III and V"},{"comment":"The fitting methodology removes the lowest and highest 10% of samples before histogramming and then includes a free normalization alpha in Eq. (8). This is internally inconsistent with the later statement in Sec. III B that 'alpha is not a free parameter but is set such that the model and empirical data are normalised equivalently.' If alpha is indeed free, the shape comparison is weakened because alpha absorbs overall normalization mismatch and can shift the best-fit N; if alpha is fixed, then Eq. (8) as written is not the actual loss function. The footnote's qualitative claim of robustness to Delta_cut and Nbins is not a substitute for a systematic scan, especially because the extracted N_B(t) values feed directly into the central N ~ 2/B result. Please clarify the role of alpha and provide a quantitative robustness study for the truncation/binning choices, at least for the asymptotic","section":"Sec. III B, Eq. (8) and footnote 1"},{"comment":"The acceptance criterion for the fits is |delta_{B,t}(x)| < 0.3 in 'most bins' (footnote 4), and the reported average residual is typically below 0.2. This is a weak goodness-of-fit measure, particularly because the bins are highly correlated and the zero-bin singularity is excluded from the criterion. The claim that the O(N) distribution 'accurately describes' the QCD distributions over all t deserves a more quantitative test, such as a chi-square per degree-of-freedom, a Kolmogorov-Smirnov test, or a likelihood-based comparison, with the number of effective degrees of freedom clearly stated. Without such a test, the fitted N values could be absorbing systematic shape mismatches rather than representing a physically meaningful parameter.","section":"Sec. III B and Fig. 3"},{"comment":"The extraction of N_B(infinity) for B=alpha uses only a constant fit over t in [25,48], because the time dependence is noisy. This is reasonable, but it makes the B=4 point particularly sensitive to the quality of the underlying histograms and to the binning/truncation choices. Given the duplication in Table V and the lack of a systematic binning study, the B=4 value of 0.433 is currently not independently supported. Please either provide the correct alpha table, or present the alpha fits in a form that allows the reader to verify N_alpha(t -> infinity).","section":"Sec. III C, Table I"}],"minor_comments":[{"comment":"Only 14 bootstrap resamples are used to estimate uncertainties. While the paper notes that doubling the number of bootstraps gives negligible differences, the 17%/83% quantiles from 14 samples are noisy. It would be helpful to state the seed/protocol and, if possible, increase to at least 100 resamples for the final fits.","section":"Sec. III B, step 3"},{"comment":"The lower panel of Fig. 5 shows log10(omega_+/-), but the text describes the parameters as omega_+ and omega_-. The axis label should be explicit about which parameter is plotted and in what units; currently it is easy to misread the two curves.","section":"Sec. III C, Fig. 5"},{"comment":"The phase analysis using von Mises distributions is a nice complement, but the R^2 values >0.99 for t>=4 should be supplemented by the fitted kappa uncertainties or a residual plot. The current Fig. 7 shows only the central values.","section":"Sec. III D"},{"comment":"The paper would benefit from a short description of how the correlation function signs are handled in the histogram: the text says the real part is used, but the effective mass discussion refers to 'absolute values' and 'phase of the correlation function.' Clarify exactly what quantity is histogrammed for the zero-bin region.","section":"General"},{"comment":"The tables are large and the uncertainty notation is compact. A note defining the 17%/83% quantile convention (as in Table I) would help readers interpret the many asymmetric uncertainties.","section":"Captions of Tables III-V"}],"recommendation":"major_revision","confidential_remarks":"The duplication between Table V and Table III is almost certainly a copy-paste error, but it is a critical one: it removes the verifiable tabulated support for the B=4 anchor of the central N ~ 2/B scaling. The other concerns (free alpha, 10% truncation, weak GOF criterion) are fixable but need to be addressed carefully. I would advise the editor not to accept the paper until the alpha table is corrected and a systematic robustness study is provided. The paper's central idea is interesting and worth publishing, but the current form does not yet allow the reader to check the main result."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick read: the paper does something genuinely useful—it takes the O(N)-model correlation-function distribution from Yunus & Detmold and actually compares it to empirical lattice QCD histograms for baryon number B=1..4. The fit quality is impressive: the model captures the shape of the distributions across t, the residuals are small, and the phase analysis with von Mises fits is a nice extra. That descriptive result alone is worth knowing, and the comparison to earlier proposed distributions is fair and useful.\n\nThe central claim, however, is the N(t->infty) ~ 2/B scaling, and here the manuscript has a serious problem. Table V, labelled 'Fit parameters for the alpha', is numerically identical to Table III for the deuteron. I checked: every entry matches, including N=2.49 at t=1 and N=1.053 at t=48. So the paper does not actually contain the alpha fit parameters shown in Fig. 5. The B=4 point is exactly what makes the inverse-linear relation look well resolved, and it is currently unsupported by the tabulated data. Either the table was copied in error or the figure uses a different set of fits; either way, a referee cannot verify the anchor of the scaling claim.\n\nThere is also a smaller internal inconsistency about the normalization alpha. Section IIIB says alpha is determined in the L1 fit (Eq. 8), but then later says alpha is not a free parameter and is set by equivalent normalization. Those two statements cannot both be true. If alpha is free, shape mismatch can be absorbed, which weakens the interpretation of the fitted N. The authors need to clarify which procedure they actually used.\n\nThe acceptance criterion (|residual| < 0.3 in most bins) is loose, and the robustness of the 10% truncation and 1600-bin choice is only qualitatively asserted in footnote 1. These are not fatal on their own, but they matter because the asymptotic N is extracted from the same histograms that are being fit. None of this kills the qualitative claim that the O(N) distribution describes the QCD histograms—that seems robust from the figures—but the N~2/B relation is a fitted scaling, not an independent check, and its B=4 value is presently unverifiable.\n\nThe paper deserves peer review: the empirical observation is interesting, the errors are fixable, and the field would benefit from seeing the corrected version. But it should not be accepted as is; the Table V duplication in particular must be fixed and the alpha normalization clarified.","headline":"The O(N) distribution really does describe these QCD histograms, but the B=4 anchor of the N~2/B scaling is not verifiable as written because the alpha fit table duplicates the deuteron table.","tokens_in":26571,"tokens_out":3592,"would_cite":false,"duration_ms":41330,"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":"Lattice QCD correlation functions for nuclei follow an O(N)-model distribution, with effective N≈2/B.","keywords":["lattice QCD","nuclear correlation functions","O(N) model","signal-to-noise problem","baryon number","Monte Carlo distributions","two-point correlation functions","von Mises distribution"],"falsifier":"Discard the 10% tail cut or fix the histogram normalization to unit integral and refit N_B(t) at late t; the 2/B claim fails if the asymptotic N values change by more than the quoted 17–83% bootstrap uncertainties, or if normalized residuals grow above the reported 0.2 level.","tokens_in":25596,"feed_emoji":"⚛️","tokens_out":11319,"duration_ms":114054,"temperature":0.7,"pith_summary":"Lattice QCD correlation functions for the proton, deuteron, triton, and helium-4 are dominated at large times by Monte-Carlo noise, but this paper finds that the full probability distribution of that noise has a simple shape. The paper fits the empirically sampled histograms of the zero-momentum two-point correlation functions to the analytic distribution derived for the O(N) model—a field theory with N equivalent components—and reports that the fit works for all four baryon numbers at every studied temporal separation, with typical normalized residuals below 0.2. The fitted value of the parameter N is non-integer and, at large separations, asymptotes to about 2/B, so a deuteron behaves like an O(1) distribution, a triton like O(0.6), and an alpha like O(0.4). If the result holds, larger nuclei would push the effective description toward the O(N→0) limit, whose statistical mechanics is connected to self-avoiding random walks.","feed_headline":"Nuclear correlation functions match O(N) model, N≈2/B","feed_subtitle":"Proton, deuteron, triton and helium-4 samples fit one distribution whose N falls as baryon number grows.","key_machinery":"The object doing the work is the analytic distribution P(x;ω+,ω−,N) of Eq. (2) for the zero-momentum O(N)-invariant two-point function, written in terms of a modified Bessel function of the second kind. It has two scale parameters ω± that become equal at late times and an effective field count N that controls the shape near x=0 and the tails. The paper turns this into a fitting template: Monte-Carlo samples are concatenated, the lowest and highest 10% are removed, a 1600-bin histogram is built, and an L1-norm fit with a free normalization extracts N and ω± for each B and t. The extracted N is the quantity that carries the result—its large-time value is the 2/B scaling.","core_discovery":"The central claim is that the real part of the zero-momentum QCD two-point correlation function, sampled by lattice Monte Carlo, obeys the same probability law as the simplest two-point function of the O(N) model. For each baryon number B∈{1,2,3,4} and each time separation, the empirical histogram is fit to Eq. (2), the analytic distribution with parameters N, ω+, and ω−, and the fits reproduce the histogram's shape and magnitude at all studied times. The paper's second claim is that the best-fit N at large time saturates to N≈2/B, while N(t) approaches that value exponentially with a scale near the pion mass or the QCD scale. The phase of the correlation function is separately described by","pith_inferences":["Testable extension: apply the same fit to meson (B=0) correlation functions; if N≈2/B continues to B=0, the effective N would diverge or lose meaning, which would distinguish the scaling from an accidental fit.","The results come from a single lattice ensemble at mπ≈450 MeV; a check at physical quark masses and different lattice spacings would show whether N≈2/B is a universal feature of QCD or an artifact of this ensemble.","The fitted normalization and 10% tail cut are analysis choices that could affect N; an unbinned likelihood fit on the complete sample would either confirm or weaken the 2/B relation.","Moment ratios (for example variance and kurtosis) implied by the asymptotic O(N) distribution could be compared directly with the raw samples, avoiding histogram resolution entirely."],"forward_implications":["If the O(N) form is correct, the Monte-Carlo noise in nuclear correlation functions has a known analytic model, so estimators for the underlying path integrals can be designed using the distribution rather than only the mean.","The 2/B asymptotic makes multi-baryon correlators approach the O(N→0) model as B grows, linking nuclear correlation-function statistics to the self-avoiding-walk universality class.","The exponential approach of N_B(t) to its asymptote gives a data-driven way to identify when the signal-to-noise problem has entered its asymptotic regime for each baryon number.","Because the distributions broaden and become strongly non-Gaussian with t, any error propagation for nuclear correlators should use the full distribution or its fitted form."],"supporting_citations":[{"why":"Derives the analytic O(N)-model distribution in Eq. (2) from correlated mean-field arguments; this is the functional form fitted to all the QCD data.","marker":"[18]"},{"why":"Earlier work establishing correlation-function distributions in scalar-field and O(N)-model settings, which motivates applying these forms to QCD.","marker":"[17]"},{"why":"Generalises the Parisi-Lepage signal-to-noise heuristic to baryon number B, defining the exponential signal-to-noise problem that the distribution analysis clarifies.","marker":"[3]"},{"why":"Provides the lattice-QCD ensemble and correlation-function samples analysed in this work.","marker":"[8]"},{"why":"Previous statistical study of QCD baryon correlation-function phase distributions; supplies a baseline and the von Mises form used for the phase analysis.","marker":"[13]"},{"why":"Proposes alternative candidate distributions for QCD correlation functions that the paper finds less accurate than the O(N) form.","marker":"[14]"}],"fun_headline_variants":["QCD correlation functions match O(N) with N≈2/B","Lattice QCD fits O(N) statistics: N scales as 2/B","Nuclear correlation distributions follow O(N), N≈2/B","Baryon correlations obey O(N) law, N≈2/B from fits"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The load-bearing premise is that the empirical histogram produced by trimming 10% of samples, binning into 1600 bins, and allowing a free normalization in the L1 fit preserves the true baryon-dependent shape of the correlation-function distribution; if the trimmed tails or the fitted normalization carry that dependence, the O(N) descriptions and the N≈2/B scaling could be artifacts.","fun_headline_variants_meta":{"raw":{"variants":["QCD correlation functions match O(N) with N≈2/B","Lattice QCD fits O(N) statistics: N scales as 2/B","Nuclear correlation distributions follow O(N), N≈2/B","Baryon correlations obey O(N) law, N≈2/B from fits"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000201,"raw_usage":{"total_tokens":1207,"prompt_tokens":724,"completion_tokens":483,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":468,"completion_tokens_details":{"reasoning_tokens":404}},"tokens_in":468,"tokens_out":483,"duration_ms":5597,"temperature":1.0,"reasoning_tokens":404,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T15:05:15.722342+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Discard the 10% tail cut or fix the histogram normalization to unit integral and refit N_B(t) at late t; the 2/B claim fails if the asymptotic N values change by more than the quoted 17–83% bootstrap uncertainties, or if normalized residuals grow above the reported 0.2 level.","supporting_citations":[],"review_version":1}