{"id":"a971b2ab-6f7a-46a7-bbe3-74017b5cd6fb","arxiv_id":"2505.23491","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"The paper claims charged particle multiplicity distributions follow from maximizing Shannon entropy with a fixed mean, yielding exponential tails and, after Poisson-Gamma mixing, the negative binomial distribution.","lead":"This paper applies the principle of maximum entropy to charged particle multiplicity data, claiming that the measured exponential and negative binomial shapes follow from only normalization and a fixed average count. A broad reader might care because it offers a statistical-inference story for a widely used empirical distribution and connects final-state multiplicity to parton entanglement entropy.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The exponential 'derivation' depends on choosing continuous support and importing the mean from data; with the actual discrete support, POME yields a geometric distribution, so the claimed assumption-free derivation does not hold.","rationale":"The reader's weakest assumption correctly identifies the continuous-support and imported-mean premise as the load-bearing point, and I agree that this premise fails to support the abstract's strong claim. My analysis sharpens the concern: the choice of continuous support is not merely a technical convenience. It changes the mathematical result. Under the same normalization and mean constraints, the discrete max-entropy distribution is geometric, not exponential. Therefore the paper's Eq. (5) is not a consequence of POME alone; it is a consequence of POME plus a particular, unjustified modeling choice. Since the later Gamma/NBD argument in Sec. 4 starts from convolutions of these exponentials, the claimed 'natural explanation' of the NBD shape inherits the same unsupported continuum assumption. The n < 3 exclusion in Fig. 1 further weakens the empirical support by omitting the low-multiplicity region where the discrete and continuous predictions differ most. Separately, Eq. (3) is not a valid variational result on an infinite interval, so the initial-state uniform distribution is not derived from a well-posed entropy maximization. These issues are internal to the paper's argument, not merely disagreements with prior literature. The paper could be rehabilitated by treating the multiplicity as discrete, fitting over the full range, and either deriving or carefully testing the relation k = N. As written, however, the central claim of an assumption-free derivation is not supported, and the reader's REJECT verdict is appropriate. My stress-test therefore does not change the verdict.","tokens_in":4248,"tokens_out":4116,"duration_ms":40664,"concrete_test":"Re-derive Eq. (5) with the support n ∈ {0,1,2,...} rather than [0,∞), keeping only normalization and mean μ as constraints. The max-entropy solution is then the geometric distribution p(n) = (1/(1+μ)) (μ/(1+μ))^n. Compare this geometric distribution to the same LHCb data used in Fig. 1 over the full n range, including n < 3, using the same fitting procedure and χ²/NDF. If the geometric form fits at least as well as the exponential with the same μ and no additional cut, the exponential and the resulting Gamma/NBD convolution are consequences of the chosen continuous support, not of POME alone. If the geometric fit is dramatically worse, the n < 3 cut and continuum approximation are load-bearing and need explicit justification.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim (abstract; Sec. 2.2) is that the exponential multiplicity distribution follows from POME 'without any a priori physical assumption.' This is not supported by the derivation in Eqs. (2)-(5). The support [0,∞) and the mean constraint are themselves assumptions; the mean μ is not derived but fitted to the data in Sec. 3. More decisively, the continuous support is not innocent: charged multiplicities are integers. If one maximizes Shannon entropy on the actual support n ∈ {0,1,2,...} subject to normalization and mean μ, the unique solution is the geometric distribution p(n) = (1/(1+μ)) (μ/(1+μ))^n, not the exponential p(n) = λ e^{-λ n} of Eq. (5). The exponential form can only be recovered as an approximation, and the paper provides no error estimate or justification for it. The comparison in Fig. 1 also excludes n < 3, removing the region where the discrete and continuous forms differ most; this is an extra a priori assumption. Because Sec. 4 builds the Gamma/NBD argument on convolution of exponentials, that extension inherits the unsupported continuum assumption. Additionally, the normalization-only maximization in Eq. (3) has no finite maximizer on [0,∞), so the initial-state uniform distribution is also not obtained from a well-posed variational problem.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes that charged-particle multiplicity distributions measured in high-energy collisions follow from the Principle of Maximum Entropy (POME) with no a priori physical assumptions. It treats multiplicity n as a continuous variable on [0,∞), first maximizing Shannon entropy under normalization alone and claiming to recover a uniform distribution p(n)=1/N for the initial partonic state, then adding a mean constraint and obtaining the exponential distribution p(n)=λe^{-λn}. The exponential is fitted to LHCb multiplicity data in narrow rapidity windows, and the extracted Shannon entropy S=1-ln(λ) is reported to be approximately constant. The paper then extends the argument by convolving exponential distributions k times to obtain a Gamma distribution and, via a Poisson-Gamma mixture, a negative binomial distribution (NBD), which it presents as a natural explanation of the observed NBD shape.","tokens_in":4554,"tokens_out":2541,"duration_ms":25997,"significance":"If the central claim were sound, the paper would offer a strikingly parsimonious derivation of the exponential and NBD forms of multiplicity distributions, connecting them to information-theoretic maximal entropy without dynamical input. The manuscript does contain a correct statement of the standard max-entropy result for a continuous distribution on [0,∞) with fixed mean, and the comparison to LHCb data, though limited, is a useful sanity check. However, the central claim as stated is not supported: the normalization-only maximization on an infinite domain is ill-posed, the discrete nature of charged multiplicities is ignored without any error estimate, and the apparent 'constant entropy' is a monotone function of the fitted mean, so it is not an independent prediction. These issues directly undermine the advertised assumption-free derivation, so the paper does not, in its present form, establish its main conclusion.","major_comments":[{"comment":"The normalization-only maximization on [0,∞) is not well defined. A constant density p(n)=e^{α-1} on an infinite interval cannot be normalized to unity: the integral ∫_0^∞ dn diverges, so no finite Lagrange multiplier enforces the constraint. Consequently, the claimed derivation of p(n)=1/N has no valid variational formulation. This is load-bearing because it is used to recover the uniform partonic distribution in Eq. (1) and to identify N with the number of states.","section":"Eqs. (2)-(3), Sec. 2.1"},{"comment":"The exponential form p(n)=λe^{-λn} is obtained by maximizing continuous entropy on [0,∞) with a fixed mean, but charged multiplicities are integers. On the actual support n∈{0,1,2,...}, the same max-entropy problem with normalization and mean μ yields the geometric distribution p(n)=(1/(1+μ))(μ/(1+μ))^n, not the exponential. The paper provides no estimate of the error introduced by the continuum approximation, and the fit in Fig. 1 excludes n<3, precisely the region where discrete and continuous forms differ most. This unsupported discretization is an a priori assumption that contradicts the claim of an assumption-free derivation.","section":"Eqs. (4)-(5), Sec. 2.2 and Fig. 1"},{"comment":"The identification of the number of convolutions k with the number of partonic states N in Eq. (1) is asserted without derivation. Convolving exponential distributions k times gives a Gamma distribution, but k is a free parameter (the NBD shape parameter) that is not fixed by POME. The statement that 'the origin of the NBD shape lies in the convolution of maximal entropy distributions repeated k=N times' therefore rests on an additional assumption, not on the maximization principle itself.","section":"Sec. 4, Eq. (7)"},{"comment":"The reported constant Shannon entropy is not an independent prediction. From Eq. (5), S=1-ln(λ)=1+ln(μ), so the extracted entropy is a monotone function of the fitted mean μ. If μ is approximately constant across rapidity windows, S is constant by construction; the agreement with H1 therefore provides no independent confirmation of the max-entropy picture beyond the constancy of the fitted mean.","section":"Sec. 3 and Eq. (5)"}],"minor_comments":[{"comment":"There is a typo: 'he essential idea' should be 'The essential idea'. Also, the phrase 'by its nature' is used twice in one sentence in Sec. 2.","section":"Sec. 1"},{"comment":"The axis labels and data-range notation in the left panel are difficult to read due to formatting of the probability axis; the plot would benefit from standard logarithmic axis labels and an explicit statement of the fitting range.","section":"Fig. 1"},{"comment":"The sentence 'The Gamma distribution is known to be the scaling function of the negative binomial distribution' is unclear; it would be more precise to state that the Gamma distribution is the continuous mixing distribution in a Poisson-Gamma mixture.","section":"Sec. 4"},{"comment":"The relation k_Γ = k_NBD and λ_Γ = k_NBD/⟨n⟩ is stated without derivation; a brief derivation would help the reader follow the parameter mapping.","section":"Sec. 4, Eq. (7)"}],"recommendation":"reject","confidential_remarks":"The manuscript is a conference proceedings contribution with an attractive idea, but the central derivation is mathematically ill-posed for the initial-state claim (non-normalizable uniform density on an infinite interval) and the exponential-to-geometric discretization issue is not addressed. The 'constant entropy' finding is a restatement of the constancy of the fitted mean. These are load-bearing problems that cannot be fixed by local revisions without substantially changing the paper's claims, so I recommend rejection. The paper could be reconsidered if the authors rederive the results on the discrete support, provide error bounds for the continuous approximation, and explicitly characterize k and μ as data-fitted rather than derived quantities."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear Colleague,\n\nThe short version: the claim that measured multiplicity distributions are derived from maximum entropy with no physical assumptions doesn't survive the discrete nature of n. On the actual support {0,1,2,...} with a fixed mean, POME gives the geometric distribution p_n = (1/(1+μ))(μ/(1+μ))^n, not the exponential p(n)=λ e^{-λn} of Eq. (5). The exponential only follows if you treat n as continuous on [0,∞), and the paper never justifies that approximation or quantifies its error. The fit in Fig. 1 excludes n<3, precisely the region where discrete and continuous forms differ most.\n\nWhat's good: the paper is clearly written, the fits are transparent, and the author is honest about several limitations. The observation that narrow-rapidity windows have nearly constant mean, so an exponential tail is plausible, is sensible. The Poisson-Gamma route to the NBD is textbook and correctly presented. The extracted \"constant entropy\" is not an independent result—S = 1 + ln μ, so it simply restates that the fitted means are roughly constant.\n\nThe soft spots are load-bearing. The initial-state derivation in Eqs. (2)-(3) is not well-posed: a uniform density on [0,∞) is not normalizable, so the variational problem has no solution. In the final state, the mean constraint is imported from data, not derived, so \"without any a priori physical assumption\" is false on multiple counts. The NBD argument builds on the exponential convolution and identifies k=N without derivation, then admits that the data require non-integer k. These are not minor quibbles; they undermine the central claim.\n\nThe paper is honest and the author is thinking seriously, but the main derivation is mathematically flawed. A corrected version could be built by doing the discrete max-entropy calculation (geometric distribution) and comparing over the full n range without cutting low n, and by clearly stating which inputs are data-derived constraints.\n\nRecommendation: this does not deserve a serious referee as is. The novelty is low (Refs. [15,16] already applied POME to multiplicity), and the normalization error in the initial-state part is disqualifying. I'd desk reject, but invite a resubmission that fixes the support issue and drops the overreach in the abstract.","headline":"The paper's central derivation doesn't hold: on the discrete support of multiplicity, max entropy gives a geometric, not exponential, distribution, and the fit cuts the low-n region where they differ.","tokens_in":5080,"tokens_out":5053,"would_cite":false,"duration_ms":44910,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Maximum entropy alone yields measured particle multiplicity shapes","keywords":["charged particle multiplicity","maximum entropy","Shannon entropy","negative binomial distribution","Gamma distribution","Poisson-Gamma mixture","rapidity windows","LHCb data"],"falsifier":"Measure a charged-particle multiplicity distribution in a narrow rapidity window with high statistics and test whether $\\ln P(n)$ is strictly linear in $n$ when the low-multiplicity points ($n<3$) that the current fit excludes are included; a significant curvature or a low-$n$ excess would falsify the exponential prediction, and a fitted mean that drifts across sub-windows would falsify the constant-mean assumption on which it rests.","tokens_in":3998,"feed_emoji":"⚛️","tokens_out":6368,"duration_ms":64020,"temperature":0.7,"pith_summary":"The paper sets out to show that the measured charged-particle multiplicity distributions in high-energy collisions can be derived from the principle of maximum entropy alone, with no dynamical input beyond normalization and a fixed mean. In narrow rapidity windows, where the mean is approximately constant, the maximum-entropy distribution is the exponential $p(n)=\\lambda e^{-\\lambda n}$, and the paper finds that this fits LHCb data with an acceptable $\\chi^2$ while the extracted Shannon entropy stays roughly constant across rapidity. For wider rapidity windows, the paper argues that convoluting $k$ such exponentials yields a Gamma distribution, and the Poisson-Gamma mixture turns into the negative binomial distribution, offering a derivation of the well-known NBD shape that is more general than the clan model. A sympathetic reader would care because the result would mean that two ubiquitous empirical features of multiplicity data are consequences of information-theoretic indifference, not of specific production mechanisms.","feed_headline":"Maximum entropy alone yields measured multiplicity shapes","feed_subtitle":"Normalization and a fixed mean produce exponential tails and the negative binomial shape.","key_machinery":"The machinery is the variational entropy functional $F[p,\\alpha,\\beta] = -\\int_0^\\infty p(n)\\ln p(n)\\,dn + \\alpha\\left(\\int_0^\\infty p(n)\\,dn - 1\\right) + \\beta\\left(\\int_0^\\infty n p(n)\\,dn - \\mu\\right)$, whose maximization produces $p(n)=\\lambda e^{-\\lambda n}$ with $\\lambda=1/\\mu$; the normalization-only version of the same functional gives the uniform $p(n)=1/N$ used for the initial partonic state. The second ingredient is the Poisson-Gamma mixture identity: if $Y\\sim\\Gamma(k,\\lambda)$ and $X\\sim\\mathrm{Poisson}(\\mu)$ with $\\mu=Y$, then $P(X=x\\mid\\mu)=\\mathrm{NBD}(k,\\bar n=k/\\lambda)$, which converts the $k$-fold convolution of exponentials into the observed negative binomial shape.","core_discovery":"The central discovery claimed is that maximum Shannon entropy, subject only to normalization and a fixed mean $\\mu$, reproduces the charged-particle multiplicity data: in a narrow rapidity window the unique maximizer is the exponential distribution $p(n)=\\lambda e^{-\\lambda n}$ with $\\lambda=1/\\mu$, and this form describes the measured LHCb multiplicities in 0.5-unit rapidity windows, with the fitted entropy consistent with a constant. The same principle is then extended to wider windows: if the exponential is the building block of each narrow slice, then the wider-window distribution is a $k$-fold convolution of exponentials, i.e., a Gamma distribution $\\Gamma(k,\\lambda)$; mixing that Gamma with a Poisson yields the negative binomial distribution, with parameter relations $k_\\Gamma=k_{\\mathrm{NBD}}$ and $\\lambda_\\Gamma=k_{\\mathrm{NBD}}/\\langle n\\rangle$. Thus the paper claims that the famous NBD shape of charged multiplicities is not an input but an output of the maximum-entropy requirement.","pith_inferences":["The derivation implies that any collision system with a fixed charged multiplicity mean in a small phase-space window should show the same exponential shape, so the exponential tail is not a distinctive signature of the underlying production mechanism but a baseline expected from entropy maximization.","If the constant-mean assumption were relaxed to allow a mean that varies within the window, the resulting maximum-entropy distribution would be a generalized exponential whose shape could be used to measure the rapidity dependence of the mean directly from the data.","The same machinery could be applied to joint multiplicity distributions in multiple rapidity windows by imposing cross-window correlations as additional constraints, potentially exposing short- and long-range correlation parameters without a dynamical model.","One testable extension is to apply the Gamma-Poisson mixture with the parameter relations of Eq. (7) to existing wide-window data sets and check whether the predicted $k$--$\\langle n\\rangle$ correlation matches the measured NBD parameter correlation."],"forward_implications":["The exponential tail observed in narrow rapidity windows becomes a generic information-theoretic prediction, requiring no model of particle production.","The same maximum-entropy argument yields the negative binomial distribution in wider windows, making the NBD shape a consequence of repeated exponential convolutions rather than of clan dynamics.","The NBD parameter $k$ should grow linearly with the width of the rapidity window and the entropy should saturate, both of which the paper notes are consistent with existing data.","The extracted Shannon entropy in narrow windows should be constant across rapidity, offering a direct final-state counterpart to the initial-state entanglement entropy conjectured in earlier work."],"supporting_citations":[{"why":"Establishes the maximum-entropy variational method that the paper uses to derive the exponential distribution.","marker":"[13]"},{"why":"Extends the maximum-entropy formalism to quantum systems, supporting the principle's general validity.","marker":"[14]"},{"why":"Supplies the LHCb charged-particle multiplicity data in 0.5-unit rapidity windows against which the exponential fit is tested.","marker":"[17]"},{"why":"Earlier work whose uniform initial-state distribution is recovered as the normalization-only case of the same entropy maximization.","marker":"[4]"},{"why":"H1 measurement showing roughly constant entropy in ep collisions, used as the comparison for the extracted constant final-state entropy.","marker":"[12]"},{"why":"The clan model that the paper contrasts with its Gamma-convolution explanation of the negative binomial shape.","marker":"[20]"}],"fun_headline_variants":["Maximal entropy derives known multiplicity shapes","Entropy principle yields negative binomial multiplicity distributions","Maximum entropy alone explains charged particle multiplicities","No assumptions beyond mean: entropy gives multiplicity shapes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the multiplicity count $n$ can be treated as a continuous variable on $[0,\\infty)$ and that normalization plus a fixed mean, with the mean value imported from data, are the only constraints; if a continuous distribution with only those two constraints is not the right description of the measurement, the exponential form and the NBD explanation lose their stated foundation.","fun_headline_variants_meta":{"raw":{"variants":["Maximal entropy derives known multiplicity shapes","Entropy principle yields negative binomial multiplicity distributions","Maximum entropy alone explains charged particle multiplicities","No assumptions beyond mean: entropy gives multiplicity shapes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000255,"raw_usage":{"total_tokens":1501,"prompt_tokens":804,"completion_tokens":697,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":420,"completion_tokens_details":{"reasoning_tokens":642}},"tokens_in":420,"tokens_out":697,"duration_ms":8602,"temperature":1.0,"reasoning_tokens":642,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:44:44.539134+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure a charged-particle multiplicity distribution in a narrow rapidity window with high statistics and test whether $\\ln P(n)$ is strictly linear in $n$ when the low-multiplicity points ($n<3$) that the current fit excludes are included; a significant curvature or a low-$n$ excess would falsify the exponential prediction, and a fitted mean that drifts across sub-windows would falsify the constant-mean assumption on which it rests.","supporting_citations":[{"cited_title":"Giovanniniet al., Nucl","cited_arxiv_id":null,"evidence_quote":"The clan model that the paper contrasts with its Gamma-convolution explanation of the negative binomial shape."}],"review_version":1}