REVIEW 4 major objections 5 minor 1 cited by
The method of mixed events for higher cumulants of conserved charges
T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read The paper claims that the conventional mixed-event background for higher cumulants of conserved charges should be replaced by global-pool sampling, which makes the mixed sample's cumulants independent of the original charged-particle…
desk verdict Useful comparison of mixed-event schemes, but the 'best' claim rests on an unexamined criterion: erasing the original Nc distribution also removes non-critical global fluctuations that a background should preserve. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the mixed-event probability distribution for the number of charged particles. Method-III replaces per-event charged fractions with a single pool fraction, so the mixed distribution is a convolution of the original multiplicity distribution with a binomial distribution whose success probability is the global charged fraction. Because that probability is constant, the original charged-particle distribution cannot enter the mixed-event cumulants at any statistics, which is the property the paper identifies as the criterion for best.
What would settle it
Generate events with a known non-critical source of kappa-sigma-squared, for example an event-by-event fluctuation in the charged fraction tied to multiplicity, and compare the non-critical signal recovered by subtracting the conventional and global-pool backgrounds; if the global-pool background does not reproduce the known non-critical kappa-sigma-squared while the conventional one does, the paper's claim that the global-pool method is best fails.
Extended reading notes
Core claim
The paper's central claim is that Method-III, in which all particles of the original events are put into one pool and each mixed event is built by drawing its multiplicity of particles at random from that pool, yields mixed-event cumulants that are independent of the original distribution of charged-particle numbers at all tested statistics, including as few as ten thousand events. In the conventional Method-I, the probability of drawing a charged particle from an event depends on that event's charged fraction, so the original distribution continues to influence the mixed sample's kappa-sigma-squared until statistics are large. Because the pool probability is a single constant, Method-III erases that influence. Methods that forbid reusing original particles are found to be unnecessary constraints, and the authors conclude that the global-pool method is the best mixed-event prescription for cumulant backgrounds.
Load-bearing premise
The load-bearing premise is that the original distribution of event charged-particle counts is an unwanted influence that a mixed-event background should erase; if fluctuations in that distribution carry physical global effects that the background is supposed to mirror, the global-pool method would subtract real physics and calling it best would not follow.
Editorial extensions
If this is right
- Replacing the conventional per-event selection with global-pool sampling removes the bias that the original charged-particle distribution imprints on mixed-event cumulants at low statistics.
- At the lowest statistics tested, the global-pool method produces mixed-event cumulants that are indistinguishable for very different original charged-particle distributions, so background estimates no longer depend on the shape of that distribution.
- The no-reuse constraints of the other two methods are unnecessary, since they add no improvement over their with-reuse counterparts.
- A global-pool mixed sample preserves the multiplicity distribution and the mean charged-particle number within statistical error, retaining the global non-critical features the background is designed to estimate.
Reading between the lines
- The paper's criterion of best is maximal randomness, meaning independence from the original charged-particle distribution; a complementary criterion would ask whether the background reproduces all non-critical correlations, in which case the global-pool method's erasure of the joint charged-number and multiplicity distribution would be a drawback rather than a virtue.
- A directly testable extension is to simulate centrality or volume fluctuations in the charged fraction and compare how faithfully the conventional and global-pool backgrounds recover the known non-critical kappa-sigma-squared, which would show whether global-pool sampling over-subtracts.
- Because the global-pool method has a closed analytic form, it could be used to derive a correction converting measured cumulants into estimates of the critical excess, a construction the paper does not pursue.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper considers four methods for constructing mixed-event samples to serve as backgrounds for higher cumulants of conserved charges in heavy-ion collisions. Method-I selects, for each mixed event, Nch original events and takes one particle from each; Method-II removes used particles from the original sample; Method-III pools all particles from all original events and draws Nch particles from the pool with replacement; Method-IV draws without replacement. The authors derive expressions for the mixed-event Nc distributions, and test the methods on toy models with fixed multiplicity and either Poisson or two-delta Nc distributions. They find that all four methods preserve the mean charged-particle number and remove the correlation between Nc and the event label. They then show that the kappa-sigma-squared cumulant of the mixed sample for Method-III is independent of the original Nc distribution even at low statistics, while Method-I only approaches the analytic expectation at high statistics. The paper concludes that Method-III, being the 'most random or least constrained', is the best method, and recommends it over the conventional Method-I.
Significance. The manuscript addresses a practical question of direct relevance to the STAR/RHIC cumulant program: how to construct mixed-event backgrounds without introducing statistical bias at limited event counts. The paper's explicit formulation of four methods and its transparent toy-model comparisons are useful, and the derivation of the probability distributions, despite some technical ambiguities, correctly identifies a real difference between the methods: Method-III eliminates the event-by-event variation of the charged fraction and thus converges much faster. If the recommended criterion—that the background should be independent of the original Nc distribution—is accepted, the conclusion is well-supported by the simulations. However, the paper does not justify that criterion, and it is in tension with the usual purpose of mixed-event backgrounds, which is to reproduce non-critical global/systematic effects such as centrality-bin-width and volume fluctuations. The analytic baseline in Fig. 1 is not derived, and Methods-II and -IV are not shown.
major comments (4)
- [Sec. III C, Sec. IV, Abstract] The conclusion that Method-III is 'best' rests on the assumption that the influence of the original distribution of Nc on mixed-event cumulants is a defect to be removed. This assumption is not stated or defended. In Sec. I, the paper itself identifies centrality bin width and initial size fluctuations as non-critical global/systematic effects that the mixed-event background should help subtract; those effects produce event-by-event fluctuations in the charged fraction f_i = Nc_i/Nch_i, not just in Nch. Method-I/II preserve these fluctuations, because the mixed-event Nc|Nch distribution is a mixture of binomials with event-specific probabilities f_i, whereas Method-III forces a single global probability p_c and therefore removes the variance of f_i entirely. If f_i fluctuations are non-critical, Method-III would underestimate the background, leaving those non-critical fluctuations in the subtracted cumulants. The paper provides no physics argument that the full shape of P^o(Nc) beyond its mean is a critical correlation to be erased. The ranking of methods is therefore a choice of criterion rather than a derived result.
- [Fig. 1, Sec. III C] The blue horizontal line labeled 'expectation of analytic calculation' is introduced without derivation. Because the central comparison uses this line as the reference value to which Method-I should converge, the paper should provide the explicit analytic expression for the asymptotic kappa-sigma-squared of the mixed sample in the fixed-multiplicity toy model, and show how it depends on the original Nc distribution for Methods I and III. Without this, the claim that Method-I is 'influenced by the original distribution of Nc' at low statistics is only a simulation observation, and the reader cannot assess the size of the bias or the convergence rate claimed.
- [Sec. III C, Fig. 1] The text states that the statistics dependences for Methods II and IV are 'similar to Fig. 1(a) and (b), respectively' but does not show them. Since the paper's purpose is to compare four methods, the omission prevents the reader from verifying the conclusion that the no-replacement constraint is 'trivial and unnecessary'. A complete comparison, for example as additional panels in Fig. 1 or as a table of the four methods' kappa-sigma-squared values at representative statistics, should be added.
- [Sec. II, Eqs. (2)-(6)] The probability formulas for Methods I and II are not properly defined. In the stated algorithm, an event is selected randomly from the original sample, so the probability of drawing a charged particle from the selected event is simply that event's charged fraction f_i = Nc_i/Nch_i; the factors P^o(Nc) and P^o(Nch) in Eqs. (3)-(4) do not represent the event-selection probability. As written, these formulas appear to use the product of one-dimensional marginal distributions, which is not the joint probability of drawing an event with given (Nc,Nch). The derivation of P^m(Nc) should be corrected or clarified, since it is the basis for the claim that P^o(Nc) enters the mixed-event distribution in Method-I.
minor comments (5)
- [Sec. II] The phrase 'Method-VI' should be 'Method-IV' in the sentence 'Obviously, Methods-II and Method-VI are more restrictive...'.
- [Abstract] The phrase 'It is showed' should be 'It is shown'.
- [Eq. (10)] The notation 'C(N e,N e, c, )' is confusing; please use 'C(N^e, N_c^e)'.
- [Fig. 1 caption] The legend text 'Poisson, δ2, )c(NoP' is garbled; it should read 'Poisson' and 'two-δ functions'.
- [Tables I and II] The reported uncertainties should specify whether they are standard deviations of the distribution of the mean or standard errors of the mean.
Circularity Check
No significant circularity: Method-III's independence from the original Nc distribution follows directly from its defining pool construction, and the paper's ranking rests on an explicit evaluative criterion rather than a circular derivation.
full rationale
The paper's core comparison is not circular. Method-III is defined by placing all particles in a global pool and drawing with a fixed charged-particle probability pc = sum Nc_i / sum Nch_i (Eqs. 7-9); therefore the mixed-event charged distribution conditional on Nch is binomial and indeed independent of the original Nc distribution by construction. Method-I's dependence on P_o(Nc) similarly follows from its per-event sampling probabilities in Eqs. (2)-(4). The observed difference in Fig. 1 is a direct mathematical consequence of these definitions, not a fitted or self-referential prediction. No parameter is fitted to data and then renamed as a prediction, no uniqueness theorem from the authors' prior work is used to force the conclusion, and the small self-citations (e.g., Refs. [15,16] for the sub-sample averaging convention, and the coming paper Ref. [31]) are not load-bearing for the central claim. The main weakness is that the recommendation of Method-III as 'best' equates 'best' with erasing the full original Nc distribution, which is an unexamined evaluative criterion rather than a circular step; a different physical criterion could favor Method-I. That is an assumption issue, not a circularity issue.
Assumptions & free parameters
assumptions (5)
- domain assumption Mixed-event samples should preserve the original event multiplicity distribution and the mean of charged particles.
- ad hoc to paper The influence of the original Nc distribution on mixed-event cumulants is a defect to be removed rather than a signal to be preserved.
- domain assumption Global and systematic non-critical fluctuations are independent of inner correlations and can be represented by a sample with correlations switched off.
- standard math Independent random draws from a large pool with a fixed charged fraction produce uncorrelated particles.
- domain assumption Poisson and two-delta toy models are representative enough to test the methods at low statistics.
Cite this review
Pith. "Pith review of The method of mixed events for higher cumulants of conserved charges." pith.science (2026). https://pith.science/paper/IYXA3TC4
@misc{pith2026190805465,
author = {Pith},
title = {Pith review of: The method of mixed events for higher cumulants of conserved charges},
year = {2026},
howpublished = {\url{https://pith.science/paper/IYXA3TC4}},
note = {Machine review of arXiv:1908.05465}
}
read the original abstract
Higher cumulants of conserved charges are sensitive observables of quantum chromodynamics phase transitions. The sample of mixed events provides a background to estimate non-critical effects of cumulants. Four possible methods for constructing the sample of mixed events are suggested. The effectiveness of each method is examined. It is showed that the method of most random or least constrain is the best, rather than the conventional method.
Figures
Forward citations
Cited by 1 Pith paper
-
Subtracting non-critical fluctuations in higher cumulants of conserved charges
Dynamical cumulants defined as original minus mixed-event cumulants remove statistical and detector-related fluctuations in AMPT simulations, leaving cleaner critical-point-sensitive signals.
Reference graph
Works this paper leans on
-
[1]
Athanasiou, K
C. Athanasiou, K. Rajagopal, and M. Stephanov. Phys.Rev.D, 82:074008, (2010)
2010
- [2]
-
[3]
Stephanov
M.A. Stephanov. Phys.Rev.Lett, 107:052301, (2011)
2011
-
[4]
Hatta and M.A
Y. Hatta and M.A. Stephanov. Phys.Rev.Lett, 91:102003, (2003)
2003
-
[5]
Cheng, P
M. Cheng, P. Hegde, and C. et al. Jung. Eur.Phys.J.C, 71:1694, (2011)
2011
-
[6]
Morita, B
K. Morita, B. Friman, and K. Redlich. Phys.Lett.B, 741:178–183, (2015)
2015
-
[7]
Asakawa, S
M. Asakawa, S. Ejiri, and M. Kitazawa. Phys.Rev.Lett, 103:262301, (2009)
2009
-
[8]
Adamczyk et al.(STAR Collaboration)
L. Adamczyk et al.(STAR Collaboration). Phys.Rev.Lett, 113:092301, (2014)
2014
Show all 31 references
-
[9]
Adamczyk et al.(STAR Collaboration)
L. Adamczyk et al.(STAR Collaboration). Phys.Rev.Lett, 112:032302, (2014)
2014
-
[10]
Gupta, X
S. Gupta, X. Luo, and B et al. Mohanty. Science, 332:1525–1528, (2011)
2011
-
[11]
arXiv, 1007:2613, (2010)
M.M.Aggarwal et al.(STAR Collaboration). arXiv, 1007:2613, (2010)
2010
-
[12]
X. Luo. Nucl.Phys.A, 956:75–82, (2016)
2016
-
[13]
X. Luo. Nucl.Phys.A, 904-905:911c–914c, (2013)
2013
-
[14]
arXiv, 1906:00936v1, 2019
Adam Bzdak, ShinIchi Esumi, Volker Koch, Jinfeng Liao, Mikhail Stephanov, and Nu Xu. arXiv, 1906:00936v1, 2019
1906
-
[15]
J.Phys.G:Nucl.Part.Phys, 38:11504, (2011)
Lizhu Chen, Xue Pan, Fengbo Xiong, Lin Li, Na Li, Zhiming Li, Gang Wang, and Yuanfang Wu. J.Phys.G:Nucl.Part.Phys, 38:11504, (2011)
2011
-
[16]
J.Phys.G:Nucl.Part.Phys, 42:065103, (2015)
Lizhu Chen, Zhiming Li, Xia Zhong, Yuncun He, and Yuanfang Wu. J.Phys.G:Nucl.Part.Phys, 42:065103, (2015)
2015
-
[17]
Phys.Rev.C, 89:014904, (2014)
Xue Pan, Fan Zhang, Zhiming Li, Lizhu Chen, Mingmei Xu, and Yuanfang Wu. Phys.Rev.C, 89:014904, (2014)
2014
-
[18]
Csernai, and Ben- Hao Sa
Dai-Mei Zhou, Ayut Limphirat, Yu-liang Yan, Cheng Yun, Yu-peng Yan, Xu Cai, Laszlo p. Csernai, and Ben- Hao Sa. Phys.Rev.C, 85:064916, (2012)
2012
-
[19]
Bzdak and V
A. Bzdak and V. Koch. Phys.Rev.C, 91:027901, (2015)
2015
-
[20]
Phys.Rev.C, 86:044904, (2012)
Adam Bzdak and Volker Koch. Phys.Rev.C, 86:044904, (2012)
2012
-
[21]
X. Luo. J.Phys.C.S, 316:012003, (2011)
2011
-
[22]
Phys.Rev.C, 94:024918, (2016)
Lijia Jiang, Pengfei Li, and Huichao Song. Phys.Rev.C, 94:024918, (2016)
2016
-
[23]
Phys.Rev.C, 97:014902, 2018
Jixing Li, Hao-jie Xu, and Huichao Song. Phys.Rev.C, 97:014902, 2018
2018
-
[24]
(NA49 Collaboration)
S.V.Afanasev et al. (NA49 Collaboration). Phys.Lett.B, 491:59–66, (2000)
2000
-
[25]
Phys.Rev.C, 77:064908, (2008)
C.Alt et al (NA49 Collaboration). Phys.Rev.C, 77:064908, (2008)
2008
-
[26]
Phys.Rev.C, 116:172302, (2016)
V.Khachatryan et al.(CMS Collaboration). Phys.Rev.C, 116:172302, (2016)
2016
-
[27]
(NA49 Collaboration)
S.V.Afanasev et al. (NA49 Collaboration). Phys.Rev.Lett, 86:1965, (2001)
2001
-
[28]
70:034902, (2004)
T.Anticic et al. 70:034902, (2004)
2004
-
[29]
(STAR Collaboration)
L.Adamczyk et al. (STAR Collaboration). Phys.Rev.C, 88:014903, (2013)
2013
-
[30]
(STAR Collaboration)
L.Adamczyk et al. (STAR Collaboration). Phys.Rev.C, 93:014907, (2016)
2016
-
[31]
arXiv, 1908.05470, (2019)
Fan Zhang, Zhiming Li, Lizhu Chen, Xue Pan, Mingmei Xu, Yeyin Zhao, and Yuanfang Wu. arXiv, 1908.05470, (2019)
2019 arXiv
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.