arxiv: 2605.13710 · v1 · submitted 2026-05-13 · 🧮 math.ST · stat.TH

Recognition: unknown

Pattern-based tests for two-dimensional copulas

L. Baringhaus , R. Gr\"ubel

Authors on Pith no claims yet

Pith reviewed 2026-05-14 17:36 UTC · model grok-4.3

classification 🧮 math.ST stat.TH

keywords copulapermutonpattern frequencyfunctional central limit theoremgoodness-of-fit testnonparametric testrank plotbootstrap

0 comments

The pith

A functional central limit theorem for pattern frequencies in bivariate rank plots enables nonparametric copula tests.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proves a functional central limit theorem for the frequencies with which specific patterns appear in the rank plots of i.i.d. bivariate samples. These plots correspond to discrete copulas, and their pattern counts converge in the topology used to define permutons. The limit theorem supplies the asymptotic distribution needed to build goodness-of-fit tests, two-sample tests, and symmetry tests for the underlying copula. A bootstrap method is introduced to compute critical values, and the same approach is applied to parametric families including the Farlie-Gumbel-Morgenstern class.

Core claim

In the context of two-dimensional random samples, pattern frequencies in rank plots obey a functional central limit theorem. This result serves as the basis for nonparametric goodness-of-fit tests, for two-sample tests, and for tests of symmetry. It includes a bootstrap variant for critical values and extends to parametric examples such as the Farlie-Gumbel-Morgenstern class and a family of delay copulas.

What carries the argument

Functional central limit theorem for pattern frequencies, establishing convergence of the empirical pattern process to a Gaussian limit in the permuton space.

Load-bearing premise

The two-dimensional samples consist of independent and identically distributed observations drawn from a continuous bivariate distribution.

What would settle it

Drawing repeated samples from a known continuous bivariate distribution such as the standard bivariate normal and checking whether the suitably normalized pattern frequency process converges in distribution to the claimed Gaussian process would test the theorem; failure to observe the predicted convergence would falsify the central claim.

read the original abstract

In statistics permutations typically arise in the context of rank plots for two-dimensional data. Such plots can also be interpreted as discrete copulas. In discrete mathematics, typically in the context of the description of large (non-random) objects, two-dimensional copulas appear as limits of permutations and are then known as permutons if the topology refers to the convergence of pattern frequencies. We obtain a functional central limit theorem for such pattern frequencies in the context of two-dimensional random samples. The result serves as the basis for nonparametric goodness-of-fit tests, for two-sample tests, and for tests of symmetry. This includes a suitable variant of the bootstrap for obtaining critical values. Pattern-based procedures are also of interest in a parametric context. We consider two examples, the Farlie-Gumbel-Morgenstern class and a family of delay copulas. We discuss implementation aspects of the resulting procedures and we provide a simulation study that supplements the theoretical results in the nonparametric case.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

A functional CLT for pattern frequencies in bivariate rank plots grounds nonparametric copula tests with a bootstrap.

read the letter

The main new result is a functional central limit theorem for the frequencies of fixed patterns in the rank permutation coming from i.i.d. continuous bivariate samples. The limit lives in the permuton topology, and the authors use it to construct tests for goodness of fit, two-sample equality, and symmetry, together with a bootstrap for critical values. They also sketch how the same idea works inside two parametric families (Farlie-Gumbel-Morgenstern and delay copulas) and supply a simulation study for the nonparametric procedures. The reduction to a uniform random permutation under the null is clean and standard, so the limiting covariance does not depend on unknown parameters. That is the part that feels genuinely useful rather than just technical. The simulation study is mentioned explicitly, which helps, though its scope and power results will need a close look in review. The parametric examples are short, but they serve mainly to illustrate implementation rather than to carry the paper. One minor soft spot is that the bootstrap consistency argument for the functional statistic is only sketched in the abstract; any gaps there would matter for the tests. Overall the assumptions are the usual ones (i.i.d., continuous margins) and the citation pattern looks appropriate for the subfield. This is for readers who work on nonparametric dependence testing or permutation-based methods in copula statistics. A statistician looking for new tools that stay distribution-free would get concrete value. The theoretical grounding is strong enough that the paper deserves a serious referee rather than a desk rejection.

Referee Report

1 major / 3 minor

Summary. The paper derives a functional central limit theorem for the empirical frequencies of fixed-size patterns in the rank permutation induced by i.i.d. samples from a continuous bivariate distribution. These frequencies are interpreted via the permuton topology on discrete copulas. The limiting Gaussian process is then used to construct nonparametric goodness-of-fit tests, two-sample tests, and symmetry tests, with a bootstrap procedure supplied for critical values. Parametric illustrations are given for the Farlie-Gumbel-Morgenstern family and a class of delay copulas, accompanied by implementation remarks and a simulation study in the nonparametric setting.

Significance. If the FCLT is valid, the work supplies a combinatorial, pattern-based route to copula inference that directly exploits the exchangeability properties of uniform random permutations. This approach may detect localized dependence structures missed by classical rank-correlation or distance-based tests. The bootstrap construction and simulation evidence are practical strengths; the grounding in standard limit theorems for pattern counts under the i.i.d. continuous assumption is technically clean.

major comments (1)

[Section 3] The functional CLT is stated for pattern frequencies in the permuton topology, yet the manuscript does not display the explicit form of the covariance kernel of the limiting process (presumably in the main theorem of Section 3). Without this kernel, it is unclear whether the bootstrap is asymptotically valid for all pattern classes or only for those with finite support; a concrete expression or reference to its derivation is needed to confirm that the tests are parameter-free under the null.

minor comments (3)

[Abstract] The abstract refers to 'a suitable variant of the bootstrap' without naming the resampling scheme (e.g., permutation bootstrap versus multiplier bootstrap). This detail should appear in the abstract or the first paragraph of the introduction.
[Simulation study] In the simulation study, the reported empirical sizes and powers are given only for a few pattern sizes; adding a table that varies both pattern size and sample size would make the finite-sample behavior easier to assess.
[Section 2] Notation for the pattern-counting functional and the permuton metric should be introduced once in a dedicated subsection rather than piecemeal across the theoretical development.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading of the manuscript and for the constructive comment. We address the point below and will incorporate the requested clarification in the revised version.

read point-by-point responses

Referee: [Section 3] The functional CLT is stated for pattern frequencies in the permuton topology, yet the manuscript does not display the explicit form of the covariance kernel of the limiting process (presumably in the main theorem of Section 3). Without this kernel, it is unclear whether the bootstrap is asymptotically valid for all pattern classes or only for those with finite support; a concrete expression or reference to its derivation is needed to confirm that the tests are parameter-free under the null.

Authors: We agree that an explicit display of the covariance kernel would improve readability and confirm the properties of the bootstrap. In the revised manuscript we will augment the statement of the main functional CLT (Theorem 3.1) with the concrete form of the covariance kernel: for two fixed patterns A and B the limiting covariance is given by the difference between the joint probability that a uniform random permutation contains both patterns at the same locations and the product of the marginal probabilities, which follows directly from the standard multinomial structure of pattern counts in i.i.d. continuous samples. Under any fixed null copula (including independence) this kernel is completely determined by the uniform measure and therefore parameter-free. Because the permuton topology is metrized by the supremum over a finite collection of patterns, the process lives in a finite-dimensional space for any practical test; the bootstrap is therefore asymptotically valid by the standard continuous-mapping argument for finite-dimensional Gaussian limits. We will also add a short derivation sketch and a reference to the classical results on pattern statistics in random permutations. revision: yes

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The derivation rests on the i.i.d. continuous bivariate sampling assumption directly implying that the rank plot is distributed as a uniform random permutation, whose pattern frequencies then admit a functional CLT in the permuton topology by standard exchangeability and convergence arguments for permutation statistics. The limiting covariance is parameter-free under the null, the bootstrap is applied directly to the empirical pattern counts without refitting, and the tests for goodness-of-fit, two-sample, and symmetry follow from the same limiting process. No equation reduces a claimed prediction to a fitted input by construction, no uniqueness theorem is imported from self-citation, and no ansatz is smuggled via prior work; the central result is self-contained against external probabilistic benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The work rests on standard mathematical results for empirical processes and functional convergence; no free parameters, ad-hoc axioms, or invented entities are indicated in the abstract.

axioms (1)

standard math Standard assumptions for functional central limit theorems in empirical processes
Invoked to obtain the limit theorem for pattern frequencies.

pith-pipeline@v0.9.0 · 5457 in / 1167 out tokens · 51970 ms · 2026-05-14T17:36:36.297905+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

52 extracted references · 38 canonical work pages

[1]

In: Proceedings of the 2021 ACM-SIAM Symposium on Discrete Algorithms (SODA), pp

Bahadur, R.R. (1971).Some Limit Theorems in Statistics.SIAM, Philadelphia. https://doi.org/10.1137/1. 9781611970630.ch1

work page doi:10.1137/1 1971
[2]

Baringhaus, L. (1994). On a modification of the Hoeffding–Blum–Kiefer–Rosenblatt independence criterion. Comm. Statist. Simulation Comput.23683–689. https://doi.org/10.1080/03610919408813193

work page doi:10.1080/03610919408813193 1994
[3]

Baringhaus, L., Franz, C. (2004). On a new multivariate two-sample test.J. Multivariate Anal.88190–206. https://doi.org/10.1016/S0047-259X(03)00079-4

work page doi:10.1016/s0047-259x(03)00079-4 2004
[4]

Baringhaus, L., Franz, C. (2010). Rigid motion invariant two-sample tests.Statist. Sinica201333–1361. Tests for copulas 25

2010
[5]

Baringhaus, L., Grübel, R. (2024). Random permutations generated by delay models and estimation of delay distributions.Electron. J. Stat.18167–190. https://doi.org/10.1214/23-EJS2205

work page doi:10.1214/23-ejs2205 2024
[6]

Beran, R., Millar, P.W. (1986). Confidence sets for a multivariate distribution.Ann. Statist.14431–443. https: //doi.org/10.1214/aos/1176349931

work page doi:10.1214/aos/1176349931 1986
[7]

Bergsma, W., Dassios, A. (2014). A consistent test of independence based on a sign covariance related to Kendall’s tau.Bernoulli201006–1028. https://doi.org/10.3150/13-BEJ514

work page doi:10.3150/13-bej514 2014
[8]

Blum, J.R.,Kiefer, J., Rosenblatt, M. (1961). Distribution free tests of independence based on the sample distribution function.Ann. Math. Statist.32485–498. https://doi.org/10.1214/aoms/1177705055

work page doi:10.1214/aoms/1177705055 1961
[9]

Bo, L., Genton, M.G. (2013). Nonparametric identification of copula structures.J. Amer. Statist. Assoc.108 666–675. https://doi.org/10.1080/01621459.2013.787083

work page doi:10.1080/01621459.2013.787083 2013
[10]

Chan, T. F. N., Král’, D., Noel, J. A., Pehova, Y., Sharifzadeh, M., Volec, J. (2020). Characterization of quasirandom permutations by a pattern sum.Random Structures Algorithms57920–939. https://doi.org/10. 1002/rsa.20956

2020
[11]

G., Zhu, L

Chen, F., Meintanis, S. G., Zhu, L. (2019). On some characterizations and multidimensional criteria for testing homogeneity, symmetry and independence.J. Multivariate Anal.173125–144. https://doi.org/10.1016/ j.jmva.2019.02.006

2019
[12]

Chernoff, H. (1956). Large sample theory: Parametric case.Ann. Math. Statist.271–22. https://doi.org/10. 1214/aoms/1177728347

work page arXiv 1956
[13]

Csörg ˝o, S. (1986). Testing for normality in arbitrary dimension.Ann. Statist.14708–723. https://doi.org/10. 1214/aos/1176349948

work page arXiv 1986
[14]

(2010).On the Asymptotic Behaviour of the Estimator of Kendall’s Tau

Dengler, B. (2010).On the Asymptotic Behaviour of the Estimator of Kendall’s Tau. Doctoral thesis, Tech- nische Universität Wien

2010
[15]

(1998)Decoupling

de la Peña, V., Giné, E. (1998)Decoupling. From dependence to independence.Springer, New York. https: //doi.org/10.1007/978-1-4612-0537-1

work page doi:10.1007/978-1-4612-0537-1 1998
[16]

S., Dassios, A., Bergsma, W

Dhar, S. S., Dassios, A., Bergsma, W. (2016). A study of the power and robustness of a new test for indepen- dence against contiguous alternatives.Electron. J. Stat.10330–351. https://doi.org/10.1214/16-EJS1107

work page doi:10.1214/16-ejs1107 2016
[17]

Drton, M., Han, F. Shi, H. (2020). High-dimensional consistent independence testing with maxima of rank correlations.Ann. Statist.483206–3227. https://doi.org/10.1214/19-AOS1926

work page doi:10.1214/19-aos1926 2020
[18]

Even-Zohar, C., Leng, C. (2021). Counting small permutation patterns.Proceedings of the 2021 ACM-SIAM Symposium on Discrete Algorithms (SODA)2288–2302. https://dl.acm.org/doi/10.5555/3458064.3458200

work page doi:10.5555/3458064.3458200 2021
[19]

Farlie, D. J. G. (1960). The performance of some correlation coefficients for a general bivariate distribution. Biometrika47307–323. https://doi.org/10.1093/biomet/47.3-4.307

work page doi:10.1093/biomet/47.3-4.307 1960
[20]

Goodness-of-fit tests for copulas.J

Fermanian, J.-D.(2005). Goodness-of-fit tests for copulas.J. Multivariate Anal.95119–152. https://doi.org/ 10.1016/j.jmva.2004.07.004

work page doi:10.1016/j.jmva.2004.07.004 2005
[21]

Fermanian, J.-D., Radulovi ´c, D., Wegkamp, M. (2004). Weak convergence of empirical copula processes. Bernoulli10847–860. https://doi.org/10.3150/bj/1099579158

work page doi:10.3150/bj/1099579158 2004
[22]

(1987).Seminar on Empirical Processes.Springer, Basel

Gaenssler, P., Stute, W. (1987).Seminar on Empirical Processes.Springer, Basel. https://doi.org/10.1007/ 978-3-0348-6269-1

1987
[23]

Genest, C., Verret, F. (2005). Locally most powerful rank tests of independence for copula models.J. Non- parametr. Stat.17521–539. https://doi.org/10.1080/10485250500038926

work page doi:10.1080/10485250500038926 2005
[24]

M., Rasch, M

Gretton, A., Borgwardt, K. M., Rasch, M. J., Schölkopf, B., Smola, A. (2012). A kernel two-sample test.J. Mach. Learn. Res.13723–773

2012
[25]

Grübel, R. (2024). Ranks, copulas, and permutons.Metrika87155–182. https://doi.org/10.1007/ s00184-023-00908-2

2024
[26]

Gumbel, E. J. (1958).Statistics of Extremes. Columbia University Press, New York. https://doi.org/10.7312/ gumb92958

1958
[27]

Hoppen, C., Kohayakawa, Y., Moreira, C.G., Ráth, B., Sampaio, R.M. (2013). Limits of permutation se- quences.J. Combin. Theory Ser. B10393–113. https://doi.org/10.1016/j.jctb.2012.09.003

work page doi:10.1016/j.jctb.2012.09.003 2013
[28]

Hoeffding, W. (1948). A non-parametric test of independence.Ann. Math. Stat.19546–557. https://doi.org/ 10.1214/aoms/1177730150

work page doi:10.1214/aoms/1177730150 1948
[29]

Janson, S. (1984). The asymptotic distributions of incomplete𝑈-statistics.Z. Wahrsch. Verw. Gebiete66 495–505. https://doi.org/10.1007/BF00531887 26L. Baringhaus and R. Grübel

work page doi:10.1007/bf00531887 1984
[30]

Janson, S., Nakamura, B., Zeilberger, D. (2015). On the asymptotic statistics of the number of occurrences of multiple permutation patterns.J. Comb.6117–143. https://dx.doi.org/10.4310/JOC.2015.v6.n1.a8

work page doi:10.4310/joc.2015.v6.n1.a8 2015
[31]

Kallenberg, W.C.M., Ledwina, T. (1987). On local and nonlocal measures of efficiency.Ann. Statist.15 1401–1420. https://doi.org/10.1214/aos/1176350601

work page doi:10.1214/aos/1176350601 1987
[32]

Kourouklis, S. (1989). On the relation between Hodges–Lehmann efficiency and Pitman efficiency.Canad. J. Statist.17311–318. https://doi.org/10.2307/3315526

work page doi:10.2307/3315526 1989
[33]

(2005).Testing Statistical Hypotheses

Lehmann, E., Romano, J. (2005).Testing Statistical Hypotheses. Wiley, New York. https://doi.org/10.1007/ 0-387-27605-X

2005
[34]

(2012).Large Networks and Graph Limits

Lovász, L. (2012).Large Networks and Graph Limits. American Mathematical Society Colloquium Publica- tions 60. Amer. Math. Soc., Providence, RI. https://doi.org/10.1090/coll/060

work page doi:10.1090/coll/060 2012
[35]

Meintanis, S. G. (2016). A review of testing procedures based on the empirical characteristic function.South African Statist. J.50, 1–14. https://doi.org/10.37920/sasj.2016.50.1.1

work page doi:10.37920/sasj.2016.50.1.1 2016
[36]

Morgenstern, D. (1956). Einfache Beispiele zweidimensionaler Verteilungen.Mitt.-Bl. Math. Statistik8234– 235

1956
[37]

Nandy, P., Weihs, L., Drton, M. (2016). Large-sample theory for the Bergsma-Dassios sign covariance.Elec- tron. J. Stat.102287–231. https://doi.org/10.1214/16-EJS1166

work page doi:10.1214/16-ejs1166 2016
[38]

Nikitin, Ya. Yu. (1995).Asymptotic Efficiency of Nonparametric Tests.Cambridge University Press, Cam- bridge. https://doi.org/10.1017/CBO9780511530081

work page doi:10.1017/cbo9780511530081 1995
[39]

Yu, Ponikarov, E.V

Nikitin, Ya. Yu, Ponikarov, E.V. (2001). Rough asymptotics of probabilities of Chernoff type large deviations for von Mises functionals and U-statistics.Amer. Math. Soc. Transl. Ser. 2203107–146. https://doi.org/10.1090/ trans2/203/04

2001
[40]

Pinelis, I. (2023). Asymptotic relative efficiency of the Kendall and Spearman correlation statistics.Theory Probab. Appl.68111–122. https://doi.org/10.1137/S0040585X97T991313

work page doi:10.1137/s0040585x97t991313 2023
[41]

Rémillard, B., Scaillet, O. (2009). Testing for equality between two copulas.J. Multivariate Anal.100377–

2009
[42]

https://doi.org/10.1016/j.jmva.2008.05.004

work page doi:10.1016/j.jmva.2008.05.004 2008
[43]

Shi, H., Drton, M., Han, F. (2022). On the power of Chatterjee’s rank correlation.Biometrika109317–333. https://doi.org/10.1093/biomet/asab028

work page doi:10.1093/biomet/asab028 2022
[44]

2004, InterStat, 5 Sz´ ekely, G

Székely, G. J., Rizzo, M. L. (2013). Energy statistics: a class of statistics based on distances.J. Statist. Plann. Inference1431249–1272. https://doi.org/10.1016/j.jspi.2013.03.018

work page doi:10.1016/j.jspi.2013.03.018 2013
[45]

Tsirelson, V.S. (1975). The density of the distribution function of the maximum of a Gaussian process.Theory Probab. Appl.20847–856. https://doi.org/10.1137/1120092

work page doi:10.1137/1120092 1975
[46]

(1998).Asymptotic Statistics

van der Vaart, A.W. (1998).Asymptotic Statistics. Cambridge University Press, Cambridge. https://doi.org/ 10.1017/CBO9780511802256

work page doi:10.1017/cbo9780511802256 1998
[47]

Springer, New York

van der Vaart, A.W., Wellner, J.A. (1996).Weak Convergence and Empirical Processes.Springer, New York. https://doi.org/10.1007/978-1-4757-2545-2

work page doi:10.1007/978-1-4757-2545-2 1996
[48]

(1981).Probability Distributions on Linear Spaces.North Holland, New York

Vakhania, N.N. (1981).Probability Distributions on Linear Spaces.North Holland, New York

1981
[49]

Weihs, L., Drton, M., Meinshausen, N. (2018). Symmetric rank covariances: a generalized framework for nonparametric measures of dependence.Biometrika105547–562. https://doi.org/10.1093/biomet/asy021

work page doi:10.1093/biomet/asy021 2018
[50]

Wieand, H. S. (1976). A condition under which the Pitman and Bahadur approaches to efficiency coincide. Ann. Statist.41003–1011. https://doi.org/10.1214/aos/1176343600

work page doi:10.1214/aos/1176343600 1976
[51]

(1995).Mathematische Statistik II.Teubner, Stuttgart

Witting, H., Müller–Funk, U. (1995).Mathematische Statistik II.Teubner, Stuttgart. https://doi.org/10.1007/ 978-3-322-90152-1

1995
[52]

Yanagimoto, T. (1970). On measures of association and a related problem.Ann. Inst. Statist. Math.2257–63. https://doi.org/10.1007/BF02506323

work page doi:10.1007/bf02506323 1970