pith. sign in

arxiv: 2604.10254 · v1 · submitted 2026-04-11 · ❄️ cond-mat.stat-mech

Algorithmic overlaps as thermodynamic variables: from local to cluster Monte Carlo dynamics in critical phenomena

Ian Pil\'e , Youjin Deng , Lev Shchur This is my paper

Pith reviewed 2026-05-10 15:24 UTC · model grok-4.3

classification ❄️ cond-mat.stat-mech
keywords Monte Carlo algorithmscluster algorithmscritical phenomenaIsing modelPotts modelFortuin-Kasteleyn clustersorder parameterconfiguration overlap
0
0 comments X

The pith

The overlap of successive clusters or configurations in cluster Monte Carlo algorithms behaves as an order parameter that tracks the thermodynamics of phase transitions.

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

This paper examines the spatial overlap between successive spin configurations generated by the Metropolis, Swendsen-Wang, and Wolff algorithms applied to the Ising and three-state Potts models. It finds that the overlap of two successive Wolff clusters directly reflects critical behavior and functions as an order parameter for the algorithm dynamics. The variation in overlap between consecutive full configurations plays the same role for the Swendsen-Wang algorithm. No analogous critical signature appears in the local Metropolis dynamics, where behavior is instead set by the spin-flip acceptance rate. The authors connect these geometric overlaps to the Fortuin-Kasteleyn clusters and conclude that the overlaps mirror the underlying thermodynamic phase transition in all examined cases.

Core claim

The central claim is that the overlap between two successive Wolff clusters, and the variation in overlap between two consecutive lattice configurations under Swendsen-Wang updates, each reflect the critical behavior of the model and can be treated as order parameters for the algorithm's dynamics; these geometric quantities are naturally tied to the percolation properties of Fortuin-Kasteleyn clusters and therefore reproduce the thermodynamics of the phase transition, whereas local Metropolis dynamics show no such overlap signature and are controlled instead by the acceptance rate.

What carries the argument

The spatial overlap (or its variation) between successive spin configurations or clusters, which acts as a geometric proxy whose critical scaling mirrors that of the thermodynamic order parameter through its connection to Fortuin-Kasteleyn clusters.

If this is right

  • Wolff-cluster overlap can be tracked during a run to diagnose proximity to criticality without extra cost.
  • Swendsen-Wang overlap variation provides a geometric diagnostic of the Fortuin-Kasteleyn percolation transition.
  • Local Metropolis dynamics remain governed solely by acceptance rate even near criticality, with no overlap-based critical signature.
  • The geometric-thermodynamic link holds for both the Ising and three-state Potts universality classes.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Overlap monitoring might supply a lightweight way to locate critical points in simulations of other discrete spin models.
  • The correspondence between geometric overlap and thermodynamic exponents could be used to relate algorithmic autocorrelation times more directly to known critical exponents.
  • Similar overlap constructions could be tested in continuous-time or event-chain Monte Carlo schemes to see whether they likewise capture criticality.

Load-bearing premise

The critical signatures observed in the overlaps are intrinsic to the algorithms and models rather than artifacts of the chosen lattice sizes, temperatures, or measurement protocols.

What would settle it

Repeating the overlap measurements on lattices several times larger than those used or at temperatures well away from criticality and checking whether the order-parameter-like scaling in the overlaps disappears would test the claim.

Figures

Figures reproduced from arXiv: 2604.10254 by Ian Pil\'e, Lev Shchur, Youjin Deng.

Figure 2
Figure 2. Figure 2: Wolff algorithm, Ising model. The average fraction [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 1
Figure 1. Figure 1: Dependence of U (W) 2 on temperature for several values of size [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: Wolff algorithm, Ising model. The energy per spin [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 6
Figure 6. Figure 6: Wolff updates, Ising model. Scaled variance [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Wolff updates, three-state Potts model. Average [PITH_FULL_IMAGE:figures/full_fig_p005_7.png] view at source ↗
Figure 10
Figure 10. Figure 10: Wolff updates, three-state Potts model. Rescaled [PITH_FULL_IMAGE:figures/full_fig_p006_10.png] view at source ↗
Figure 9
Figure 9. Figure 9: Wolff updates, three-state Potts model. Cluster [PITH_FULL_IMAGE:figures/full_fig_p006_9.png] view at source ↗
Figure 13
Figure 13. Figure 13: Swendsen–Wang updates, Ising model. Top: vari [PITH_FULL_IMAGE:figures/full_fig_p007_13.png] view at source ↗
Figure 12
Figure 12. Figure 12: Swendsen–Wang updates, Ising model. Mean two [PITH_FULL_IMAGE:figures/full_fig_p007_12.png] view at source ↗
Figure 14
Figure 14. Figure 14: Swendsen–Wang updates, Ising model. Rescaled [PITH_FULL_IMAGE:figures/full_fig_p008_14.png] view at source ↗
Figure 16
Figure 16. Figure 16: Swendsen–Wang updates, three-state Potts model. [PITH_FULL_IMAGE:figures/full_fig_p008_16.png] view at source ↗
Figure 15
Figure 15. Figure 15: Swendsen–Wang updates, three-state Potts model. [PITH_FULL_IMAGE:figures/full_fig_p008_15.png] view at source ↗
Figure 17
Figure 17. Figure 17: Swendsen–Wang updates, three-state Potts model. [PITH_FULL_IMAGE:figures/full_fig_p009_17.png] view at source ↗
Figure 19
Figure 19. Figure 19: Metropolis updates, Ising model. Top: two-step [PITH_FULL_IMAGE:figures/full_fig_p009_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Metropolis updates, Ising model. Local linear fit [PITH_FULL_IMAGE:figures/full_fig_p010_20.png] view at source ↗
Figure 22
Figure 22. Figure 22: Metropolis updates, Ising model. Top: The vari [PITH_FULL_IMAGE:figures/full_fig_p010_22.png] view at source ↗
Figure 21
Figure 21. Figure 21: Metropolis updates, Ising model. U2 versus Metropolis acceptance rate, L = 512. The red dashed line indicates the point where U2(a) starts to deviate from the lin￾ear dependence at a = 0.347 in the high-temperature phase. The overlap variance exhibits a peculiarity in the crit￾ical region at Tc, a pronounced peak is observed whose height decreases with the system size ( [PITH_FULL_IMAGE:figures/full_fig_… view at source ↗
Figure 25
Figure 25. Figure 25: Metropolis updates, three-state Potts model. Lo [PITH_FULL_IMAGE:figures/full_fig_p011_25.png] view at source ↗
Figure 23
Figure 23. Figure 23: Metropolis updates, three-state Potts model. Top: [PITH_FULL_IMAGE:figures/full_fig_p011_23.png] view at source ↗
Figure 24
Figure 24. Figure 24: Metropolis updates, Potts model. U2 versus Metropolis acceptance rate, L = 512. Red dashed line indi￾cates the point where U2(a) starts to deviate from the linear dependence at a = 0.397 Near the transition, the energy dependence is again locally linear, but with a slope −0.573(5) ( [PITH_FULL_IMAGE:figures/full_fig_p011_24.png] view at source ↗
Figure 28
Figure 28. Figure 28: Wolff updates, three-state Potts model. Multi [PITH_FULL_IMAGE:figures/full_fig_p012_28.png] view at source ↗
Figure 27
Figure 27. Figure 27: Wolff updates, Ising model. Multistep cluster [PITH_FULL_IMAGE:figures/full_fig_p012_27.png] view at source ↗
Figure 30
Figure 30. Figure 30: Swendsen–Wang updates, three-state Potts model. [PITH_FULL_IMAGE:figures/full_fig_p013_30.png] view at source ↗
read the original abstract

We investigate the spatial overlap of successive spin configurations in Markov chain Monte Carlo simulations using the local Metropolis algorithm and the Swendsen-Wang and Wolff cluster algorithms. We examine the dynamics of these algorithms for two models in different universality classes: the Ising model and the Potts model with three components. The overlap of two successive Wolff clusters reflects critical behavior and can be used as an order parameter for the algorithm's dynamics. In the case of the Swendsen-Wang algorithm, similar behavior is demonstrated by the variation in the overlap of two consecutive lattice configurations, which behaves like order parameter. Nothing similar is observed for the Metropolis algorithm, and the dynamics in the critical region are determined by the spin flip frequency, which is equivalent to the acceptance rate. Thus, the critical behavior of Wolff cluster overlap and the variation of configuration overlap in the Swendsen-Wang algorithm are naturally related to the critical behavior of geometric objects - Fortuin-Kastelein clusters. Interestingly, in all cases, the geometric quantity - configuration overlap or its variation - reflects the thermodynamics of the phase transition.

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.

Referee Report

2 major / 2 minor

Summary. The manuscript investigates spatial overlaps between successive spin configurations generated by local Metropolis, Swendsen-Wang, and Wolff cluster Monte Carlo algorithms applied to the Ising and 3-state Potts models. It claims that the overlap between successive Wolff clusters exhibits critical behavior usable as an order parameter for the algorithm dynamics; that the variation in overlap between consecutive lattice configurations plays an analogous role for Swendsen-Wang; and that neither feature appears for Metropolis, whose dynamics are instead controlled by the acceptance rate. These geometric quantities are asserted to be naturally tied to Fortuin-Kasteleyn clusters and to reflect the underlying thermodynamics of the phase transition.

Significance. If the central observations survive finite-size and protocol controls, the work would supply a concrete geometric diagnostic that links algorithmic efficiency directly to the critical properties of FK clusters. This could furnish a new, model-independent way to quantify how cluster algorithms suppress critical slowing down and might open avenues for designing or analyzing dynamics via overlap statistics rather than conventional autocorrelation times.

major comments (2)
  1. [Numerical results (throughout)] The abstract and reported numerical observations on Ising and 3-state Potts models supply no finite-size scaling analysis, no L-extrapolation of the overlap quantities, and no control runs performed away from Tc. Without these, it remains possible that the apparent order-parameter behavior simply tracks the growth of the correlation length inside the chosen finite-L and T window rather than constituting an intrinsic algorithmic-thermodynamic link.
  2. [Discussion of results] The claim that Wolff-cluster overlap and SW configuration-overlap variation are 'naturally related to the critical behavior of geometric objects—Fortuin-Kasteleyn clusters' requires explicit demonstration that the measured overlaps are insensitive to the precise definition of the measurement window and to the cluster-labeling protocol; the current presentation leaves open whether post-hoc choices in how overlaps are computed affect the reported critical signatures.
minor comments (2)
  1. [Abstract and § on Metropolis] The abstract states that 'nothing similar is observed for the Metropolis algorithm' but does not quantify how the acceptance-rate dependence was measured or compared against the overlap statistics; a brief table or plot of acceptance rate versus temperature would clarify the contrast.
  2. [Methods] Notation for the overlap quantities (e.g., whether they are normalized by volume or by cluster size) is not defined in the provided abstract and should be introduced with an equation at first use.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading of our manuscript and the constructive comments. We address each major point below and have revised the manuscript to incorporate additional analyses that strengthen the claims.

read point-by-point responses

  1. Referee: [Numerical results (throughout)] The abstract and reported numerical observations on Ising and 3-state Potts models supply no finite-size scaling analysis, no L-extrapolation of the overlap quantities, and no control runs performed away from Tc. Without these, it remains possible that the apparent order-parameter behavior simply tracks the growth of the correlation length inside the chosen finite-L and T window rather than constituting an intrinsic algorithmic-thermodynamic link.

    Authors: We agree that the absence of explicit finite-size scaling and off-criticality controls leaves the interpretation open to the concern raised. In the revised manuscript we have added data for multiple lattice sizes, performed L-extrapolations of the key overlap quantities, and included control simulations at temperatures both above and below Tc. These additional runs show that the order-parameter-like signatures in the Wolff-cluster overlap and the Swendsen-Wang configuration-overlap variation are absent or qualitatively different away from criticality, supporting an intrinsic link to the phase transition rather than a simple finite-size correlation-length effect. revision: yes

  2. Referee: [Discussion of results] The claim that Wolff-cluster overlap and SW configuration-overlap variation are 'naturally related to the critical behavior of geometric objects—Fortuin-Kasteleyn clusters' requires explicit demonstration that the measured overlaps are insensitive to the precise definition of the measurement window and to the cluster-labeling protocol; the current presentation leaves open whether post-hoc choices in how overlaps are computed affect the reported critical signatures.

    Authors: We acknowledge that the original presentation did not include systematic robustness checks. The revised manuscript now contains an explicit subsection that varies the measurement window (different Monte Carlo step separations between configurations) and compares two independent cluster-labeling implementations. The critical signatures remain stable under these variations. We have also added a direct comparison of the overlap statistics with the Fortuin-Kasteleyn cluster-size distribution and percolation observables, making the geometric connection more explicit. revision: yes

Circularity Check

0 steps flagged

No significant circularity; claims rest on direct numerical observations

full rationale

The paper reports Monte Carlo simulation results for overlaps in Metropolis, Swendsen-Wang, and Wolff algorithms applied to Ising and 3-state Potts models. Central statements (overlap or its variation behaving like an order parameter near criticality, and its relation to Fortuin-Kasteleyn clusters) are presented as outcomes of those measurements rather than as algebraic identities or fitted quantities. The FK-cluster link is invoked as a pre-existing standard property of the cluster algorithms, not derived or justified inside the paper via self-citation or redefinition. No parameter fitting, ansatz smuggling, or uniqueness theorems appear in the provided text; the derivation chain is therefore empirical and self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The work relies on standard assumptions of Markov chain Monte Carlo ergodicity and the known equivalence of cluster algorithms to Fortuin-Kasteleyn representations, with no new free parameters or invented entities introduced.

axioms (2)
  • standard math Markov chain Monte Carlo algorithms generate equilibrium configurations according to the Boltzmann distribution.
    Invoked implicitly when treating successive configurations as samples from the equilibrium ensemble.
  • domain assumption Fortuin-Kasteleyn clusters underlie the critical behavior of the Ising and Potts models.
    Used to connect geometric overlaps to thermodynamic critical phenomena.

pith-pipeline@v0.9.0 · 5494 in / 1392 out tokens · 20407 ms · 2026-05-10T15:24:57.298561+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

34 extracted references · 34 canonical work pages

  1. [1]

    The mean acceptance rate⟨P acc⟩, averaged over all trial moves in a sweep, which is known to behave as a thermodynamic function of temperature in a number of models [25]

    work page

  2. [2]

    (8) below with ∆t=nsweeps

    Ann-stepoverlapbetween configurations sepa- rated bynsuccessive sweeps, defined in Eq. (8) below with ∆t=nsweeps. Alongside these algorithmic observables, we compute standard thermodynamic quantities from the energy time series. The heat capacity is obtained from the fluctua- tion–dissipation relation, C= N T 2 ⟨ϵ2⟩ − ⟨ϵ⟩2 ,(5) and serves as a reference a...

    work page

  3. [3]

    Dependence ofU (W) 2 on temperature for several values of size

    Ising model with Wolff updates Figure 1. Dependence ofU (W) 2 on temperature for several values of size. Figure 1 shows the dependence ofU (W) 2 on tem- perature for several lattice size values. In the high- temperature phase, the overlapU (W) 2 is zero in the thermodynamic limit, since the cluster radius decreases rapidly with increasing temperature. In ...

    work page

  4. [4]

    Figure 3

    Plotting a graph ofU (W) 2 versus energy, initially calcu- lated as a function of temperature, is possible due to the relationship between energy and temperature, see figure 3. Figure 3. Wolff algorithm, Ising model. The energy per spin versus temperature. We can gain additional insight into how the overlap valueU (W) 2 vanishes at the critical point as t...

    work page

  5. [5]

    critical exponents

    Three state Potts model with Wolff updates We present a similar analysis for the three-state Potts model, focusing on how the observed magnitude of cluster overlap behaves during the transition. The figure 7 shows the dependence of the average over- lapU (W) 2 on the energy per spin. The transition is marked by a sharp drop in the overlap, which becomes i...

    work page

  6. [6]

    Binder,Monte Carlo Methods in Statistical Physics (Springer, Berlin, 1986)

    K. Binder,Monte Carlo Methods in Statistical Physics (Springer, Berlin, 1986)

    work page 1986

  7. [7]

    D. P. Landau and K. Binder,A Guide to Monte Carlo Simulations in Statistical Physics, 4th ed. (Cambridge University Press, Cambridge, 2014)

    work page 2014

  8. [8]

    Weigel, Simulating Spin Models on GPU, Computer Physics Communications182, 1833 (2011)

    M. Weigel, Simulating Spin Models on GPU, Computer Physics Communications182, 1833 (2011)

    work page 2011

  9. [9]

    Bisson, M

    M. Bisson, M. Bernaschi, M. Fatica, N. G. Fytas, I. G.- A. Pemart´ ın, V. Mart´ ın-Mayor, and A. Vasilopoulos, Massive-Scale Simulations of 2D Ising and Blume–Capel Models on Rack-Scale Multi-GPU Systems, Computer Physics Communications315, 109690 (2025)

    work page 2025

  10. [10]

    Lin and J

    Y. Lin and J. B. Kogut, Parallel Cluster Algorithms for Spin Systems on Hybrid Architectures, Journal of Com- putational Physics347, 270 (2017)

    work page 2017

  11. [11]

    R. H. Swendsen and J.-S. Wang, Nonuniversal Critical Dynamics in Monte Carlo Simulations, Physical Review Letters58, 86 (1987)

    work page 1987

  12. [12]

    Wolff, Collective Monte Carlo Updating for Spin Sys- tems, Physical Review Letters62, 361 (1989)

    U. Wolff, Collective Monte Carlo Updating for Spin Sys- tems, Physical Review Letters62, 361 (1989)

    work page 1989

  13. [13]

    C. M. Fortuin and P. W. Kasteleyn, On the Random- Cluster Model. I. Introduction and Relation to Other Models, Physica57, 536 (1972)

    work page 1972

  14. [14]

    C. M. Fortuin, On the Random-Cluster Model. II. The Percolation Model, Physica58, 393 (1972)

    work page 1972

  15. [15]

    H. G. Evertz, The Loop Algorithm, Advances in Physics 52, 1 (2003)

    work page 2003

  16. [16]

    B. A. Berg and T. Neuhaus, Multicanonical Algorithms for First Order Phase Transitions, Physical Review Let- ters68, 9 (1992)

    work page 1992

  17. [17]

    Wang and D

    F. Wang and D. P. Landau, Efficient, Multiple-Range Random Walk Algorithm to Calculate the Density of States, Physical Review Letters86, 2050 (2001)

    work page 2050

  18. [18]

    P. M. C. de Oliveira, T. J. P. Penna, and H. J. Herrmann, Broad Histogram Method, Brazilian Journal of Physics 26, 677 (1996)

    work page 1996

  19. [19]

    Hukushima and K

    K. Hukushima and K. Nemoto, Exchange Monte Carlo Method and Application to Spin Glass Simulations, Jour- nal of the Physical Society of Japan65, 1604 (1996)

    work page 1996

  20. [20]

    D. J. Earl and M. W. Deem, Parallel Tempering: Theory, Applications, and New Perspectives, Physical Chemistry Chemical Physics7, 3910 (2005)

    work page 2005

  21. [21]

    Trebst, D

    S. Trebst, D. A. Huse, and M. Troyer, Optimizing the Ensemble for Equilibration in Broad-Histogram Monte Carlo Simulations, Physical Review E70, 046701 (2004)

    work page 2004

  22. [22]

    R. E. Belardinelli and V. D. Pereyra, Fast Algorithm to Calculate Density of States, Physical Review E75, 046701 (2007)

    work page 2007

  23. [23]

    L. Y. Barash, M. A. Fadeeva, and L. N. Shchur, Control of Accuracy in the Wang–Landau Algorithm, Physical Review E96, 043307 (2017), arXiv:1706.06097

    work page arXiv 2017

  24. [24]

    Machta, Population Annealing with Weighted Aver- ages: A Monte Carlo Method for Rough Free-Energy Landscapes, Physical Review E82, 026704 (2010)

    J. Machta, Population Annealing with Weighted Aver- ages: A Monte Carlo Method for Rough Free-Energy Landscapes, Physical Review E82, 026704 (2010)

    work page 2010

  25. [25]

    Weigel, L

    M. Weigel, L. Y. Barash, L. N. Shchur, and W. Janke, Understanding Population Annealing Monte Carlo Sim- ulations, Physical Review E103, 053301 (2021), arXiv:2102.06611

    work page arXiv 2021

  26. [26]

    L. Y. Barash, M. Weigel, M. Borovsk´ y, W. Janke, and L. N. Shchur, GPU Accelerated Population Annealing Algorithm, Computer Physics Communications220, 341 (2017), arXiv:1703.03676

    work page arXiv 2017

  27. [27]

    Rose and J

    N. Rose and J. Machta, Equilibrium Microcanonical An- nealing for First-Order Phase Transitions, Physical Re- view E100, 063304 (2019)

    work page 2019

  28. [28]

    Mozolenko and L

    V. Mozolenko and L. Shchur, Blume–Capel Model Analysis with a Microcanonical Population Anneal- ing Method, Physical Review E109, 045306 (2024), arXiv:2402.18985

    work page arXiv 2024

  29. [29]

    Mozolenko, M

    V. Mozolenko, M. Fadeeva, and L. N. Shchur, Compar- ison of the Microcanonical Population Annealing Algo- rithm with the Wang–Landau Algorithm, Physical Re- view E110, 045301 (2024)

    work page 2024

  30. [30]

    Burovski, W

    E. Burovski, W. Janke, M. Guskova, and L. Shchur, Ac- ceptance Rate Is a Thermodynamic Function in Local Monte Carlo Algorithms, Physical Review E100, 063303 (2019)

    work page 2019

  31. [31]

    Onsager, Crystal Statistics

    L. Onsager, Crystal Statistics. I. A Two-Dimensional Model with an Order-Disorder Transition, Physical Re- view65, 117 (1944)

    work page 1944

  32. [32]

    F. Y. Wu, The Potts Model, Reviews of Modern Physics 54, 235 (1982)

    work page 1982

  33. [33]

    Metropolis, A

    N. Metropolis, A. W. Rosenbluth, M. N. Rosenbluth, A. H. Teller, and E. Teller, Equations of state calcula- tions by fast computing machines, J. Chem. Phys.21, 1087 (1953)

    work page 1953

  34. [34]

    R. J. Glauber, Time-dependent statistics of the Ising model, J. Math. Phys.4, 294 (1963)

    work page 1963