Lattice $\phi^{4}$ field theory as a multi-agent system of financial markets

Dimitrios Bachtis

REVIEW 3 major objections 5 minor 34 references

Lattice $\phi^{4}$ field theory as a multi-agent system of financial markets

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A lattice $\phi^4$ field theory with frustrated dynamics is introduced as a multi-agent model of financial markets, and the paper claims it reproduces fat-tailed returns and clustered volatility comparable to FTSE 100 data.

desk verdict A plausible continuous-field extension of spin-based market models, but the sampler as written does not target the stated action, so the stylized facts are not yet attributable to the model; deserves review with heavy revision. read the letter →

arxiv 2411.15813 v1 pith:WBXL6A6S submitted 2024-11-24 cond-mat.dis-nn cs.CEcs.LGcs.MAhep-lat

Dimitrios Bachtis This is my paper · ORCID

classification cond-mat.dis-nncs.CEcs.LGcs.MAhep-lat

keywords phi4latticefieldtheorymulti-agentsystemfinancialmarketstylizedfactsfat-tailedreturnsclusteredvolatilityfrustrateddynamicsGibbssamplingFerrenberg-Swendsenacceptance-rejection

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to establish that a $\phi^4$ lattice field theory, normally a model of scalar fields in particle physics, can be reinterpreted as a multi-agent system of a financial market. Each lattice site is an agent, the continuous field value encodes the direction and strength of a buy or sell decision, and two competing interactions push agents to imitate their neighbors or to resist the consensus when it becomes extreme. The author argues that this frustrated dynamics produces, without fitting to the FTSE series, the two canonical empirical regularities: fat-tailed distributions of returns and clustered volatility. The payoff of the claim is a physics-derived generative model of market behavior whose continuous degrees of freedom can represent heterogeneous agent convictions, something binary Ising-type models cannot do.

What carries the argument

The argument is carried by a constructive decomposition of the single-site conditional distribution. The paper writes $p(\phi_i|\phi_j,j\in n_i)=f_1(\phi_i)f_2(\phi_i)/\int f_1 f_2$, with $f_1$ the Gaussian $\mathcal{N}(\mu,\sigma^2)$ where $\mu=(c_1+h)/(2c_2)$, $\sigma=1/\sqrt{2c_2}$, $c_1=\sum_{j\in n_i}\phi_j$, $c_2=2+m^2/2$, and $f_2(\phi_i)=\exp(-(\lambda/4)\phi_i^4)$. A candidate $\phi_i'$ is drawn from the Gaussian and accepted with probability $f_2(\phi_i')$ through a Ferrenberg–Swendsen step, recomposing the quartic theory from a Gaussian proposal. The frustration term $h$ shifts the Gaussian mean; when the absolute magnetization becomes large, $h$ dominates and forces sign flips, producing the intermittent transitions that the paper reads as bubbles and crashes.

What would settle it

Compute the exact single-site conditional distribution of $\phi_i$ given all other fields under the action that includes $h(\phi_i)=-a\phi_i|\frac{1}{V}\sum_j\phi_j|$; the quadratic coefficient in its exponent will not be $c_2$ once the dependence of the magnetization on $\phi_i$ is taken into account. A direct check would compare the marginal distribution of $\phi_i$ and the autocorrelation of $|r_{\phi^4}(t)|$ generated by the paper's update with those obtained from an exact sampler, such as a full Metropolis step or the correct conditional distribution. If the marginals disagree, the observed fat tails and clustered volatility are artifacts of a sampler that does not simulate the stated model.

Watch

Extended reading notes

Core claim

The central claim is that the multi-agent $\phi^4$ theory, defined by the ferromagnetic nearest-neighbor action together with the frustration term $h(\phi_i) = -a \phi_i |\frac{1}{V}\sum_j \phi_j|$, reproduces financial stylized facts. With couplings $m^2=-3.0$, $\lambda=0.7$, $a=5.0$ on a $64\times64$ lattice, the model yields intermittent phases with large fluctuations, a fat-tailed return histogram with excess kurtosis $\kappa_{\phi^4}=3.37$, a complementary cumulative distribution of absolute returns displayed alongside the $\theta=2.3$ power-law comparison used for spin models, and slow decay of the autocorrelation of $|r_{\phi^4}(t)|$. The paper interprets metastable configurations as expectation bubbles and turbulent configurations as chartist-dominated phases, and claims that the continuous nature of $\phi_i$, whose sign encodes buy or sell direction and whose magnitude encodes conviction, gives the framework representational capacity beyond Ising-based agent models.

Load-bearing premise

The load-bearing premise is that the Monte Carlo update used to generate every configuration targets the true probability distribution of the model, even though the frustration term involves the same variable that is being updated, and the paper does not check the sampler's correctness.

Editorial extensions

If this is right

If the central claim is correct, the $\phi^4$ field theory becomes a viable generative model for financial time series that exhibits fat tails and volatility clustering endogenously, from agent interactions rather than from external shocks.
The model generalizes the Ising spin model and shares its universality class, so it is expected to match the successes of Ising-based market models while additionally encoding continuous conviction strengths.
The constructive decomposition suggests that the same Gibbs-plus-acceptance/rejection procedure could turn other lattice field theories into multi-agent market models.
Solving the inverse problem over $m^2$, $\lambda$, and $a$ would in principle allow the model to be calibrated to empirical return distributions and used to generate synthetic FTSE-like data.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

An extension the author leaves implicit is to define trading volume as $\sum_i |\phi_i|$ and test whether the model reproduces the empirical volume–volatility correlation and the leverage effect; the continuous magnitudes are natural volume proxies.
A direct ablation not performed in the paper would switch off the frustration term $h$ and check whether the fat tails and volatility clustering disappear, isolating the mechanism the paper attributes to frustrated dynamics.
Because the return definition uses magnetization differences, a natural follow-up is to test whether multi-period returns obey the cubic-law scaling $\theta\approx3$ observed in real markets, rather than only the single-period $\theta=2.3$ comparison shown.
The same constructive sampling route could be applied to $O(N)$-symmetric field theories to build multi-asset or multi-sector agent models with correlated continuous opinions.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a readable proof-of-concept that a phi4 scalar field with a Bornholdt-style frustration term produces fat-tailed returns and clustered volatility for one hand-picked parameter set. The continuous degree of freedom is a genuinely new twist in econophysics agent models—no one in the cited literature has done this before. But the paper's main numerical claim is undercut by an inexact sampler, and there are a few avoidable slips. It is not ready for prime time, but it earns a serious referee.

What is new and good: the model itself. Mapping agents to continuous phi values and reading |phi| as conviction is a natural extension of Ising-based spin markets, and the frustration term h = -a phi |M| is a sensible continuous analogue of Bornholdt's competing interaction. The Markov-field decomposition into a Gaussian and quartic factor, recomposed with Ferrenberg-Swendsen, is standard for the bare phi4 theory. The paper is honest: it says the kurtosis depends on sample size and couplings, and it ships code.

The soft spots, in order of seriousness. First, the sampler as written does not target the stated distribution. Eq. (7) includes the updated site in the magnetization, but Eq. (8) only shifts the Gaussian mean. The exact conditional has a sign-dependent |phi_i + S| term that modifies the quadratic coefficient, not just the mean. Unless h is computed from sites other than i—which Eq. (7) doesn't say—this is not a Gibbs sampler for the phi4 + frustration action. That makes the attribution of the stylized facts to the model unproven; they could be artifacts of the update rule. A standard Metropolis or an exact conditional with h excluding site i would settle it. Second, Eq. (9) writes the log-return as the ratio of log prices rather than a difference. Minor but embarrassing. Third, kurtosis is quoted without uncertainties, there is no comparison with the corresponding Ising/Bornholdt model, and the abstract promises a critical-exponent investigation that never appears in the results.

Who is this for? People working on statistical-physics models of markets, who will see the continuous-field extension as a plausible step forward and will want the sampler fixed. The central idea is sound enough for a serious referee to engage with, and code availability helps. My recommendation: send it to review, with the expectation of heavy revision. The author should fix the sampler, correct the return formula, add error bars, and show a baseline comparison before the empirical claims can be trusted.

Referee Report

3 major / 5 minor

Summary. The manuscript introduces a two-dimensional lattice phi^4 field theory augmented by a fictitious field h = -a phi_i |(1/V) sum_i phi_i|, interpreted as a multi-agent model of financial markets in which agents imitate neighbors (the standard phi^4 kinetic term) and also follow or oppose the majority/minority opinion (the frustration term). The author proposes a constructive decomposition of the local conditional probability into a Gaussian factor and a quartic factor, recomposed via a Ferrenberg-Swendsen accept/reject step, and reports numerical results on an L=64 lattice with m^2=-3.0, lambda=0.7, a=5.0. The paper compares model-generated return series and histograms with FTSE 100 daily data, reports excess kurtosis kappa_phi4=3.37 versus kappa_FTSE=10.42, shows a complementary cumulative distribution with a line proportional to theta=2.3, and displays a slowly decaying autocorrelation of absolute returns. The author concludes that the continuous-field model reproduces fat-tailed returns and clustered volatility and offers representational capacity beyond Ising-based agent models.

Significance. If the central claim were established, the paper would offer a novel continuous-field extension of spin-based multi-agent market models, with a plausible mechanism for intermittent dynamics and bubbles. The manuscript has clear strengths: the code is publicly available, the Markov-random-field decomposition is explicit, and the model produces visibly non-Gaussian return distributions and volatility clustering from a simple local update. The comparison with FTSE data, while qualitative, is a useful sanity check. However, the significance hinges on whether the Monte Carlo sampler actually samples the stated phi^4-plus-frustration model; the current derivation does not establish this, and the quantitative claims (kurtosis, power-law exponent) are presented without uncertainties or fitted values. The paper is therefore a promising idea whose validation is incomplete.

major comments (3)

[Sec. III, Eqs. (7)-(8)]
[Sec. IV, Eq. (9)]
[Sec. IV, kurtosis and Fig. 4]

minor comments (5)

[Sec. IV, Fig. 5]
[Sec. IV, model parameters]
[Sec. I, Ref. [25]]
[Sec. III, after Eq. (7)]
[Sec. II, Eq. (4)]

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the model is hand-parameterized and compared to empirical data, not fitted from it.

full rationale

The paper's derivation chain is self-contained. The lattice action Eq. (1) and the frustration term Eq. (7) define the model; the conditional update in Eqs. (3)-(8) is presented as a constructive sampling scheme, and the parameters m2=-3.0, lambda=0.7, a=5.0 are fixed before simulation, not inferred from FTSE observables. The stylized facts (fat tails, clustered volatility, kurtosis values, power-law scaling) are measured from the simulated configurations and then compared with the empirical FTSE series; the FTSE data enter only as a benchmark for comparison, and no fitted parameter is renamed as a prediction. Citations [25]-[27] are from the author's prior work and are used as background for the Markov property of phi4, but the conditional distribution is written out explicitly in Eq. (3) from the local action, so the argument does not reduce to those citations. The potential inexactness of the Gibbs/rejection sampler in Sec. III noted in the skeptical review is a correctness or bias concern, not a circularity: the observed stylized facts are not equivalent to the model's inputs by construction, even if the sampler does not target the intended distribution.

Assumptions & free parameters 3 free parameters · 6 assumptions · 1 invented entities

The model depends on three hand-chosen couplings, a domain-specific mapping from magnetization to returns, and an unverified sampler assumption about how the self-dependent field h enters the Gaussian proposal. No new physical entity is postulated, but the fictitious field is an invented interaction term whose validity is judged only by the qualitative output of the simulation.

free parameters (3)

m^2 (squared mass coupling) = -3.0
Chosen by hand to place the system in the broken-symmetry phase; the observed kurtosis and volatility clustering depend on this choice.
lambda (quartic coupling) = 0.7
Chosen by hand together with m^2; it sets the strength of the phi^4 term and influences the fatness of the return distribution.
a (fictitious field strength) = 5.0
Chosen by hand to control the frustration from the majority and minority compulsion term; it is not calibrated to FTSE data.

assumptions (6)

standard math The phi4 lattice theory on a square lattice is a Markov random field and satisfies the local Markov property, so Gibbs sampling by conditional distributions is justified.
Invoked in Sec. II around Eq. 3 to decompose p(phi_i | neighbors); this is a standard property of lattice field theories with finite-range actions.
standard math For c2 > 0 and sufficiently small c3, sampling from Gaussian f1 and accepting with f2 via a Ferrenberg-Swendsen step recovers the conditional distribution of the phi4 theory.
Stated in Sec. II as the condition for the constructive decomposition and recomposition; it underlies the sampler used for all simulations.
domain assumption The difference in magnetization between successive configurations can be interpreted as stock returns.
Adopted from spin-model econophysics, specifically Ref. [18], and used in Eq. 10 to define r_phi4(t).
ad hoc to paper The term h = -a phi_i |(1/V) sum_i phi_i| encodes the compulsion to follow or oppose majority and minority opinions.
Introduced in Eq. 7 with no derivation from market microstructure; its behavioral interpretation as fundamentalist and chartist pressure is assumed.
ad hoc to paper The sampler remains valid when h is included by shifting only the mean of the Gaussian proposal, with the quadratic coefficient c2 left unchanged.
This is the fragile premise in Eqs. 7-8: h contains the site being updated, so the conditional variance should change unless h is evaluated from other sites only.
domain assumption FTSE 100 daily closing values are a valid empirical benchmark for stylized facts.
Used in Sec. IV for comparison; the reliability and preprocessing of the empirical data are not discussed.

invented entities (1)

Fictitious majority and minority field h(phi_i) = -a phi_i |(1/V) sum_i phi_i|
purpose: Creates frustrated ferromagnetic and antiferromagnetic dynamics so agents sometimes flip against the collective opinion; used to represent fundamentalists and chartists.
It is a modeling construct, not an observable; no separate falsifiable prediction outside the simulation is provided, and its strength a is hand-chosen.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Lattice $\phi^{4}$ field theory as a multi-agent system of financial markets." pith.science (2026). https://pith.science/paper/WBXL6A6S

@misc{pith2026241115813,
  author       = {Pith},
  title        = {Pith review of: Lattice $\phi^4$ field theory as a multi-agent system of financial markets},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WBXL6A6S}},
  note         = {Machine review of arXiv:2411.15813}
}

abstract

We introduce a $\phi^{4}$ lattice field theory with frustrated dynamics as a multi-agent system to reproduce stylized facts of financial markets such as fat-tailed distributions of returns and clustered volatility. Each lattice site, represented by a continuous degree of freedom, corresponds to an agent experiencing a set of competing interactions which influence its decision to buy or sell a given stock. These interactions comprise a cooperative term, which signifies that the agent should imitate the behavior of its neighbors, and a fictitious field, which compels the agent instead to conform with the opinion of the majority or the minority. To introduce the competing dynamics we exploit the Markov field structure to pursue a constructive decomposition of the $\phi^{4}$ probability distribution which we recompose with a Ferrenberg-Swendsen acceptance or rejection sampling step. We then verify numerically that the multi-agent $\phi^{4}$ field theory produces behavior observed on empirical data from the FTSE 100 London Stock Exchange index. We conclude by discussing how the presence of continuous degrees of freedom within the $\phi^{4}$ lattice field theory enables a representational capacity beyond that possible with multi-agent systems derived from Ising models.

Figures

Figures reproduced from arXiv: 2411.15813 by the authors.

**Figure 2.** FIG. 2. Returns [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 5.** FIG. 5. Autocorrelation function for the absolute returns [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 34 canonical work pages

[1]

Bouchaud and M

J.-P. Bouchaud and M. Potters, Theory of Financial Risk and Derivative Pricing: From Statistical Physics to Risk Management, Cambridge Books No. 9780521741866 (Cambridge University Press, 2009)

work page 2009
[2]

Mandelbrot, The Variation of Certain Speculative Prices, The Journal of Business 36, 394 (1963)

B. Mandelbrot, The Variation of Certain Speculative Prices, The Journal of Business 36, 394 (1963)

work page 1963
[3]

E. F. Fama, Mandelbrot and the stable paretian hypoth- esis, The Journal of Business 36, 420 (1963)

work page 1963
[4]

R. N. Mantegna and H. E. Stanley, Scaling behaviour in the dynamics of an economic index, Nature 376, 46 (1995)

work page 1995
[5]

Ghashghaie, W

S. Ghashghaie, W. Breymann, J. Peinke, P. Talkner, and Y. Dodge, Turbulent cascades in foreign exchange mar- kets, Nature 381, 767 (1996)

work page 1996
[6]

Galluccio, G

S. Galluccio, G. Caldarelli, M. Marsili, and Y.-C. Zhang, Scaling in currency exchange, Physica A: Statistical Me- chanics and its Applications 245, 423 (1997)

work page 1997
[7]

Lux and M

T. Lux and M. Marchesi, Scaling and criticality in a stochastic multi-agent model of a financial market, Na- ture 397, 498 (1999)

work page 1999
[8]

S. BORNHOLDT, Expectation bubbles in a spin model of markets: Intermittency from frustration across scales, International Journal of Modern Physics C 12, 667 (2001), https://doi.org/10.1142/S0129183101001845

work page doi:10.1142/s0129183101001845 2001

Show all 34 references

[9]

Giardina and J.-P

I. Giardina and J.-P. Bouchaud, Bubbles, crashes and intermittency in agent based market models, The Euro- pean Physical Journal B - Condensed Matter and Com- plex Systems 31, 421 (2003)

2003
[10]

Zhou and D

W.-X. Zhou and D. Sornette, Self-organizing ising model of financial markets, The European Physical Journal B 55, 175 (2007)

2007
[11]

P. Bak, M. Paczuski, and M. Shubik, Price variations in a stock market with many agents, Physica A: Statistical Mechanics and its Applications 246, 430 (1997)

1997
[12]

I. G. J-P. Bouchaud and M. Mezard, On a uni- versal mechanism for long-range volatility cor- relations, Quantitative Finance 1, 212 (2001), https://doi.org/10.1088/1469-7688/1/2/302

2001 doi
[13]

M. Levy, H. Levy, and S. Solomon, Microscopic Simu- lation of the Stock Market: the Effect of Microscopic Diversity, Journal de Physique I 5, 1087 (1995)

1995
[14]

Cont and J.-P

R. Cont and J.-P. Bouchaud, Herd behavior and aggre- gate fluctuations in financial markets, Macroeconomic Dynamics 4, 170–196 (2000)

2000
[15]

Bouchaud and R

J.-P. Bouchaud and R. Cont, A langevin approach to stock market fluctuations and crashes, The European Physical Journal B - Condensed Matter and Complex Systems 6, 543 (1998)

1998
[16]

A. M. Y. M. A Corcos, J-P Eckmann and D. Sornette, Imitation and contrarian behaviour: hyperbolic bubbles, crashes and chaos, Quantitative Finance 2, 264 (2002), https://doi.org/10.1088/1469-7688/2/4/303

2002 doi
[17]

Challet and Y.-C

D. Challet and Y.-C. Zhang, Emergence of coopera- tion and organization in an evolutionary game, Physica A: Statistical Mechanics and its Applications 246, 407 (1997)

1997
[18]

Kaizoji, S

T. Kaizoji, S. Bornholdt, and Y. Fujiwara, Dynamics of price and trading volume in a spin model of stock mar- kets with heterogeneous agents, Physica A: Statistical Mechanics and its Applications 316, 441 (2002)

2002
[19]

S. M. Krause and S. Bornholdt, Spin models as mi- crofoundation of macroscopic market models, Physica A: Statistical Mechanics and its Applications 392, 4048 (2013)

2013
[20]

Takaishi, Dynamical cross-correlation of multiple time series Ising model, Evolutionary and Institutional Eco- nomics Review 13, 455 (2016)

T. Takaishi, Dynamical cross-correlation of multiple time series Ising model, Evolutionary and Institutional Eco- nomics Review 13, 455 (2016)

2016
[21]

TAKAISHI, Simulations of financial mar- kets in a potts-like model, International Jour- nal of Modern Physics C 16, 1311 (2005), https://doi.org/10.1142/S0129183105007923

T. TAKAISHI, Simulations of financial mar- kets in a potts-like model, International Jour- nal of Modern Physics C 16, 1311 (2005), https://doi.org/10.1142/S0129183105007923

2005 doi
[22]

Bouchaud, Crises and collective socio-economic phe- nomena: Simple models and challenges, Journal of Sta- tistical Physics 151, 567 (2013)

J.-P. Bouchaud, Crises and collective socio-economic phe- nomena: Simple models and challenges, Journal of Sta- tistical Physics 151, 567 (2013)

2013
[23]

Sornette, Physics and financial economics (1776–2014): puzzles, ising and agent-based mod- els, Reports on Progress in Physics 77, 062001 (2014)

D. Sornette, Physics and financial economics (1776–2014): puzzles, ising and agent-based mod- els, Reports on Progress in Physics 77, 062001 (2014)

2014
[24]

Geman and D

S. Geman and D. Geman, Stochastic relaxation, gibbs distributions, and the bayesian restoration of images, IEEE Transactions on Pattern Analysis and Machine In- telligence P AMI-6, 721 (1984)

1984
[25]

Bachtis, G

D. Bachtis, G. Aarts, and B. Lucini, Quantum field- theoretic machine learning, Phys. Rev. D 103, 074510 (2021)

2021
[26]

Bachtis, G

D. Bachtis, G. Aarts, and B. Lucini, Quantum field theo- ries, markov random fields and machine learning, Journal of Physics: Conference Series 2207, 012056 (2022)

2022
[27]

Bachtis, G

D. Bachtis, G. Aarts, and B. Lucini, Machine learning with quantum field theories, PoS LA TTICE2021, 201 (2022)

2022
[28]

A. M. Ferrenberg and R. H. Swendsen, Optimized monte carlo data analysis, Phys. Rev. Lett. 63, 1195 (1989)

1989
[29]

Schaich and W

D. Schaich and W. Loinaz, Improved lattice measurement of the critical coupling in ϕ4 2 theory, Phys. Rev. D 79, 056008 (2009)

2009
[30]

Bachtis, G

D. Bachtis, G. Aarts, F. Di Renzo, and B. Lucini, Inverse renormalization group in quantum field theory, Phys. Rev. Lett. 128, 081603 (2022)

2022
[31]

Bachtis, Reducing finite-size effects with reweighted renormalization group transformations, Phys

D. Bachtis, Reducing finite-size effects with reweighted renormalization group transformations, Phys. Rev. E 109, 014125 (2024)

2024
[32]

Pagan, The econometrics of financial markets, Journal of Empirical Finance 3, 15 (1996)

A. Pagan, The econometrics of financial markets, Journal of Empirical Finance 3, 15 (1996)

1996
[33]

Bachtis, Disordered lattice glass ϕ4 quantum field the- ory (2024), arXiv:2407.06569 [hep-lat]

D. Bachtis, Disordered lattice glass ϕ4 quantum field the- ory (2024), arXiv:2407.06569 [hep-lat]

2024 arXiv
[34]

https://github.com/dbachtis/phifm

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Bouchaud and M

J.-P. Bouchaud and M. Potters, Theory of Financial Risk and Derivative Pricing: From Statistical Physics to Risk Management, Cambridge Books No. 9780521741866 (Cambridge University Press, 2009)

work page 2009

[2] [2]

Mandelbrot, The Variation of Certain Speculative Prices, The Journal of Business 36, 394 (1963)

B. Mandelbrot, The Variation of Certain Speculative Prices, The Journal of Business 36, 394 (1963)

work page 1963

[3] [3]

E. F. Fama, Mandelbrot and the stable paretian hypoth- esis, The Journal of Business 36, 420 (1963)

work page 1963

[4] [4]

R. N. Mantegna and H. E. Stanley, Scaling behaviour in the dynamics of an economic index, Nature 376, 46 (1995)

work page 1995

[5] [5]

Ghashghaie, W

S. Ghashghaie, W. Breymann, J. Peinke, P. Talkner, and Y. Dodge, Turbulent cascades in foreign exchange mar- kets, Nature 381, 767 (1996)

work page 1996

[6] [6]

Galluccio, G

S. Galluccio, G. Caldarelli, M. Marsili, and Y.-C. Zhang, Scaling in currency exchange, Physica A: Statistical Me- chanics and its Applications 245, 423 (1997)

work page 1997

[7] [7]

Lux and M

T. Lux and M. Marchesi, Scaling and criticality in a stochastic multi-agent model of a financial market, Na- ture 397, 498 (1999)

work page 1999

[8] [8]

S. BORNHOLDT, Expectation bubbles in a spin model of markets: Intermittency from frustration across scales, International Journal of Modern Physics C 12, 667 (2001), https://doi.org/10.1142/S0129183101001845

work page doi:10.1142/s0129183101001845 2001