Controlling Cost in Sandpile Models Through Local Adjustment of Drive

Behnam Parsaeifard; Saman Moghimi-Araghi

REVIEW 4 major objections 5 minor 39 references

Controlling Cost in Sandpile Models Through Local Adjustment of Drive

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read A local, neighbor-based rule for choosing where to add sand reduces large-avalanche probability in the BTW and Manna sandpile models, and an intermediate search frequency or depth minimizes total avalanche plus search cost.

desk verdict A plausible local control scheme for sandpiles whose quantitative cost optimum rests on an unstated normalization assumption. read the letter →

arxiv 1908.04991 v1 pith:SH3FDSYK submitted 2019-08-14 cond-mat.stat-mech

Behnam Parsaeifard , Saman Moghimi-Araghi This is my paper · ORCID

classification cond-mat.stat-mech PACS 05.65.+b

keywords self-organizedcriticalitysandpilemodelsBTWmodelMannadrivemodificationavalanchecostfinite-sizescalingcontrol

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 asks whether a purely local change in how a sandpile is driven—checking the height of a randomly selected site and its neighbors, then dropping the grain on the lowest (Min) or highest (Max) site among them—can shrink the largest avalanches in two classic self-organized criticality models, BTW and Manna. It finds that Max drive modestly suppresses large avalanches in BTW while keeping its critical scaling, whereas Min drive removes ordinary single-exponent criticality from BTW and lets avalanches reach sizes comparable to the whole lattice. The Manna model, by contrast, keeps its critical exponents under both drives. Assigning a concave power-law cost to avalanche size and a linear cost to each local search, the paper reports an intermediate drive frequency ($\mu^* \approx 0.58$ for BTW Max, $\approx 0.68$ for Manna Max at $L = 512$, $\alpha = 1.5$, $b/c = 10^5$) and an intermediate search depth ($n = 3$ for the BTW costs tested) that minimize total cost. If correct, this means catastrophic-event risk in such systems can be steered downward using only local information, with a finite overhead for the search.

What carries the argument

The central object is the local neighborhood drive: at each step select one site at random, inspect the height of that site and its (first, then possibly second) neighbors, and deposit the grain on the site of minimum or maximum height among those inspected. The argument is carried by the probability distribution $p_{\mu,n}(s)$ of avalanche sizes under this drive, the cost functional $\langle\text{cost}\rangle = \int p_{\mu,n}(s) c s^{\alpha} \, ds + b$ times the search frequency or depth, and the finite-size scaling ansatz $p(s) = s^{-\tau} g(s/L^D)$, which turns the average cost into a moment $\langle s^{\alpha}\rangle = L^{\alpha+\tau-1} f_\alpha(\mu)$ that grows with system size while the search cost $b\mu$ stays size-independent. That tension between a size-growing avalanche benefit and a size-independent search overhead is what produces the interior optimum in $\mu$ and $n$.

What would settle it

A direct check is to rerun the BTW Max simulations for $L = 512$, $\alpha = 1.5$, $b/c = 10^5$ and record explicitly the number of drive events that produce no avalanche; if $p(s)$ in Eq. (2) is conditional on avalanches, recomputing the average over all drive events should remove or shift the minimum at $\mu \simeq 0.58$. Also, directly measuring $f_\alpha(\mu)$ for two system sizes and checking whether $\mu^*$ moves toward 1 as $L$ increases would settle the large-system claim.

Watch

Extended reading notes

Core claim

The authors claim that replacing purely random sand addition by a local Max drive—always adding to the highest site among the selected site and its neighbors—reduces the probability of very large avalanches in the BTW model without changing its universality class, while the corresponding Min drive pushes BTW away from ordinary criticality, amplifying its multi-scaling and shifting the avalanche-size cutoff to sizes comparable to the whole lattice. In the Manna model both drives leave the critical exponents essentially unchanged. With a cost function $A(s) = c s^{\alpha}$ ($\alpha > 1$) for avalanche size $s$ plus a search cost $b$ times the frequency or depth of the local search, the total cost as a function of the mixing probability $\mu$ has an interior minimum ($\mu^* \simeq 0.58$ for BTW Max and $\simeq 0.68$ for Manna Max at $L = 512$, $\alpha = 1.5$, $b/c = 10^5$), while the avalanche cost alone decreases monotonically with $\mu$. For large enough systems the moment integral scales as $L^{\alpha+\tau-1}$, so full Max drive wins in the thermodynamic limit. The optimal search depth for BTW is $n = 3$ for the tested costs, with the avalanche cost decaying exponentially in $n$ in two stages corresponding to nearest and next-nearest neighbors.

Load-bearing premise

The quoted optimal values of $\mu$ and $n$ assume that the avalanche-cost moment and the per-search cost $b\mu$ are measured per drive event in the same normalization; if the size distribution is normalized over avalanches only, the balance between the two terms changes and the interior minimum may be an artifact.

Editorial extensions

If this is right

In both models a Max drive at full frequency lowers the expected avalanche cost to about two-thirds of the ordinary random-drive value for the BTW parameters examined, and the reduction is larger for larger $\alpha$.
Because the avalanche-moment contribution grows with $L^{\alpha+\tau-1}$ while search cost is independent of $L$, for sufficiently large systems the optimal strategy is always full maximum drive ($\mu = 1$), while for small systems the search cost can make any modification counterproductive.
The Min drive is counterproductive for cost in BTW: it raises mean cost by up to a factor of four and makes the avalanche distribution develop a peak near $s \sim 14$–$16$ with a cutoff extended to about $10^6$ on a $256 \times 256$ lattice.
For BTW, the avalanche cost as a function of search depth $n$ falls exponentially in two stages (decay constants $n_0 \simeq 3.2$ for nearest neighbors and $n_0 \simeq 12.2$ for next-nearest neighbors), so most of the benefit is obtained from shallow local searches; for Manna the initial drop is even steeper but no clean exponential form is found.
The Manna model's critical exponents under Min drive ($\tau = 1.27 \pm 0.01$, $D = 2.73 \pm 0.05$) are unchanged by the drive modification, whereas BTW's Min-drive data fail a single-exponent finite-size collapse, indicating the modification amplifies BTW's multi-scaling.

Reading between the lines

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

If the reported interior optimum survives careful normalization, the same trade-off should appear in any SOC system whose avalanche moment grows with system size and whose search cost is constant: an optimal intervention frequency that tends to 1 as $L$ grows, with a crossover that could be measured directly by extracting $f_\alpha(\mu)$ at fixed $L$.
The two-stage exponential decay of cost with $n$ suggests the relevant control variable is not simply the number of sites inspected but the topological distance reached by the search; a testable extension would compare searches that inspect $n$ sites all at distance 1 against searches that inspect fewer sites at distance 2.
A practical extension the paper does not pursue is time-dependent drive: alternating periods of Max drive with periods of random drive may prevent the slow build-up of maximum-height clusters that Min drive induces, potentially reducing cost further than any fixed $\mu$.
One could also test the paper's cost picture on systems with absorbing states or non-conservative dynamics, where the local Max rule may have a different effect on the avalanche cutoff and therefore on the optimal search depth.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 decent numerical study of a genuinely local drive protocol for BTW and Manna sandpiles. The qualitative finding—Max drive suppresses large avalanches, Min drive makes them worse in BTW but Manna is robust—is credible and worth knowing. The quantitative optimization in Section IV is the soft spot: the optimum search frequency and depth are computed from a cost integral whose normalization is never pinned down.

What's new: [34] used non-local knowledge of all height-2 sites on a graph; here the drive only looks at the chosen site and its neighbors. They check both BTW and Manna, report a robustness asymmetry, and add a search-depth cost tradeoff. That's a real extension. The Manna FSS collapse is clean, and the claim that the Manna model absorbs the perturbation is supported by the height-relaxation experiment in Section III.

Credit where due: the cost function explicitly avoids the negative-cost pathology of [34] by assigning zero benefit to no-avalanche events. The exponential decay of avalanche cost with search depth is a nice empirical structure, and they fit it with maximum likelihood.

The main problem: Eq. (2) defines the mean avalanche cost as the integral of p_{mu,n}(s) A(s) ds, and they add b*mu as a per-event search cost. But p(s) is never specified as conditional on avalanches or over all drive events including silent additions. In sandpiles, many additions don't topple; in Min drive that's the whole point. If p is conditional, the periodic cost should be P_av * integral + b*mu, where P_av depends on mu and L. If p already includes silent events, the plotted PDFs must be renormalized accordingly. So the quoted optima (mu* ~0.58, mu* ~0.68, n*=3) are not quantitatively defensible as they stand. That's not a fatal flaw in the qualitative story, but it means the headline numbers from Section IV shouldn't be trusted until the normalization is cleared up.

Minor soft spots: no error bars on any cost curve; no code or data release, so the numerics aren't independently checkable; and the conclusion's 'criticality is gone' for BTW Min outruns the paper's own evidence—the body says multi-scaling with a cutoff exponent consistent with the original BTW. It's a fair statement that single-scaling criticality is lost, but 'gone' is too strong.

Bottom line: this paper deserves a serious referee. The local protocol and Manna robustness are useful, and the normalization issue is fixable. I'd send it to review, and I'd push the authors to state the normalization, add error bars, and soften the BTW Min conclusion.

Referee Report

4 major / 5 minor

Summary. This manuscript studies the BTW and Manna sandpile models under a locally modified drive: with probability μ the sand grain is added to the minimum or maximum height among the selected site and its neighbors, and alternatively the search depth n is varied. The authors report that maximum drive shifts avalanche-size distributions toward smaller sizes, that minimum drive in the BTW model produces much larger avalanches and no clean finite-size collapse, while the Manna model remains robust to both drive methods. They define an avalanche cost A(s)=c s^α, add a search cost b per local search, and compute an average total cost from measured avalanche distributions. They report an optimal μ*≈0.58 for the BTW Max model (L=512, α=1.5, b/c=10^5), μ*≈0.68 for the Manna Max model, and an optimal search depth n*=3 for the BTW cost parameters, concluding that full maximum drive is increasingly beneficial in large systems.

Significance. If the results were fully supported, the proposed local drive mechanism would be an appealing way to reduce large avalanche risk in SOC models without requiring non-local information, and the reported distinction between the fragile BTW model and the robust Manna model would be a useful contribution to the control of self-organized critical systems. The finite-size-scaling data collapse and quoted exponents for the Manna models, with error bars, are credible strengths, as is the qualitative observation that search-depth performance saturates exponentially in the BTW case. However, the cost-optimization part is the paper's central quantitative claim and is currently not supported to the required standard because the normalization in Eq. (2) is ambiguous and the cost curves have no reported uncertainties. With a corrected normalization, corrected scaling argument, and error estimates, the paper could become a solid contribution.

major comments (4)

[§IV, Eq. (2)] Equation (2) defines the mean avalanche cost as ∫ p_{μ,n}(s) A(s) ds, and this is added to the per-drive-event search cost bμ. The normalization of p_{μ,n}(s) is never specified. If p is conditional on avalanches, the expected cost per drive event is P_av(μ,n,L) ∫ p(s) A(s) ds + bμ, where P_av is the probability that a drive event triggers an avalanche; P_av depends on μ, n, and L and is substantially less than 1 in sandpile models. If p is normalized over all drive events, including silent additions, then it has a probability mass at s=0 and the plotted PDFs and moments must be renormalized accordingly. Because the two normalizations differ by the factor P_av, the reported optimal values μ*≈0.58, μ*≈0.68, and n*=3 are not quantitatively supported until the normalization is specified and the optimization is repeated with the avalanche probability explicitly included.
[§IV, moment scaling argument] The text uses the relation ⟨s^α⟩_μ = L^{α+τ−1} f_α(μ) to argue that for large L the search-cost term bμ is negligible and full maximum drive is optimal. With the finite-size scaling form p(s)=s^{−τ} g(s/L^D) stated in Sec. III, the moment scales as L^{D(α+1−τ)}, not L^{α+τ−1}; the exponent as written does not involve D and has the wrong dependence on τ. This invalidates the large-system argument as presented and needs to be corrected before the conclusion about full drive in large systems can be accepted.
[§III A and §V] The conclusion states that 'in the BTW the criticality is gone with Minimum drive', but Sec. III A reports for the Min BTW model no clean data collapse, a size-dependent slope, and an extrapolated exponent τ∞=1.6±0.3 that is consistent with ordinary BTW within its large error bars. The evidence supports a qualitative change with stronger multiscaling, but not necessarily a loss of criticality. Please state the criterion used for 'criticality is gone' and either support it with a quantitative test or soften the conclusion.
[§IV, Figs. 7–12] The quoted optimal parameters μ* and n* are read from cost curves with no error bars or statistical uncertainties. In Fig. 8 the minimum is shallow, with the cost varying by only a few percent over a wide range of μ, so without bootstrap or repetition errors the reported optimum values are not robust. Please provide error estimates on all cost curves and, ideally, release the simulation data or code so the optima can be independently verified.

minor comments (5)

[§IV, Eq. (1)] The text says the cost function should be concave and that α>1 makes it concave, but A(s)=c s^α with α>1 is convex in s. If the intended assumption is increasing marginal cost (risk aversion), the wording should say 'convex'; if diminishing marginal cost is intended, the model requires α<1.
[§IV, paragraph after Eq. (2)] The sentence 'we have shown that with the new drive method, the system is still critical' is too broad: it is not true for the Min BTW model, for which Fig. 3 explicitly shows no single-exponent power law and no clear data collapse.
[General presentation] There are several small presentation issues: 'Tbaldi' should be 'Tebaldi'; the range '10^{-6} - 10^{-5}' should use a consistent dash; and the captions of Figs. 7 and 8 should state the units of the vertical axis and the value of c/b more clearly.
[§IV, Fig. 10 and Fig. 11] The exponential fits for A(n) report n01=3.2 and n02=12.2 and the saturation values A∞1 and A∞2, but no uncertainties are given for these fit parameters, and the cost for searching n neighbors is not written as an equation. Please add confidence intervals and state the search-cost model explicitly.
[Reproducibility] No code or processed data are provided. For a numerical study whose central claims are quantitative optima, a data/code repository would substantially improve reproducibility and allow referees to check the normalization issue raised above.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the optimum drive parameters are obtained by direct simulation and minimization of an explicitly defined cost, not derived from inputs that already contain them.

full rationale

The paper does not exhibit a circular derivation. The central quantitative results are obtained by direct simulation: for each drive modification probability mu and search depth n, the avalanche-size distribution p_{mu,n}(s) is measured, inserted into the explicitly defined cost functional Eq. (2), and the resulting empirical cost curve is minimized, with a per-search cost b*mu or b*n added. This is an optimization of a measured quantity, not a prediction of a quantity from an input that already contains it. The only self-citation, Ref. [18], appears in a list of references on the universality-class controversy and does not support any load-bearing step; it is not used to forbid alternatives or to define the model. The cost form A(s)=c*s^alpha is adopted from the external Ref. [34] and then evaluated independently here. The exponential fits in Fig. 10 are descriptive fits of the cost-versus-search-depth data, and the stated optimum n=3 in Fig. 11 is directly the minimum of the plotted total cost, so no fitted parameter is renamed as a prediction. There is therefore no circular step, and the score is 0. A separate normalization ambiguity in Eq. (2), concerning whether p_{mu,n}(s) is conditional on avalanches or over all drive events, is a potential quantitative weakness, but it is a matter of correctness of the cost estimate, not circularity.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

No new physical entities are introduced. The free parameters are modeling choices for the cost function, the arbitrary search-cost scale, and the fitted exponents and decay constants. The main hidden assumptions are the finite-size scaling form for modified models and the additivity of avalanche and search costs.

free parameters (4)

alpha (cost exponent) = 1.5
Chosen ad hoc so that A(s)=c s^alpha is concave; the optimal mu and n depend on this exponent.
search cost ratio b/c = 10^5 (with b=1 in the simulations)
Arbitrary scale used in Figs. 8-11; it controls whether an interior optimum mu* or n* exists.
BTW search-depth decay parameters = n01=3.2, A_inf1=45560, n02=12.2, A_inf2=44180
Fitted by maximum likelihood to the simulated avalanche cost versus search depth in Fig. 10; used to argue that deeper searches saturate.
Critical exponents tau and D = Max BTW 1.28+/-0.01, 2.7+/-0.1; Min Manna 1.27+/-0.01, 2.73+/-0.05; Min BTW tau_inf=1.6+/-0.3, D=2.8+/-0.1
Obtained from finite-size scaling of simulated PDFs and used to support criticality and the moment-scaling argument.

assumptions (4)

domain assumption The avalanche-size distribution obeys the finite-size scaling form p(s)=s^(-tau) f(s/L^D) for the modified models.
Invoked in Section III to extract exponents and in Section IV to derive moment scaling; for the BTW Min model the paper itself reports no clean data collapse, so the assumption is questionable there.
domain assumption The moment scales as <s^alpha>_mu = L^(alpha+tau-1) f_alpha(mu), with f_alpha independent of L and decreasing convex in mu.
Used in Section IV to argue that large systems favor full Max drive; the L-independence of f_alpha is not proven.
ad hoc to paper Total cost is the sum of average avalanche cost and per-search cost, with p in Eq. (2) normalized over the same events as the search cost.
The paper does not state whether p(s) includes silent drive events; this normalization is load-bearing for the optimality claims.
domain assumption Periodic boundary conditions with bulk dissipation epsilon in [10^-6, 10^-5] keep the systems critical and give qualitatively the same results.
Stated in Section IV as a boundary-effect check; the main simulations use open boundaries.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Controlling Cost in Sandpile Models Through Local Adjustment of Drive." pith.science (2026). https://pith.science/paper/SH3FDSYK

@misc{pith2026190804991,
  author       = {Pith},
  title        = {Pith review of: Controlling Cost in Sandpile Models Through Local Adjustment of Drive},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SH3FDSYK}},
  note         = {Machine review of arXiv:1908.04991}
}

read the original abstract

In this paper we consider sandpile models and modify the drive mechanisms to control the size of avalanches. The modification to the drive mechanism is local. We have studied the scaling behavior of the BTW and Manna models. We have found that the BTW model is more sensitive to the modification than the Manna model. Furthermore we have assigned a cost function to each avalanche and have found an optimum value for the modification to arrive at the lowest cost.

Figures

Figures reproduced from arXiv: 1908.04991 by the authors.

**Figure 2.** FIG. 2: The probability distribution function of ordinary, Min and Max [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** FIG. 3: Probability distribution function of avalanches of Min model fo [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Figures from the paper (9 more)

**Figure 4.** Figure 4: FIG. 4: Average height of the Min Manna and Max Manna model for L= [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

**Figure 5.** Figure 5: FIG. 5: Comparison of probability distribution function of the Min, Ma [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

**Figure 6.** Figure 6: FIG. 6: The Min Manna data collapse. The parameters [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

**Figure 7.** Figure 7: FIG. 7: Mean cost of the BTW system as function of [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

**Figure 8.** Figure 8: FIG. 8: Mean total cost of the BTW system with maximum drive impleme [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]

**Figure 9.** Figure 9: FIG. 9: Mean cost as a function of [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]

**Figure 10.** Figure 10: FIG. 10: The avalanche cost [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]

**Figure 11.** Figure 11: FIG. 11: Total cost as a function of the number of neighbors being [PITH_FULL_IMAGE:figures/full_fig_p011_11.png]

**Figure 12.** Figure 12: FIG. 12: The avalanche cost as a function of the number of neighbo [PITH_FULL_IMAGE:figures/full_fig_p011_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 38 canonical work pages

[34]

Brummitt, and Raissa M

Pierre-Andre Noel, Charles D. Brummitt, and Raissa M. D ’Souza Phys. Rev. Lett. 111, 078701(2013)

work page 2013
[1]

P. Bak, C. Tang, and K. Wiesenfeld, Phys. Rev. Lett. 59, 381 (1987)

1987
[2]

Bak, and C

P. Bak, and C. Tang. Journal of Geophysical Research: Sol id Earth, 94.B11 15635(1989)

work page 1989
[3]

B. D. Malamud, G. Morein, and D. L. Turcotte, Science 281, 1840 (1998)

work page 1998
[4]

G. L. Mamede, N. A. M. Araujo, C. M. Schneider, J. C. de Arau jo, and H. J. Herrmann, Proc. Natl. Acad. Sci. U.S.A. 109, 7191 (2012)

work page 2012
[5]

Haimovici, E

A. Haimovici, E. Tagliazucchi, P. Balenzuela, and D. R. C hialvo, Phys. Rev. Lett. 110, 178101 (2013)

work page 2013
[6]

Dupoyet, H

B. Dupoyet, H. Fiebig, and D. Musgrove, Physica (Amsterd am) 390A, 3120 (2011). 12

work page 2011
[7]

Dobson, B

I. Dobson, B. A. Carreras, V. E. Lynch, and D. E. Newman, Ch aos 17, 026103 (2007)

work page 2007

Show all 39 references

[8]

Dupoyet, Brice, H. R. Fiebig, and D. P. Musgrove. Physica A: Statistical Mechanics and its Applications 390.18 (2011): 3120-3135; Haldane, A. G. and May, R. M. Syste mic risk in banking ecosystems. Nature 469, 351355 (2011)

2011
[9]

Nature Reviews Genetics 12.1 (2011): 56-68

Barabasi, Albert-Laszlo, Natali Gulbahce, and Joseph L oscalzo. Nature Reviews Genetics 12.1 (2011): 56-68

2011
[10]

E. V. Ivashkevich, D. V. Ktitarev and V. B. Priezzhev, Ph ysica A 209, 347 (1994)

1994
[11]

Dhar and S

D. Dhar and S. S. Manna, Phys. Rev. E 49, 2684 (1994)

1994
[12]

Lubeck and K

S. Lubeck and K. D. Usadel, Phys. Rev. E 55 4095 (1997)

1997
[13]

D. V. Ktitarev, S. Lubeck, P. Grassberger, V. B. Priezzh ev, Phys. Rev. E 61, P. 81 (2000)

2000
[14]

De Menech, A

M. De Menech, A. L. Stella, and C. Tebaldi, Phys. Rev. E 58 , 2677 (1998)

1998
[15]

Manna, J

S.S. Manna, J. Phys. A 24 (1991) L363

1991
[16]

Ben-Hur, and O

A. Ben-Hur, and O. Biham. Phys. Rev. E 53(2) (1996): R1317

1996
[17]

Chessa, H

A. Chessa, H. E. Stanley, A Vespignani, S. Zapperi, Phys . Rev. E 59, N1, R12 (1999)

1999
[18]

Asasi, S

H. Asasi, S. Moghimi-Araghi, M. N. Najaﬁ,Physica A: Sta tistical Mechanics and its Applications 419 (2015): 196-202

2015
[19]

Nakanishi, K

H. Nakanishi, K. Sneppen. Phys. Rev. E 55(4) (1997): 4012

1997
[20]

Lubeck, and P

S. Lubeck, and P. C. Heger. Phys. Rev. E 68(5) (2003): 056102

2003
[21]

Dhar, Physica A: Statistical Mechanics and its Appli cations 270(1) (1999): 69-81

D. Dhar, Physica A: Statistical Mechanics and its Appli cations 270(1) (1999): 69-81

1999
[22]

Vespignani, S

A. Vespignani, S. Zapperi, and L. Pietronero. Phys. Rev . E 51(3) (1995): 1711

1995
[23]

Chessa, A

A. Chessa, A. Vespignani, and S. Zapperi. Computer phys ics communications 121 (1999): 299-302

1999
[24]

Pietronero, A

L. Pietronero, A. Vespignani, and S. Zapperi. Phys. Rev . Lett. 72(11) (1994): 1690

1994
[25]

Christensen and Z

K. Christensen and Z. Olami, Phys. Rev. E 48, 3361 (1993)

1993
[26]

Lubeck, Phys

S. Lubeck, Phys. Rev. E 61(1) (2000): 204

2000
[27]

A. L. Stella, and M. De Menech. Physica A: Statistical Me chanics and its Applications 295(1) (2001): 101- 107

2001
[28]

Dickman, and J

R. Dickman, and J. M. M. Campelo. Phys. Rev. E 67(6) (2003): 066111

2003
[29]

Karmakar, , S

R. Karmakar, , S. S. Manna, and A. L. Stella. Phys. Rev. Le tt. 94(8) (2005): 088002

2005
[30]

De Menech, Phys

M. De Menech, Phys. Rev. E 70,(2004) 028101

2004
[31]

Tebaldi, M

C. Tebaldi, M. De Menech, and A. L. Stella. Phys. Rev. Let t. 83(19), (1999) 3952-3955

1999
[32]

De Menech, and A

M. De Menech, and A. L. Stella, Phys. Rev. E 62(4), (2000) R4528- R4531

2000
[33]

Jo urnal of Physics A: Mathematical and Theoretical 48.40 (2015): 405003

Falk, Johannes, Marco Winkler, and Wolfgang Kinzel. Jo urnal of Physics A: Mathematical and Theoretical 48.40 (2015): 405003

2015
[35]

D. O. Cajueiro and R. F. S. Andrade, Eur. Phys. J. B 77, 291 (2010)

2010
[36]

D’Agostino, A

G. D’Agostino, A. Scala, V. Zlatic, and G. Caldarelli, E urophys. Lett. 97, 68 006 (2012)

2012
[37]

Qi and S

J. Qi and S. Pfenninger, Eur. Phys. L. 111(3), 38006 (201 5)
[38]

Bhaumik and S

H. Bhaumik and S. B. Santra, Eur. Phys. L. (Europhysics L etters), 124(4), 46002(2018)

2018
[39]

M. E. J. Newman, M. Girvan, and J. D. Farmer, Phys. Rev. Le tt. 89, 028301 (2002)

2002

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [34]

Brummitt, and Raissa M

Pierre-Andre Noel, Charles D. Brummitt, and Raissa M. D ’Souza Phys. Rev. Lett. 111, 078701(2013)

work page 2013

[2] [1]

P. Bak, C. Tang, and K. Wiesenfeld, Phys. Rev. Lett. 59, 381 (1987)

1987

[3] [2]

Bak, and C

P. Bak, and C. Tang. Journal of Geophysical Research: Sol id Earth, 94.B11 15635(1989)

work page 1989

[4] [3]

B. D. Malamud, G. Morein, and D. L. Turcotte, Science 281, 1840 (1998)

work page 1998

[5] [4]

G. L. Mamede, N. A. M. Araujo, C. M. Schneider, J. C. de Arau jo, and H. J. Herrmann, Proc. Natl. Acad. Sci. U.S.A. 109, 7191 (2012)

work page 2012

[6] [5]

Haimovici, E

A. Haimovici, E. Tagliazucchi, P. Balenzuela, and D. R. C hialvo, Phys. Rev. Lett. 110, 178101 (2013)

work page 2013

[7] [6]

Dupoyet, H

B. Dupoyet, H. Fiebig, and D. Musgrove, Physica (Amsterd am) 390A, 3120 (2011). 12

work page 2011

[8] [7]

Dobson, B

I. Dobson, B. A. Carreras, V. E. Lynch, and D. E. Newman, Ch aos 17, 026103 (2007)

work page 2007