Forest Fire Model on $\mathbb{Z}_{+}$ with Delays

Satyaki Bhattacharya; Stanislav Volkov

REVIEW 3 major objections 5 minor 15 references

Forest Fire Model on $\mathbb{Z}_{+}$ with Delays

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 Any positive delay in fire spread makes an infinite forest fire inevitable almost surely.

desk verdict A genuinely new infinite-fire phase transition in a natural generalization of the one-sided forest fire model; proofs need tightening but the claims are credible. read the letter →

arxiv 2411.13419 v2 pith:SUIPNZIL submitted 2024-11-20 math.PR

Satyaki Bhattacharya , Stanislav Volkov This is my paper · ORCID

classification math.PR MSC 60G5560K3560F20

keywords forestfiremodelinfinitedelayedspreadburningtimeself-organizedcriticalityphasetransitionBorel-Cantellirandomwalk

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

The paper studies a one-dimensional forest fire model in which fires start only at the origin, burn trees for a positive random time, and take a positive random time to jump from one tree to the next. Its central claim is that if the spread delay has any chance of being positive, then almost surely some fire will burn infinitely many trees, an 'infinite fire' that cannot occur in the classic zero-delay model. The authors see this as a genuinely new phenomenon introduced by delay, and they go on to map when it happens: with delayed spread an infinite fire is the rule rather than the exception, while with instantaneous spread it can never happen.

What carries the argument

The central machinery is the sequence of successive maxima $m_1 < m_2 < \cdots$ of the rightmost burnt site, together with the $\sigma$-algebras $F_k$ generated by the process up to the time $T_k$ when site $m_k$ stops burning. Lemma 2.2 gives a uniform lower bound on the probability that cumulative delays grow at least linearly; combined with the product in (2.1) this yields a strictly positive conditional probability $\epsilon_0 \delta_2$ that the next fire is infinite, which the conditional Borel-Cantelli lemma turns into an almost-sure event.

What would settle it

For a fixed delay law with $P(\Delta>0)>0$, simulate the process and measure the conditional probability, given the history up to $T_k$, that the next fire reaching beyond $m_k$ continues forever; Theorem 2.1 predicts this probability is never below $\epsilon_0 \delta_2 > 0$. If simulations with large $k$ show it decaying to zero, the uniform lower bound is false.

Watch

Extended reading notes

Core claim

Theorem 2.1 asserts that whenever $P(\Delta > 0) > 0$, with probability one there is a fire that never stops: $P(\inf\{k : n_k = \infty\} < \infty) = 1$. The proof tracks the successive rightmost points $m_k$ reached by fires and shows, via a conditional Borel-Cantelli argument, that from any such maximum there is a uniformly positive chance that the next fire burns forever. The same framework yields a uniqueness result when burning is instantaneous and delays have positive finite variance (exactly one infinite fire), a phase transition at average delay $c/x$ with threshold $c=1$, and a no-infinite-fire theorem when spread is instantaneous.

Load-bearing premise

The induction step of Theorem 2.1 assumes that at each site, the infinitely many delayed attempts to ignite the next site will eventually occur while that next site has a tree, even though earlier fires may have burnt it; this renewal property is asserted rather than proved.

Editorial extensions

If this is right

If the main theorem holds, the model exhibits an almost-sure infinite fire under any non-degenerate delay distribution, a phenomenon absent from the zero-delay models of [9] and [15].
With instantaneous burning and non-degenerate finite-variance delays, the infinite fire is unique: no later fire can reach infinity again.
With a fixed positive delay $a>0$, the process after the infinite fire starts is exactly the 2009 zero-delay forest fire model, so the infinite fire acts as a reset point after which known finite-fire statistics apply.
For location-dependent delays with mean about $c/x$, there is a sharp threshold: $c \le 1$ suppresses infinite fires, while $c > 1$ makes one almost sure.
When spread is instantaneous, no infinite fire occurs even with positive burn times, and first-passage times grow logarithmically in the distance.

Reading between the lines

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

The paper leaves open the exact law of the time $T$ when the infinite fire begins; a natural next step is to estimate its tail as a function of the delay distribution.
Because Theorem 2.1 needs only $P(\Delta>0)>0$, heavy-tailed delays should also produce infinite fires; quantifying how the waiting time scales with the delay tail would test the robustness of the mechanism.
The constant-delay coupling suggests that after the infinite fire, the distribution of fire sizes is governed by the known zero-delay model, which could give explicit asymptotics for the number of trees burnt by later fires.
In the instant-spread setting, Conjecture 4.4 would imply $f_{n,1}=O(\log n)$; checking numerically whether $f_{n,1}/\log n$ stays bounded for large $n$ would provide evidence for the conjecture and for Theorem 4.5.

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

Quick take on arXiv:2411.13419. The paper adds delays in fire spread and nonzero burn times to the Z+ forest fire model with ignition at 0, and proves a new phenomenon: if delays are sometimes positive, an infinite fire occurs a.s., which is impossible in the zero-delay models of Comets–Menshikov–Volkov and Volkov. The cleanest result is the phase transition in Section 3.2: for spread times with mean about c/x, the infinite fire exists iff c>1. That is new and worth knowing.

The paper does a lot of things right. The graphical representation is helpful. The constant-delay coupling with Volkov's model is elegant, and the proof that there is at most one infinite fire in the i.i.d. positive-mean delay case uses a zero-mean random walk recurrence argument that is correct and short. Section 4 has a nice double-exponential lower bound on successive maxima that does not depend on the conjecture.

The soft spots are all in the proof details, not the main ideas. Theorem 2.1's induction claims that fires reach every site infinitely often; the written justification is too quick and needs a renewal argument after each burn at x+1. The repair is standard and I don't doubt the statement. In the proof of Theorem 3.6(d), the step labeled (3.4) is not justified by the argument given: conditioning on the first fire to reach n does not make the earlier-fires event independent of the current delays in the way the text asserts. The inequality itself looks true (the event that the i-th fire reaches n is increasing in the delays, while the 'small sum' event is decreasing), but it needs a clean monotonicity argument. Finally, Theorem 4.5 is explicitly conditional on Conjecture 4.4; that is clearly stated, so it is a limitation, not an error.

None of these look fatal. The central claims are supported by the sketches, and the gaps are of the kind a referee can ask to be filled without changing the results. The citation pattern is clean; the only self-citation is to the base model in [15], which is appropriate. This is a genuine advance for the forest fire / SOC subfield, not a breakthrough for all of probability.

I'd send it out for review. For a reading group, it is a good example of how a small model change creates a real phase transition, and it has several instructive Borel–Cantelli arguments.

Referee Report

3 major / 5 minor

Summary. The paper studies a generalization of the one-dimensional forest fire model with ignition at the origin, introduced in [15], by allowing non-zero burning times θ and random spread delays Δ. The main results are: (i) if P(Δ>0)>0, then an infinite fire occurs almost surely (Theorem 2.1); (ii) for θ≡0 with i.i.d. delays of positive mean and positive variance, there is exactly one infinite fire (Theorem 3.1); (iii) for location-dependent delays with asymptotic order c/x, there is a threshold at c=1 separating a.s. absence and a.s. occurrence of infinite fire (Theorem 3.6); (iv) if Δ≡0, no infinite fire occurs (Theorem 4.1), and the record maxima satisfy a doubly exponential lower bound (Theorem 4.7). A conditional O(log n) bound for the first burning time is proved under Conjecture 4.4 (Theorem 4.5).

Significance. The paper establishes a robust new phenomenon for a natural one-dimensional forest-fire model: random spread delays can cause a fire to propagate to infinity, a behavior that is impossible in the zero-delay models of [9] and [15]. The threshold result in Section 3.2 is particularly interesting, as it connects the tail behavior of the delay distribution to a percolation-type exponent. The arguments are self-contained and use standard tools (renewal, conditional Borel-Cantelli, random walk recurrence, Hoeffding's inequality). The main unconditional theorems appear correct in outline and are potentially useful for further work on self-organized critical systems. The paper is honest about the dependence of Theorem 4.5 on Conjecture 4.4, though this dependence should be made more prominent in the presentation.

major comments (3)

[Section 2, proof of Theorem 2.1] The induction step asserting that a new tree at site x+1 will be burnt by a fire from x is stated without proof; the text only notes that a tree appears eventually and that f_{x,i}→∞. Because a tree may be burnt before the next fire from x attempts to spread, one needs a renewal argument: after each burn at x+1, the appearance time of the next tree is independent exponential, and the first subsequent spread attempt from x occurs after this appearance with probability one. Please supply this argument explicitly. In addition, the invocation of the conditional Borel-Cantelli lemma should be replaced by a short proof (e.g., bounding P(E_k) by (1−ε0δ2)^{k−1}) or a precise statement of the lemma.
[Section 4, Theorem 4.5 and Lemma 4.6(b)] These results are conditional on Conjecture 4.4, which is not proved. The manuscript labels 4.5 as 'Theorem' without making the conditional nature prominent in the section or abstract. If the conjecture remains open, the authors should relabel the statement as a conditional proposition and clearly state in the introduction that the only conditional result of the paper is Theorem 4.5 and Lemma 4.6(b).
[Section 3.2, proof of (3.4)] The equality P(X|Y,D_{n,i}) = P(X|Y) is asserted without justification. It is true because X is determined by the first i−1 fires, which are independent of the i-th spread times and of the tree appearances after the i-th fire starts, but this independence should be spelled out. Please add a sentence explaining why X is jointly independent of (Y,D_{n,i}).

minor comments (5)

[Section 4, Conjecture 4.4] Conjecture 4.4 contains a typo: 'such f_{m_i,2}' should read 'such that f_{m_i,2}'.
[General notation] The same symbol F is used for the sigma-algebra F_k in Section 2 and for the event F_n in Lemma 4.3; please use different notation (e.g., \mathcal{F}_k and E_n) to avoid ambiguity.
[Lemma 2.2] The chain of inequalities leading to the definition of δ2 is hard to follow; consider presenting the conditioning argument in a separate paragraph to improve readability.
[Theorem 3.6(d)] The condition on Δ_x is rendered as 'P(lim |Δ_x|?x=0)=1' in the text; please make the mathematical expression unambiguous (presumably P(lim |Δ_x|√x=0)=1).
[Abstract] The abstract refers to arXiv:0907.1821; for consistency with the body, cite it as [15] instead.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main theorems are derived from the model definition and standard probabilistic tools, and the cited prior work is used as an external base case rather than as a hidden restatement of the new results.

full rationale

The paper's central claim, Theorem 2.1, is proven directly from the model rules and Lemma 2.2, which is itself an elementary large-deviations estimate; no fitted parameter or target quantity is renamed as a prediction. The only self-citation, to Volkov [15], appears as the base model being generalized and as a benchmark for a coupling after an infinite fire has already been established, e.g. in Theorem 3.3 and Proposition 3.4. That coupling is a direct consequence of the definitions (θ = 0, Δ = a, and the infinite fire path acting as a shifted time origin), not a circular import of the result. The use of [15, Lemma 1.1] in Section 3.1 supplies an external distribution for the zero-delay process and does not assume the delay-model conclusions. The proof of Theorem 3.6 uses independent Borel-Cantelli and Hoeffding bounds. The minor expositional gaps noted by a skeptical reader, such as making the conditional Borel-Cantelli step fully rigorous, are not circularity: they are issues of proof detail, not of conclusions being equivalent to inputs by construction. Overall, the derivation chain is self-contained relative to its stated assumptions.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No numerical parameters are fitted to data. The model takes the distributions of theta and Delta as inputs; the threshold constant c in Theorem 3.6 is part of the model assumption EDelta_x ~ c/x, not a fitted value. The only extra input beyond the model definition is the unproved Conjecture 4.4 needed for Theorem 4.5.

assumptions (5)

domain assumption Trees appear at each vacant site at rate 1 via independent exponential clocks.
Model definition, Section 1, point (1). This is the source of the e^{-t} vacancy probabilities used throughout.
domain assumption Burning times theta_{x,i} are i.i.d. non-negative with distribution theta; spread delays Delta_{x,i} are i.i.d. non-negative with distribution Delta.
Model definition, Section 1, points (3)-(4). Theorems 2.1 and 3.6 control the tails of these distributions.
domain assumption A burning tree blocks new growth at the same site and blocks fire from the left neighbor; fire spreads to x+1 only after delay Delta after x catches fire, and only if x+1 is occupied.
Model definition, Section 1, points (4)-(5). This mechanism is what makes the infinite fire possible.
standard math Standard probability tools: conditional Borel-Cantelli, Kolmogorov three-series theorem, Hoeffding's inequality, and recurrence of one-dimensional zero-mean random walks.
Used in the proofs of Theorems 2.1, 3.1, and 3.6; these are accepted background results.
ad hoc to paper Conjecture 4.4: there is c >= 1 such that a.s. f_{m_i,2} <= c log m_i for all large i.
Theorem 4.5 and Lemma 4.6(b) are conditional on this unproved conjecture, stated in Section 4 before Theorem 4.5.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Forest Fire Model on $\mathbb{Z}_{+}$ with Delays." pith.science (2026). https://pith.science/paper/SUIPNZIL

@misc{pith2026241113419,
  author       = {Pith},
  title        = {Pith review of: Forest Fire Model on $\mathbbZ_+$ with Delays},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SUIPNZIL}},
  note         = {Machine review of arXiv:2411.13419}
}

abstract

We consider a generalization of the forest fire model on $\mathbb{Z}_+$ with ignition at zero only, studied in [arXiv:0907.1821]. Unlike that model, we allow delays in the spread of the fires as well as the non-zero burning time of individual ``trees''. We obtain some general properties for this model, which cover, among others, the phenomena of an ``infinite fire'', not present in the original model.

Figures

Figures reproduced from arXiv: 2411.13419 by the authors.

**Figure 2.** The black dots correspond to the appearance of trees; the green (red resp.) segments are the periods when a site is occupied by a “healthy” (burning resp.) tree. The shaded areas represent the periods when a burning tree at site x affects site x ` 1. The solid arrows show the fire spreading to a neighbouring tree; the dotted arrows represent the situations when the fire tried to spread unsuccessfully. 1.2 Structure … view at source ↗

**Figure 3.** how the second fire reaching mi can spread very far. reaching mi for the first time is of order log i, which is quite low compared to the order of mi that exceeds expte αiu, as we show later. This indicates that the path (using the graphic representation language) of the first fire reaching mi is not “very bumpy”, and thus it might behave more like a flat path (instantaneous burn). Hence, while we do not claim fmi,2… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 15 canonical work pages

[15]

Forest fires onZ` with ignition only at0

Volkov, Stanislav. Forest fires onZ` with ignition only at0. ALEA 6 (2009), 399–414. 22

work page 2009
[9]

Generalizations of forest fires with ignition at the origin

Comets, Francis; Menshikov, Mikhail; Volkov, Stanislav. Generalizations of forest fires with ignition at the origin. J. Appl. Probab. 60 (2023), no. 2, 418–434

work page 2023
[1]

Seven- dimensional forest fires

Ahlberg, Daniel; Duminil-Copin, Hugo; Kozma, Gady; Sidoravicius, Vladas. Seven- dimensional forest fires. Ann. Inst. Henri Poincar´ e Probab. Stat. 51 (2015), no. 3, 862— 866. 21

work page 2015
[2]

Copernicus, New York (1996)

Bak, P.: How Nature Works: The Science of Self-organized Criticality. Copernicus, New York (1996)

work page 1996
[3]

A signal-recovery system: asymptotic properties, and construction of an infinite-volume process

van den Berg, J.; T´ oth, B. A signal-recovery system: asymptotic properties, and construction of an infinite-volume process. Stochastic Process. Appl. 96 (2001), no. 2, 177–190

work page 2001
[4]

4, 480–501

vandenBerg, J.; Brouwer, R.Self-destructivepercolationRandomStructuresAlgorithms 24 (2004), no. 4, 480–501

work page 2004
[5]

van den Berg, J.; J´ arai, A. A. On the asymptotic density in a one-dimensional self- organized critical forest-fire model. Comm. Math. Phys. 253 (2005), no. 3, 633–644

work page 2005
[6]

Self-organized forest-fires near the critical time

van den Berg, J.; Brouwer, R. Self-organized forest-fires near the critical time. Comm. Math. Phys. 267 (2006), no. 1, 265–277

work page 2006

Show all 15 references

[7]

Near-critical 2D percolation with heavy-tailed impurities, forest fires and frozen percolation

van den Berg, Jacob; Nolin, Pierre. Near-critical 2D percolation with heavy-tailed impurities, forest fires and frozen percolation. Probab. Theory Related Fields 181 (2021), no. 1-3, 211–290

2021
[8]

One-dimensional general forest fire processes

Bressaud, Xavier; Fournier, Nicolas. One-dimensional general forest fire processes. M´ em. Soc. Math. Fr. (N.S.) No. 132 (2013), vi+138 pp

2013
[10]

Cluster growth in the dynamical Erd˝ os- R´ enyi process with forest fires

Crane, Edward; Freeman, Nic; T´ oth, B´ alint. Cluster growth in the dynamical Erd˝ os- R´ enyi process with forest fires. Electron. J. Probab. 20 (2015), no. 101, 33 pp

2015
[11]

and Schwabl, F

Drossel, B. and Schwabl, F. Self-organized critical forest-fire model. Phys. Rev. Lett. 69, 1629–1632 (1992)

1992
[12]

Drossel, B., Clar, S., Schwabl, F.: Exact Results for the One-Dimensional Self-Organized Critical Forest-Fire Model. Phys. Rev. Lett. 71, 3739–3742 (1993)

1993
[13]

Planar lattices do not recover from forest fires

Kiss, Demeter; Manolescu, Ioan; Sidoravicius, Vladas. Planar lattices do not recover from forest fires. Ann. Probab. 43 (2015), no. 6, 3216–3238

2015
[14]

Principles of random walk

Spitzer, Frank. Principles of random walk. Second edition. Grad. Texts in Math., Vol. 34. Springer-Verlag, New York-Heidelberg, 1976

1976

Pith tools

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

[1] [15]

Forest fires onZ` with ignition only at0

Volkov, Stanislav. Forest fires onZ` with ignition only at0. ALEA 6 (2009), 399–414. 22

work page 2009

[2] [9]

Generalizations of forest fires with ignition at the origin

Comets, Francis; Menshikov, Mikhail; Volkov, Stanislav. Generalizations of forest fires with ignition at the origin. J. Appl. Probab. 60 (2023), no. 2, 418–434

work page 2023

[3] [1]

Seven- dimensional forest fires

Ahlberg, Daniel; Duminil-Copin, Hugo; Kozma, Gady; Sidoravicius, Vladas. Seven- dimensional forest fires. Ann. Inst. Henri Poincar´ e Probab. Stat. 51 (2015), no. 3, 862— 866. 21

work page 2015

[4] [2]

Copernicus, New York (1996)

Bak, P.: How Nature Works: The Science of Self-organized Criticality. Copernicus, New York (1996)

work page 1996

[5] [3]

A signal-recovery system: asymptotic properties, and construction of an infinite-volume process

van den Berg, J.; T´ oth, B. A signal-recovery system: asymptotic properties, and construction of an infinite-volume process. Stochastic Process. Appl. 96 (2001), no. 2, 177–190

work page 2001

[6] [4]

4, 480–501

vandenBerg, J.; Brouwer, R.Self-destructivepercolationRandomStructuresAlgorithms 24 (2004), no. 4, 480–501

work page 2004

[7] [5]

van den Berg, J.; J´ arai, A. A. On the asymptotic density in a one-dimensional self- organized critical forest-fire model. Comm. Math. Phys. 253 (2005), no. 3, 633–644

work page 2005

[8] [6]

Self-organized forest-fires near the critical time

van den Berg, J.; Brouwer, R. Self-organized forest-fires near the critical time. Comm. Math. Phys. 267 (2006), no. 1, 265–277

work page 2006