Pith. sign in

REVIEW 1 major objections 5 minor 7 references

Random walks in a field of soft traps and criticality for the dissipative Abelian Sandpile Model

T0 review · 1 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read A one-dimensional random walk among soft traps with spacings obeying $|I_{j+1}|=c|I_j|^2$ has finite expected lifetime for $c<1$ and infinite expected lifetime for $c>1$, placing the phase transition at $c=1$ in this trap landscape.

desk verdict Sharp c=1 threshold for super-exponentially spaced traps is a real advance; proof needs a fix in Lemma 4.1 but the bound survives. read the letter →

arxiv 2507.00562 v2 pith:LN6YVKEX submitted 2025-07-01 math-ph math.MPmath.PR

classification math-phmath.MPmath.PR MSC 60G5060J10
keywords randomwalksofttrapsexpectedsurvivaltimedissipativeabeliansandpilecriticalitytransitionsuper-exponentialtrapspacingsembeddedtailinvariance
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 simple random walk on the nonnegative integers, reflected at 0, with soft traps at sites $x_0=01$, provided $c>|I_1|^{-1}$. Because finite expected avalanche size in the dissipative abelian sandpile is equivalent to finite expected survival time of the associated trapped walk, this sharp transition at $c=1$ is claimed to say exactly how much dissipation a one-dimensional sandpile can absorb before losing criticality. The paper also proves that finiteness of the expected survival time depends only on the tail of the trap configuration, and gives an explicit example where enlarging every interval turns a critical landscape into a non-critical one.

What carries the argument

The argument runs through the embedded random walk $\xi_n$, the walk's sequence of visits to trap locations. For any prescribed itinerary of right moves, stays, and left moves between traps, Lemma 4.1 bounds the product of path probability times expected inter-arrival time against the all-right itinerary, with factors $2/c$ per stay and $2/(c^2|I_1|^2)$ per left move, plus matching lower bounds for nonnegative itineraries. The recurrence $|I_{j+1}|=c|I_j|^2$ makes these factors telescope, yielding an explicit formula for the all-right contribution, and Theorem 3.3 (finiteness depends only on the tail) lets the proof replace the initial interval by an arbitrarily distant one.

What would settle it

Evaluate $E(\tau)$ numerically for the recursion with $|I_1|=3$ on truncated systems at $c=0.9$ and $c=1.1$: if the mean survival time does not stay bounded at $c=0.9$, or does not grow without bound at $c=1.1$ as the truncation is removed, the phase boundary is not at $c=1$.

Watch

Extended reading notes

Core claim

The central claim is a sharp phase transition in the expected lifetime of the walk. For traps placed so that $|I_{j+1}|=c|I_j|^2$ with $c>|I_1|^{-1}$, Theorem 3.1 states that $E(\tau)<\infty$ when $c<1$ and $E(\tau)=\infty$ when $c>1$. The proof obtains upper and lower bounds on $E(\tau)$ that approach each other as the first interval length $|I_1|$ grows, then uses a tail-invariance result to push the first interval arbitrarily far away, making the bounds pinch at $c=1$. Through the previously established correspondence between this trapped random walk and the dissipative abelian sandpile, the same dichotomy is stated for the sandpile: non-critical (finite expected avalanche size) for $c<1$, critical (infinite expected avalanche size) for $c>1$.

Load-bearing premise

The load-bearing premise is that the correction factors in Lemma 4.1—a stay multiplies a contribution by at most $2/c$ and a left move by at most $2/(c^2|I_1|^2)$, with matching lower bounds—are exact consequences of the recursion $|I_{j+1}|=c|I_j|^2$; if those factors differ, the critical constant shifts, and the sandpile conclusion additionally assumes the cited equivalence between finite survival time and finite avalanche expectation.

Editorial extensions

If this is right

  • For $c<1$, the corresponding one-dimensional dissipative sandpile has finite expected avalanche size: dissipation disrupts criticality.
  • For $c>1$, expected avalanche size is infinite even though the trap density is zero: criticality survives a super-exponentially sparse set of dissipative sites.
  • Changing any finite initial segment of the trap landscape cannot flip finiteness of $E(\tau)$; only the tail matters.
  • The expected survival time is not monotone in interval lengths: a landscape with all intervals four times larger can have finite $E(\tau)$ where the smaller-interval landscape has infinite $E(\tau)$.

Reading between the lines

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

  • At $c=1$ itself the theorem is silent; the upper and lower bounds pinch at that point, so a natural conjecture—not made by the paper—is that $E(\tau)$ diverges exactly at $c=1$, possibly with a slowly growing partial-sum rate.
  • The same comparison method could be applied to other growth rules such as $|I_{j+1}| = c |I_j|^p$; if the correction factors still telescope, the analogue of the critical $c$ would be a function of $p$, a statement the paper does not contain.
  • In the sandpile reading, the non-monotonicity example suggests that moving dissipative sites farther apart can reduce the average avalanche size; that consequence for avalanches is implicit rather than proved here.
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.

Referee Report

1 major / 5 minor

Summary. The paper studies a reflected simple random walk on the nonnegative integers, killed with probability 1/3 at each trap site, with trap locations 0 = x0 < x1 < x2 < ... and interval lengths |I_j| = x_j - x_{j-1} obeying the recursion |I_{j+1}| = c |I_j|^2. The main result (Theorem 3.1) is that, for c > 1/|I1| and under the integrality restriction of Remark 3.1, the expected survival time is finite for c < 1 and infinite for c > 1. The proof passes through an embedded walk on traps and compares arbitrary embedded histories with the all-right history via Lemma 4.1, then establishes a tail-similarity statement (Theorem 3.3) that lets the first interval be shifted away. The paper also gives a counterexample to monotonicity of the expected survival time in interval lengths. Through the equivalence established in [6], the phase transition is interpreted as the threshold between non-critical and critical dissipative one-dimensional abelian sandpile behavior.

Significance. If the proof is repaired, the paper gives a sharp, parameter-free phase transition in this random-walk model: the upper and lower bounds in Section 4 pinch exactly at c = 1, with no fitting parameters. The explicit embedded-walk computation, the tail property that reduces the transition to the asymptotic growth of intervals, and the concrete non-monotonicity example are all valuable contributions. The sandpile conclusion is conditional on the external equivalence proved in [6], which is not re-proved here, but the random-walk theorem itself is self-contained and does not use sandpile criticality as an input. I found no circularity in the main derivation.

major comments (1)
  1. [Section 4.1, Eq. (51)] The displayed combination of (49) and (50) for the substitution sigma = -1 is algebraically incorrect in the case s < i - 2. Direct computation gives P(A_{kappa_{s,-1}})/P(A_kappa) = 2^{I(Theta=1)} |I_{f-2}| |I_{f-1}| / |I_Theta|^2, using that validity of kappa_{s,-1} forces Theta >= 1. Equation (51) instead contains the extra factor (2/3)(1/(2|I_{Theta+1}|)) in the s < i - 2 term, making the displayed expression too small by a factor that can be as large as O(|I_1|). As a consequence, the displayed inequality (52) with the bracket value 2/3 for s < i - 2 does not follow from (51) as written. The corrected ratio is still at most 2 |I_{f-2}| |I_{f-1}| / |I_1|^2, so the final bound (52) and hence Lemma 4.1(24) remain valid, but the proof as written needs a repaired calculation before the manuscript can be accepted.
minor comments (5)
  1. [Section 4.2.2, Eq. (79)] The displayed algebra in (79) is incorrect: 3 - 2(1 - 1/|I1|)^2 = 1 + 4/|I1| - 2/|I1|^2, not 1 + 4/|I1| - 1/|I1|^2. The subsequent integer-gap argument is unaffected, but the equation should be corrected.
  2. [Section 2, Eq. (28)] The notation |I_i + I(i=0)| in the stay-at-the-same-trap transition probability is ambiguous. For i = 0 the intended expression appears to be 2/3 (1 - 1/|I1|); please rewrite this formula so that the i = 0 case is explicit.
  3. [Example 5.1] The wording 'increasing the size of a single interval' is imprecise: the comparison in Example 5.1 enlarges every interval by a factor of four. The rigorous single-interval counterexample is given in Section 5.1, so the earlier example should be described as a motivation rather than as a single-interval perturbation.
  4. [Section 5.1, Eq. (107)-(109)] The numerical bounds in (107)-(109) are stated for P_+ and P_- in [1/2, 2/3], while inequality (105) only proves the range [1/3, 2/3]. Please justify that a landscape can be chosen with P_+, P_- in [1/2, 2/3], or extend the numerical verification to the full range.
  5. [Throughout] There are several small presentation issues: a summation index typo in the definition of N in (5), the duplicated equations (42)/(44) and (43)/(45), and the misspelling 'don not' in Section 5. These should be cleaned up in revision.

Circularity Check

0 steps flagged · score 0.0 of 10

The random-walk derivation is self-contained; the sandpile interpretation is cited from prior work and is not an input to the theorem.

full rationale

The central derivation chain is self-contained. The expected survival time is decomposed via the embedded random walk in equations (19) and (23), and Lemma 4.1 bounds each contribution against the all-right path contribution P(A1)E(Yi|A1^+) using the explicit transition probabilities (28) and the recursion |I_{j+1}| = c|I_j|^2. The upper bound (24) and lower bound (25) are combined with the explicit formula (26) for the all-right path, yielding convergence when (1 + 2/c + 2/(c^2|I1|^2))(c/3) < 1 and divergence when the analogous lower-bound factor exceeds 1. The threshold c = 1 is obtained by taking k large in the tail-similarity argument (Theorem 3.3), so the conditions (14) and (15) pinch to c = 1. No parameter is fitted to the quantity being predicted, and the critical value c = 1 emerges from the upper and lower bounds rather than being imposed. The sandpile-criticality correspondence is stated through the equivalence established in [6], a prior published paper by overlapping authors, but that equivalence is not used in the proof of Theorem 3.1; the random-walk theorem is mathematically independent of it. The skeptical note about an algebraic error near equation (51) is a correctness concern about an intermediate calculation and does not indicate circularity. No circular step is present.

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

The model parameters c and |I1| are variables of the theorem, not fitted constants, so no free parameters are counted. No new entities are introduced; the soft traps and the dissipative sandpile correspondence come from prior work. The main external inputs are standard one-dimensional random-walk facts and the [6] equivalence.

assumptions (4)
  • standard math Standard gambler's ruin hitting probabilities and expected hitting times for one-dimensional simple random walk (formulas (29) and (36)).
    Used throughout Lemma 4.1 to compute transition probabilities and inter-arrival expectations.
  • domain assumption The equivalence from [6] between finiteness of expected avalanche size in the dissipative abelian sandpile and finiteness of the expected survival time of the associated trapped random walk.
    Bridges the random-walk theorem to the sandpile criticality claim in the abstract; cited, not re-proved.
  • domain assumption The random walk on N reflected at the origin with soft-trap killing probability 1/3 faithfully represents the one-dimensional dissipative sandpile with stable heights 0,1 and dissipative heights 0,1,2.
    Defines the correspondence between trap dynamics and sandpile topplings, following [6].
  • domain assumption The interval lengths |I_j| are increasing in j for c>|I1|^{-1}, which is used in several inequalities in Lemma 4.1 and Theorem 3.2.
    Relies on c>|I1|^{-1} and |I1|>=1; used, for example, in (42) and (70).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Random walks in a field of soft traps and criticality for the dissipative Abelian Sandpile Model." pith.science (2026). https://pith.science/paper/LN6YVKEX

@misc{pith2026250700562,
  author       = {Pith},
  title        = {Pith review of: Random walks in a field of soft traps and criticality for the dissipative Abelian Sandpile Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LN6YVKEX}},
  note         = {Machine review of arXiv:2507.00562}
}
read the original abstract

Motivated by the dissipative abelian sandpile model, we analyze the trajectories of a one-dimensional random walk in a landscape of soft traps. These traps, placed at increasing distances from each other, correspond to dissipative sites in the associated dissipative abelian sandpile model. We identify a critical growth rate of the sizes of intervals between successive traps where there is a transition between finiteness and non-finiteness of the expected survival time of the random walk. This corresponds to a transition between non-criticality and criticality of the associated dissipative abelian sandpile model. Therefore, in this setting, we thus identify precisely how much dissipation can be added to the original abelian sandpile model in order to disrupt its criticality.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

7 extracted references · 7 canonical work pages

  1. [6]

    Ruszel, and Ellen Saada

    Frank Redig, Wioletta M. Ruszel, and Ellen Saada. Non-criticality criteria for abelian sandpile models with sources and sinks. Journal of Mathematical Physics , 59(6), 2018

  2. [1]

    Self-organized criticality: An explanation of the 1/f noise

    Per Bak, Chao Tang, and Kurt Wiesenfeld. Self-organized criticality: An explanation of the 1/f noise. Physical review letters , 59(4):381, 1987

  3. [2]

    The abelian sandpile and related models

    Deepak Dhar. The abelian sandpile and related models. Physica A: Statistical Mechanics and its applications , 263(1-4):4–25, 1999

  4. [3]

    Proof of breaking of self-organized criticality in a nonconservative abelian sandpile model

    Tomoko Tsuchiya and Makoto Katori. Proof of breaking of self-organized criticality in a nonconservative abelian sandpile model. Physical Review E , 61(2):1183, 2000

  5. [4]

    The infinite volume limit of dissipative abelian sandpiles

    Christian Maes, Frank Redig, and Ellen Saada. The infinite volume limit of dissipative abelian sandpiles. Communications in mathematical physics , 244:395–417, 2004. 28

  6. [5]

    Approaching criticality via the zero dissipation limit in the abelian avalanche model

    Antal A J´ arai, Frank Redig, and Ellen Saada. Approaching criticality via the zero dissipation limit in the abelian avalanche model. Journal of Statistical Physics , 159:1369–1407, 2015

  7. [7]

    Criticality of the abelian sandpile in dimension 1

    Adrien Rezzouk. Criticality of the abelian sandpile in dimension 1. In preparation, 2025. 29

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.