REVIEW 1 major objections 3 minor 6 references
Sharp asymptotics for the tree-completion time in cylindrical Hastings--Levitov$(0)$
T0 review · 1 major / 3 minor · reviewed 2026-08-10 · deepseek-v4-flash
Pith's one-line read For every fixed particle size λ, the expected tree-completion time in cylindrical Hastings–Levitov(0) is asymptotically log N / (2λ).
desk verdict A clean second-moment proof of the lower bound, but the load-bearing recursion is imported from an unpublished preprint with a shared author; the referee should verify that input. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The zero-colored set M_k(0) is the union of base-circle arcs that have not yet produced a tree after k particles, and Z_k = |M_k(0)| is its length. The exact first-moment recursion E[Z_{k+1} | F_k] = q Z_k, with q = 1 − a_δ/π ≈ 1 − λ/(πN), is imported from the earlier work of Chen, Procaccia and Zong. The second moment is controlled by the one-step kernel H(A) = E[Z_{k+1}^2 | M_k(0) = A], which equals the $L^{2}$ norm of the circular convolution D_δ * 1_A, where D_δ is the derivative of the inverse slit map. The key inequality is the deterministic excess bound 0 ≤ H(A) − $q^{2}$ |A|^2 ≤ ($2a_δ^{2}$/π) |A|, whose error term is O($N^{{-2}}$) and therefore negligible on the critical scale; iterating it yields the summed $L^{2}$ estimate that closes the second-moment argument.
What would settle it
Simulate CHL_N for fixed λ and increasing N, recording the index of the last particle that creates a new tree; the claim predicts that the mean of this index divided by N log N approaches π/λ. A persistent deviation from that ratio, or a direct measurement showing that the expected one-step shrinkage factor of the uncovered set depends on how the uncovered arcs are arranged, would falsify the lower bound.
Extended reading notes
Core claim
The central claim is Theorem 1.4: for every fixed λ > 0, E[ω_{N,λ}]/log N → 1/(2λ) as N → ∞, where ω_{N,λ} is the last time a new tree is born on the base circle. The new contribution is the lower bound liminf_{N→∞} E[ω_{N,λ}]/log N ≥ 1/(2λ), obtained by showing that tree births persist until nearly (π/λ) N log N particles have arrived. Concretely, with m_N = ⌊(1−ε)(π/λ) N log N⌋, the count Y_N of tree births after m_N satisfies EY_N ∼ (π/λ) N^ε and $EY_N^{2}$ ≤ (1+o(1)) (EY_N)^2; the Paley–Zygmund inequality then gives P(Y_N > 0) → 1, which forces ω_{N,λ} to be at least roughly log N/(2λ).
Load-bearing premise
The proof assumes that the expected one-step shrinkage factor of the not-yet-treed part of the base circle is exactly the same constant q for every possible arrangement of that part, importing this recursion from earlier work as a black box.
Editorial extensions
If this is right
- The conjecture of Chen, Procaccia and Zong is resolved: the expected tree-completion time is asymptotic to log N / (2λ), with no further correction to leading order.
- The expected number of particles attached by tree completion is (π/λ) N log N (1 + o(1)).
- With probability tending to one, at least one new tree is born after any fixed fraction of the critical threshold, so tree births are not cut off prematurely.
- The renormalized zero-set length Z_k / q^k is a martingale whose predictable quadratic variation up to the critical time is o(1) in expectation, meaning the untouched set stays close to its deterministic mean throughout the relevant period.
Reading between the lines
- The proof uses only the convolution structure of the inverse-slit derivative and the size of its Fourier modes, which suggests the logarithmic coefficient 1/(2λ) may be insensitive to fine details of the conformal map and could survive for other Hastings–Levitov parameters with a modified q.
- The martingale formulation suggests a stronger concentration statement: Z_k / q^k should stay near its mean up to the critical scale, so one could hope for concentration of ω_{N,λ} around its mean, not just first-order asymptotics.
- A natural next problem, not addressed here, is the behavior of the surviving infinite tree after tree completion; this paper fixes the time at which competition ends, leaving the post-competition growth rate open.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proves that for the cylindrical Hastings–Levitov(0) process with fixed particle size lambda > 0, the expected tree-completion time satisfies E[omega_{N,lambda}] / log N -> 1/(2 lambda). The upper bound is taken from a previous preprint by Chen, Procaccia, and Zong [1], and the paper's main contribution is a matching lower bound obtained by a second-moment analysis of a discrete coverage process. The key technical step is a deterministic excess bound (Lemma 3.2) for the one-step second-moment kernel, leading to a summed L2 estimate and a Paley–Zygmund argument showing that new trees continue to be born up to the cutoff m_N = floor((1-eps)(pi/lambda) N log N). The proof is concise and clearly written, but it rests on a black-box recursion imported from [1].
Significance. If the lower bound is valid, the paper resolves a conjecture from [1] and establishes the sharp first-order asymptotics for the tree-completion time. The method is noteworthy: the excess bound (9) is a clean deterministic statement about the variance of the inverse-slit derivative convolution, and the martingale reformulation in Remark 4.2 gives an instructive alternative view. The paper is transparent about its reliance on [1]; however, the extent to which the central claim depends on an unpublished, same-author preprint is a risk that the journal should weigh.
major comments (1)
- [§2.1, Eq. (13)] The exact configuration-independent first-moment recursion E[Z_{k+1} | F_k] = q Z_k is imported as a black box from [1, Sec. 5], an unpublished preprint sharing an author. This recursion is used in the proof of Proposition 2.5 (Eq. (16)), in the correlation bound (17), and in Proposition 3.3 (Eq. (25)), all of which support the Paley–Zygmund argument in Proposition 4.1. If the true conditional expectation contained a configuration-dependent correction—for instance a term of order a_delta^2 times the number of zero arcs—the cutoff scale m_N could shift and the lower bound (5) might fail. The manuscript should either provide a self-contained proof of (13) or state a precise theorem from [1] and validate that the rate q is configuration-independent. This is a load-bearing dependency, not a routine citation.
minor comments (3)
- [§1.3] The notation a_delta is used in Eq. (7) before it is defined in Eq. (10); consider defining it at first use.
- [Figure 1] The figure is reproduced from [1]; if this is a direct reproduction, permission should be obtained or a note should be added.
- [Acknowledgments] The acknowledgment that 'the main results were obtained by Eureka' is unconventional; the authors may wish to clarify the role of the AI system in deriving the proofs, or omit this statement so that the scientific content stands alone.
Circularity Check
No significant circularity: the lower bound is derived from a prior exact first-moment recursion, not from the target limit.
full rationale
The paper's new contribution is the lower bound (5), and its derivation does not assume the target asymptotic. The argument starts from the exact first-moment recursion E[Z_{k+1}|F_k]=qZ_k (Eq. 13), imported from [1], and then proves, within the paper, an elementary second-moment identity (Prop. 2.5), a deterministic excess bound on the one-step kernel (Lemma 3.2), a summed L2 estimate (Prop. 3.3), and a Paley–Zygmund conclusion (Prop. 4.1). None of these internal steps presuppose E[omega] ~ log N/(2lambda); the constant 1/(2lambda) emerges only at the end from q, the cutoff m_N, and the time-change identity (14), rather than being inserted as an input. Equation (13) is a prior result about the zero-colored length, not a disguised form of the tree-completion-time limit, so the lower bound is not equivalent to its input by construction. The heavy citation of [1], which shares an author, is concerning for provenance but, under the stated rules, counts as independent support because the cited recursion is parameter-free, has stated model assumptions, and does not include the target result. The unverified status of the preprint [1] is a correctness and robustness risk, not a circularity risk. No fitted parameter is renamed as a prediction, and no defined quantity reduces to the conclusion by definition. Therefore the circularity score is 0.
Assumptions & free parameters
assumptions (5)
- domain assumption Marked-configuration coupling of CHL_N to the coverage process on T^1, identifying new-tree births with hits of the zero-colored set.
- domain assumption Exact, configuration-independent first-moment recursion E[Z_{k+1}|F_k] = q Z_k, with q = 1 - a_δ/π.
- domain assumption Inverse-slit derivative formula D_δ(u) with ∫ D_δ = 2πq, and the image-disjointness of distinct arcs used in Lemma 3.1.
- domain assumption Deterministic monotonicity Z_{k+1} ≤ Z_k (Lemma 2.3), from inverse-slit interval estimates in [1, §2.1].
- standard math Standard probabilistic and analytic tools: Parseval's identity, Paley-Zygmund inequality, monotone convergence, Doob's inequality.
Cite this review
Pith. "Pith review of Sharp asymptotics for the tree-completion time in cylindrical Hastings--Levitov$(0)$." pith.science (2026). https://pith.science/paper/AW66PY5I
@misc{pith2026260806925,
author = {Pith},
title = {Pith review of: Sharp asymptotics for the tree-completion time in cylindrical Hastings--Levitov$(0)$},
year = {2026},
howpublished = {\url{https://pith.science/paper/AW66PY5I}},
note = {Machine review of arXiv:2608.06925}
}
abstract
Let $\mathrm{CHL}_N$ be the cylindrical Hastings--Levitov aggregation process with parameter $0$ on a cylinder of width $N$ with particles of fixed size $\lambda>0$, and let $\omega_{N,\lambda}$ be its tree-completion time --- the last time at which a new tree is born on the base circle. Chen, Procaccia and Zong proved the sharp upper bound $\mathbb{E}[\omega_{N,\lambda}]\le(1+\varepsilon)(\log N)/(2\lambda)$ and conjectured the matching limit. Here we prove the matching lower bound, and therefore \[ \lim_{N\to\infty}\frac{\mathbb{E}[\omega_{N,\lambda}]}{\log N}=\frac{1}{2\lambda} \qquad\text{for every fixed }\lambda>0 . \]
Figures
Reference graph
Works this paper leans on
-
[1]
G. Chen, E. B. Procaccia and Y. Zong,One-arm domination time in cylindrical Hastings–Levitov(0), preprint, arXiv:2507.11028, 2025
arXiv 2025
-
[2]
Durrett,Probability: Theory and Examples, 5th ed., Camb
R. Durrett,Probability: Theory and Examples, 5th ed., Camb. Ser. Stat. Probab. Math., vol. 49, Cambridge University Press, Cambridge, 2019
work page 2019
-
[3]
M. B. Hastings and L. S. Levitov,Laplacian growth as one-dimensional turbulence, Phys. D116(1998), no. 1–2, 244–252
work page 1998
-
[4]
J. Norris and A. Turner,Hastings–Levitov aggregation in the small-particle limit, Comm. Math. Phys.316 (2012), no. 3, 809–841
work page 2012
-
[5]
E. B. Procaccia and A. Zhuchenko,Cylindrical Hastings–Levitov, preprint, arXiv:2301.12737, 2023
arXiv 2023
-
[6]
T. A. Witten and L. M. Sander,Diffusion-limited aggregation, a kinetic critical phenomenon, Phys. Rev. Lett. 47(1981), no. 19, 1400–1403. School of Mathematical Sciences, University of Science and Technology of China, Hefei, Anhui 230026, P.R. China Email address:fuxm@ustc.edu.cn School of Mathematical Sciences, University of Science and Technology of Chi...
work page 1981
Reviewed August 10, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.