Pith. sign in

REVIEW 1 major objections 4 minor 14 references

Asymptotics for the harmonic descent chain and applications to critical beta-splitting trees

T0 review · 1 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper proves that the harmonic descent chain's gap to its limit has a precise power-law decay with exponent about 1.56735, and that this rate yields central limit theorems for the critical beta-splitting tree.

desk verdict Sharp exponent for the harmonic descent chain, plus CLTs for critical beta-splitting trees; the main argument is sound, with one missing but trivial monotonicity proof to add. read the letter →

arxiv 2505.24821 v2 pith:ZILP2IAG submitted 2025-05-30 math.PR math.CO

Anna Brandenberger , Byron Chin , Elchanan Mossel This is my paper
classification math.PRmath.CO MSC 60F0560C0505C0560J10
keywords harmonicdescentchaincriticalbeta-splittingtreeconvergenceratepower-lawexponentcentrallimittheoremcontractionmethodrecursivesequence
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 the harmonic descent chain, a recursion of weighted averages that controls how clade counts converge in the critical $\beta$-splitting tree model. Its main result is that the recursive sequence decreases and that the gap to its limit obeys $x_n - x = n^{-\gamma_*+o(1)}$, where $\gamma_* \approx 1.56735$ is the solution to an explicit infinite-series equation. The same exponent governs the occupation probabilities of the harmonic descent chain, so the paper answers a rate question that had been left open for this model. With these sharp error estimates, the paper derives central limit theorems for clade counts, copies of fixed fringe subtrees, and total length of the continuous-time tree. If the proof is right, a previously qualitative convergence becomes a usable, parameter-free rate.

What carries the argument

The load-bearing transformation is the recursion on consecutive differences $d_n = x_n - x_{n+1}$. Summation by parts and the explicit form of the split probabilities give $d_n = h_n^{-1}\sum_{i=k}^{n-1} \alpha'_{i,n} d_i$ with coefficients $\alpha'_{i,n}$ that are exactly computable from harmonic sums. Assuming a power law $d_i \sim i^{-\gamma}$, the paper approximates the coefficient sum by an integral using Euler-Maclaurin summation and compares the divergent part with the growth of $h_n$; the comparison forces the infinite-series equation that defines $\gamma_*$. The induction is stable because the auxiliary function $g'(x) = -(\gamma+1)x^{1-\gamma} + (1+\gamma x^{-\gamma})/(1-x)^2 + \gamma x^{-\gamma-1}\log(1-x)$ is shown to be positive on $(0,1)$ for $2<\gamma<3$, which lets the Euler-Maclaurin remainder be bounded by $f(n-1)-f(k)$ rather than by the total variation of $f$. For the central limit theorems, the key mechanism is a degenerate contraction-method lemma (Proposition 1.3) that upgrades closeness to mixtures of Gaussians into a CLT once mean and variance scale linearly.

What would settle it

Compute $g'(x)$ numerically on a fine grid for $2<\gamma<3$, $0<x<1$; any nonpositive value would break the appendix inequality that the exponent proof relies on. Independently, iterate the recursion for fixed $k$ up to large $n$ and fit $\log(x_n - x)$ against $\log n$; the slope should approach $-\gamma_*\approx -1.56735$ if the claimed exponent is correct.

Watch

Extended reading notes

Core claim

The central theorem (Theorem 1.1) concerns the sequence $x_n$ defined by $x_n = \sum_{i=1}^{n-1} x_i/(h_{n-1}(n-i))$ with $x_1=\cdots=x_{k-1}=0$ and $x_k>0$, where $h_m=\sum_{j=1}^m 1/j$. The paper proves $x_n$ is strictly decreasing and its limit $x$ satisfies $x_n - x = n^{-\gamma_*+o(1)}$, with $\gamma_*$ the root of $\sum_{i=1}^\infty (1/i - i/((i+1)(i+1-\gamma)))=0$ lying between $1.5$ and $2$; numerically $\gamma_*=1.5673537531\ldots$. It also proves two-sided bounds for growing $k$, giving $c(k,\epsilon)(n/k)^{-\gamma_*-\epsilon} \le x_n-x \le C(k,\epsilon)(n/k)^{-\gamma_*+\epsilon}$ whenever $k=o(n/\log^3 n)$. Because the normalized expected clade count $\mathbb{E}[N_n(k)]/n$ satisfies the same recursion, this yields $a(n,k)-a(k) = n^{-\gamma_*+o(1)}$ for the occupation probability of the harmonic descent chain. The sharp rate is then fed into a degenerate contraction argument to prove central limit theorems for $N_n(k)$, for the number of copies of any fixed fringe subtree, and for the total length $\Lambda_n$ of the continuous-time tree, addressing Open Problems 5, 6, and 8 of the survey on critical $\beta$-splitting trees.

Load-bearing premise

The proof of the exponent depends on a calculus inequality, proved case by case in an appendix, that a certain derivative stays positive on $(0,1)$ for every exponent between $2$ and $3$; if that inequality failed, the sign in the induction step would reverse and the claimed exponent would not be established.

Editorial extensions

If this is right

  • For fixed clade size $k$, the occupation probability of the harmonic descent chain converges at rate $n^{-\gamma_*+o(1)}$, resolving the rate part of the open problems listed for this model.
  • The normalized number of clades of size $k$ in the critical beta-splitting tree converges in distribution to a standard Gaussian.
  • The same central limit theorem holds for the number of copies of any fixed fringe subtree.
  • The total length of the continuous-time critical beta-splitting tree satisfies a central limit theorem.
  • The monotonicity result extends to recursive averages under a ratio-monotonicity condition, covering beta-splitting variants with parameter $\beta<0$.

Reading between the lines

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

  • A natural next step beyond the paper is to test whether the same universal exponent controls the dominant error in other recursively defined statistics of the critical beta-splitting tree; the mechanism here suggests it would, because the exponent is fixed by the split kernel rather than by the statistic.
  • The two-sided bounds in Theorem 1.1(v) leave only a $o(1)$ correction in the exponent; a finer expansion, for instance a logarithmic correction such as $n^{-\gamma_*}(\log n)^{-\alpha}$, would require a second-order version of the induction, which the paper does not attempt.
  • Since the monotonicity criterion in Proposition 1.2 is stated abstractly, extending the sharp rate analysis to beta-splitting variants with other parameters is a plausible route if the positivity condition can be rechecked for those kernels.
  • For occupation probabilities with clade size $k$ growing like $n^\theta$, the paper gives two-sided bounds but not a sharp exponent; mapping that intermediate regime is an open continuation of this line of study.
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 / 4 minor

Summary. The manuscript studies the harmonic descent recursion x_n = sum_{i=1}^{n-1} x_i / (h_{n-1}(n-i)) with zero initial values and a single positive initial value x_k. It proves monotonicity of the sequence (Proposition 1.2, Theorem 1.1(i)), long- and short-term error bounds (Theorem 1.1(ii),(iii)), and a sharp asymptotic x_n - x = n^{-gamma_*+o(1)} for the convergence rate, where gamma_* is the implicit root of equation (1.2) (Theorem 1.1(iv),(v)). The proof rewrites the recursion in terms of consecutive differences d_n = x_n - x_{n+1}, derives exact coefficient expressions (2.6)-(2.7), approximates the resulting sums by integrals via an Euler-Maclaurin argument (Lemma 2.1), and then inductively propagates upper and lower polynomial bounds. The asymptotic is applied to the critical beta-splitting tree: it yields rates for the occupation probabilities a(n,k)-a(k), and, through a degenerate contraction-method proposition (Proposition 1.3), central limit theorems for the number of clades of fixed size or shape (Corollary 1.4) and for the total length of the continuous-time tree (Corollary 1.5). The paper thus answers Open Problems 5 and 6 and partially answers Open Problem 8 of Aldous and Janson.

Significance. If the main theorem is correct, the paper resolves the long-standing question of the convergence rate of the harmonic descent chain and gives the corresponding rate for the occupation probabilities of the critical beta-splitting tree. A notable strength is that the exponent gamma_* is not fitted: it is the parameter-free root of the implicit equation (1.2), so Theorem 1.1(iv) is a concrete, falsifiable prediction. The proofs are detailed and largely self-contained, with explicit recurrences (2.3) and (2.6), a careful Euler-Maclaurin approximation in Lemma 2.1, a calculus verification of the required sign of g' in Appendix A, and transparent moment/variance estimates in Lemmas 4.1, 4.2, 5.1, and 5.2. The CLT applications follow a clear framework and the paper is careful to isolate the new contraction-method ingredients. The main gap identified below is a missing verification of the sign and monotonicity of the series defining gamma_*; this is a short argument, but it is load-bearing for the inductive proof of the exponent.

major comments (1)
  1. [Section 2.4, after Lemma 2.1; Theorem 1.1(iv)] The induction for parts (iv)-(v) hinges on the inequalities (2.8) and (2.9). The text states that, for large n, these are equivalent to the series S(gamma) = sum_{i>=1} (1/i - i/((i+1)(i+2-gamma))) being positive for gamma < gamma_*+1 and negative for gamma > gamma_*+1. Lemma 2.1 only proves that the relevant quantity j_n equals S(gamma)+o(1); it does not establish the sign of S(gamma). Nor does the paper prove that equation (1.2) has a unique root gamma_* in (1.5,2); the numerical value quoted in Remark 2 is not a proof. These facts are load-bearing: if the sign of S(gamma) were different on either side of the claimed threshold, the inequalities in (2.8)-(2.9) would reverse and the exponent in Theorem 1.1(iv) would not follow. The missing argument is short and should be added, for example by writing S(gamma)=F(gamma-1) with F(beta)=sum_{i>=1}(1/i - i/((i+1)(i+1-beta))), noting F'(beta)<0 on (1,2), F(1)=1, and lim_{beta up arrow 2} F(beta) = -infinity.
minor comments (4)
  1. [Equation (2.4)] The displayed identity in (2.4) appears to have index and sign errors: as written, the left-hand side has numerator x-y after putting over a common denominator, while the right-hand side claims y-x, and the indices do not match the actual products p(n+1,x)p(n,y) - p(n,x)p(n+1,y). The conclusion that alpha_{i,n}<0 is correct, but the displayed equation should be corrected to avoid confusion.
  2. [Proof of Theorem 1.1(iii), Section 2.3] In the induction step, the display contains the typo 'x_k/(m+)h_k'; it should read x_k/((m+1)h_k).
  3. [Appendix A, final paragraph] The sentence 'Notice that for gamma > 2.5, we have m < 1' should read 'm > 1', since m=(gamma-2)/(3-gamma) exceeds 1 exactly when gamma>2.5. The subsequent reasoning is consistent with the corrected inequality, but the typo makes the argument difficult to follow.
  4. [Section 3, proof of Proposition 1.3] In the estimate for the term involving the integral, the numerical bound '0.01' appears without derivation; since it is used to choose constants, a parenthetical explanation of the numerical evaluation would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the exponent γ* is derived from the recurrence via integral approximation and verified by two-sided induction, not fitted or imported from the authors' own prior work.

full rationale

The central claim Theorem 1.1(iv) is not circular. The exponent γ* is defined by the implicit equation (1.2), which arises in the proof by comparing a divergent integral approximation to the recurrence for the differences d_n = x_n - x_{n+1}; the proof then establishes matching upper and lower bounds by induction using (2.8) and (2.9), so the rate n^{-γ_*+o(1)} is not assumed as an input. No parameter is fitted to data, and no prediction is used to set a constant. The convergence of the sequence to the explicit limit a(k) is imported from published external work [5, 7, 3, 4], and the contraction-method framework is taken from Neininger and Rueschendorf [10]; these are not self-citations and are used as independent published results. The only substantive gaps noted in the argument—the asserted sign of the series defining γ* in (1.2) and the positivity of g' used in Lemma 2.1—are unproved analytic claims, not reductions of the conclusion to its own inputs, and therefore are correctness risks rather than circularity.

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

No numbers are fitted to data or chosen by hand; the exponent γ* is defined implicitly by equation (1.2), and all constants in the CLT statements arise from the recursion or from prior proved limits. No new particles, forces, dimensions, or similar entities are postulated.

assumptions (4)
  • standard math Euler-Maclaurin formula as stated in equation (2.11), with the remainder bound |R_1| ≤ ∫|f'(x)|dx.
    Used in Lemma 2.1 to approximate sums by integrals; cited from [12].
  • standard math Zolotarev metric ζ_3 is a probability metric with ideal-metric properties: convergence implies weak convergence, ζ_3(cX,cY)=|c|^3 ζ_3(X,Y), and translation invariance.
    Used throughout Section 3 and Appendices B-C; cited from [13,14,11].
  • standard math Neininger-Rüschendorf contraction method for distributional recursions with degenerate limit equations.
    The paper adapts the framework of [10, Theorem 5.1]; the metric machinery is assumed as background.
  • domain assumption Known limit a(k) = 6 h_{k-1}/(π^2(k-1)) for the harmonic descent chain occupation probability.
    Used in Section 5 to evaluate the leading constant of E[Λ_n] and to state the applications; it is a prior result for this model, not re-derived here.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Asymptotics for the harmonic descent chain and applications to critical beta-splitting trees." pith.science (2026). https://pith.science/paper/ZILP2IAG

@misc{pith2026250524821,
  author       = {Pith},
  title        = {Pith review of: Asymptotics for the harmonic descent chain and applications to critical beta-splitting trees},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZILP2IAG}},
  note         = {Machine review of arXiv:2505.24821}
}
abstract

Motivated by the connection to a probabilistic model of phylogenetic trees introduced by Aldous, we study the recursive sequence governed by the rule $x_n = \sum_{i=1}^{n-1} \frac{1}{h_{n-1}(n-i)} x_i$ where $h_{n-1} = \sum_{j=1}^{n-1} 1/j$, known as the harmonic descent chain. While it is known that this sequence converges to an explicit limit $x$, not much is known about the rate of convergence. We first show that a class of recursive sequences including the above are decreasing and use this to bound the rate of convergence. Moreover, for the harmonic descent chain we prove the asymptotic $x_n - x = n^{-\gamma_* + o(1)}$ for an implicit exponent $\gamma_*$. As a consequence, we deduce central limit theorems for various statistics of the critical beta-splitting random tree. This answers a number of questions of Aldous, Janson, and Pittel.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

  1. [1]

    D. Aldous. Probability distributions on cladograms. In D. Aldous and R. Pemantle, editors, Random Discrete Structures, volume 76 of The IMA Volumes in Mathematics and its Applications, pages 1–18, New York, NY, 1996. Springer. 1, 2

    work page 1996

  2. [2]

    The Critical Beta-splitting Random Tree II: Overview and Open Problems

    D. Aldous and S. Janson. The critical beta-splitting random tree II: Overview and open problems. arXiv preprint arXiv:2303.02529v3, 2023. 1, 3, 4

    work page Pith review arXiv 2023

  3. [3]

    The Critical Beta-splitting Random Tree III: The exchangeable partition representation and the fringe tree

    D. Aldous and S. Janson. The critical beta-splitting random tree III: The exchangeable partition representation and the fringe tree. arXiv preprint arXiv:2412.09655, 2024. 2, 3

    work page Pith review arXiv 2024

  4. [4]

    Aldous and S

    D. Aldous and S. Janson. The critical beta-splitting random tree IV: Mellin analysis of leaf height. Electronic Journal of Probability, 30:1 – 39, 2025. 2, 3

    work page 2025

  5. [5]

    Aldous, S

    D. Aldous, S. Janson, and X. Li. The harmonic descent chain. Electronic Communications in Probability, 29:1–10, 2024. 2, 3 ASYMPTOTICS FOR HARMONIC DESCENT CHAIN AND APPLICATIONS 19

    work page 2024

  6. [6]

    Aldous and B

    D. Aldous and B. Pittel. The critical beta-splitting random tree I: Heights and related results. The Annals of Applied Probability, 35(1):158–195, 2025. 1, 2

    work page 2025

  7. [7]

    A. Iksanov. The harmonic descent chain and regenerative composition structures. Electronic Communi- cations in Probability, 30:1–3, 2025. 2, 3

    work page 2025

  8. [8]

    Kolesnik

    B. Kolesnik. Critical beta-splitting, via contraction. Electronic Communications in Probability, 30:1–14,

    work page

Show all 14 references
  1. [9]

    Neininger and L

    R. Neininger and L. R¨ uschendorf. A general limit theorem for recursive algorithms and combinatorial structures. The Annals of Applied Probability, 14(1):378–418, 2004. 4

    2004

  2. [10]

    Neininger and L

    R. Neininger and L. R¨ uschendorf. On the contraction method with degenerate limit equation. The Annals of Probability, 32(3B):2838 – 2856, 2004. 3, 4, 5, 13, 14, 22

    2004

  3. [11]

    S. T. Rachev. Probability Metrics and the Stability of Stochastic Models. Wiley Series in Probability and Mathematical Statistics: Applied Probability and Statistics. John Wiley & Sons, Ltd., Chichester, 1991. 13

    1991

  4. [12]

    E. T. Whittaker and G. N. Watson. A Course of Modern Analysis. Cambridge University Press, Cambridge, fifth edition, 2021. 11

    2021

  5. [13]

    Zolotarev

    V. Zolotarev. Approximation of distributions of sums of independent random variables with values in infinite-dimensional spaces. Theory of Probability & Its Applications, 21(4):721–737, 1977. 13

    1977

  6. [14]

    E " ζ3 r In n ZIn + r n − In n Zn−In + b(n), r In n τInN1 + r n − In n τn−InN2 + b(n) ! In ## ≤ E

    V. Zolotarev. Ideal metrics in the problem of approximating distributions of sums of independent random variables. Theory of Probability & Its Applications, 22(3):433–449, 1978. 13 Appendix A. Positivity of g′(x) in section 2.4 Recall that we have g′(x) = −(γ + 1)x1−γ + 1 + γx...

    1978

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.