Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Occupation times on the legs of a diffusion spider

T0 review · 3 major / 4 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 The paper establishes that the Laplace transform of the joint moments of occupation times on the legs of a homogeneous diffusion spider obeys an explicit recursion in terms of the spider's Green kernel, and that for Bessel…

desk verdict Useful and mostly careful extension of the authors' one-dimensional occupation-time results to spiders; the r=2 proofs are convincing, but Theorem 5's full claimed generality for r>=3 rests on 'same method can be repeated' rather than a displayed induction. read the letter →

arxiv 2411.09976 v1 pith:KXEHXAAY submitted 2024-11-15 math.PR

Paavo Salminen , David Stenlund This is my paper · ORCID
classification math.PR MSC 60J6060J5560J6505A10
keywords diffusionspiderWalshBrownianmotionoccupationtimesGreen'sfunctionKac'smomentformulaBesselprocessStirlingnumbersself-similar
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 aims to make the joint law of occupation times on the legs of a diffusion spider tractable. It proves a recursion that expresses the Laplace transform of any joint moment of the occupation times in terms of lower-order moments, using only the spider's Green kernel and the Laplace transforms of its hitting times of the origin. For a Bessel spider—where each leg has the radial law of a Bessel process with parameter $\nu\in(-1,0)$—the recursion is solved in closed form: every joint moment is a finite sum over Stirling numbers. The Brownian spider is the special case $\nu=-1/2$, and the same results cover sector occupation times for Walsh Brownian motion. If the paper is right, occupation-time moments on these multiray graphs are no longer an iterative or simulation-only problem.

What carries the argument

A diffusion spider is a star graph of $R$ legs carrying the same one-dimensional diffusion on each leg, with the process choosing leg $i$ with probability $\beta_i$ at the vertex. The machinery is the spider's Green kernel $g_\lambda$ (Theorem 1) together with the coefficients $D_k^{(i)}(\lambda)=\frac{\lambda^k}{(k-1)!}\int_0^\infty g_\lambda(0,(y,i))\,E_{(y,i)}(H_0^{k-1}e^{-\lambda H_0})\,\beta_i m(dy)$ from Eq. (13); these coefficients carry all the model data. The recursive step is the generalized Kac moment formula (Proposition 1), which breaks a product of occupation-time powers into a sum over legs and lower-order moments, with the integrals over the transition density collapsing onto the $D$ coefficients. For a self-similar spider, the scaling identity $A_t^{(i)}\stackrel{d}{=}t A_1^{(i)}$ replaces Laplace transforms by direct recursions for moments at time 1. In the Bessel case the $D$ coefficients become explicit, $D_k^{(i)}=-\beta_i\binom{\nu+k-1}{k}$, and the recursion solves to a finite Stirling-number sum.

What would settle it

Compute the two-leg Brownian spider moment $E_0(A_1^{(1)}A_1^{(2)})$ with $\beta_1=\beta_2=1/2$ by direct Monte Carlo simulation of Walsh Brownian motion; the explicit formula (Corollary 3) gives $1/8$, so a result that differs from $1/8$ by more than sampling error would refute the paper's closed-form theorem.

Watch

Extended reading notes

Core claim

The paper's central claim is Theorem 3: for a homogeneous diffusion spider with $R$ legs and occupation times $A_t^{(i)}$, the Laplace transform of any joint moment satisfies $$L_t\{E_0(\prod_{i=1}^r ($A_t^{{(i)}}$)^{n_i})\}(\$\lambda$)=\sum_{i=1}^r\sum_{k=1}^{n_i}\frac{n_i!\,$D_k^{{(i)}}$(\$\lambda$)}{(n_i-k)!\,\$\lambda$^k} L_t\{E_0(\prod_{j=1}^r ($A_t^{{(j)}}$)^{n_j}/($A_t^{{(i)}}$)^k)\}(\$\lambda$),$$ with $D_k^{(i)}(\lambda)$ defined by Eq. (13) through the Green kernel. For self-similar spiders this becomes the direct recursion (16) for $E_0(\prod_i (A_1^{(i)})^{n_i})$. The paper then evaluates the $D$ coefficients for Bessel spiders, obtaining $D_k^{(i)}=-\beta_i\binom{\nu+k-1}{k}$, and solves the recursion to arrive at Theorem 5: the explicit finite sum of Eq. (32) over Stirling numbers of the first and second kind. As a corollary, the joint first moment on $r$ legs is $E_0(A_1^{(1)}\cdots A_1^{(r)})=(-\nu)^{r-1}\beta_1\cdots\beta_r$, and the Brownian case $\nu=-1/2$ simplifies further to Eq. (34).

Load-bearing premise

The load-bearing premise is that the quoted formulas—the explicit Green kernel of the spider and the explicit Bessel coefficients $D_k^{(i)}$—are correct; the paper uses them without reproving them, so an error in either would invalidate every recursion and closed-form moment that follows.

Editorial extensions

If this is right

  • Joint occupation moments on the legs of any homogeneous diffusion spider become recursively computable once the Green kernel and the Laplace transforms of first-hitting times of 0 are known.
  • For Bessel spiders the recursion closes into a finite Stirling-number sum, so moments at all orders are available in closed form rather than by iterative integration.
  • For Brownian spiders the closed form simplifies via Bessel numbers and gives concrete formulas for Walsh Brownian motion on finitely many rays.
  • The moment generating function at an independent exponential time has a rational closed form, linking the moment recursion to the known double-Laplace-transform description of the joint law.
  • For Walsh Brownian motion with a continuous angular distribution, sector occupation times follow the same formulas with $\beta_i$ equal to sector probabilities.

Reading between the lines

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

  • The paper leaves implicit that Theorem 3 is an algorithm: for any homogeneous spider with an explicit Green kernel, the $D_k^{(i)}$ integrals can be evaluated once, and the recursion then yields moments of arbitrary order without further simulation.
  • A natural testable extension is the inhomogeneous spider, where each leg has its own speed measure; the Kac-moment structure is leg-agnostic, so the recursion likely survives with leg-dependent coefficients, while the closed Bessel form would not.
  • The exponential-time moment generating function (21) could be inverted numerically in $\lambda$ to obtain densities of the joint occupation law, connecting the moment formulas to the double-Laplace-transform characterization of earlier work.
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

3 major / 4 minor

Summary. The paper studies the joint moments of occupation times on the legs of a homogeneous diffusion spider. It presents four main results: Theorem 3, a recursive formula for the Laplace transform of the joint occupation-time moments for any number r of legs; Theorem 4, a formula for the moment generating function of the occupation times at an independent exponential time; Theorem 5, an explicit Stirling-number closed form for the joint moments in the Bessel spider case; and Theorem 6, a specialized closed form for the Brownian spider. The proofs of Theorems 3 and 5 are written out only for r = 2, with the extension to larger r asserted by repetition of the same method; Theorem 4 is proved for general R. The paper also connects the results to Yano's double Laplace transform formula and to Walsh Brownian motion on sectors.

Significance. If the stated formulas are correct, the paper offers a useful extension of the single-leg occupation-time results in [20] to joint occupation times on spider legs, and the explicit formula (32) for Bessel spiders is elegant and likely to be of interest to researchers working on diffusions on graphs. The proof strategy is sound for the r = 2 case: the recursion is derived from Kac's moment formula and the explicit Green kernel (7), without circularity, and Theorem 4 is proved for general R in a clean manner. The special cases r = 1 and r = 2, and the Brownian specialization in Theorem 6, check out. However, the claimed generality of the central results is not currently supported: both Theorem 3 and Theorem 5 are proved only for two legs, and the text explicitly leaves the generalization to the reader. Since the r = 2 proof of Theorem 5 relies on nontrivial combinatorial identities involving Stirling numbers, the status of (32) for r ≥ 3 is an open question within the manuscript; similarly, the dependence of (16) on the r = 2 proof means the foundation for higher r is not fully documented.

major comments (3)
  1. [Appendix A / Theorem 5, Eq. (32)] Theorem 5 is stated for any r ∈ {1,...,R}, but the proof in Appendix A is carried out only for r = 2, ending with the statement that the procedure 'can then readily be repeated for a larger value of r'. This is a load-bearing gap: the r = 2 induction collapses the sums using two nontrivial combinatorial identities involving products of Stirling numbers of the first and second kind, and for r ≥ 3 the analogous expressions would contain products of three or more such Stirling factors. No displayed identity or argument shows that these higher-r expressions collapse to the claimed form (32). The formula for r ≥ 3 is therefore currently an unproved assertion. Please either provide a full proof for general r, or state Theorem 5 for r = 2 (with a clearly labeled conjecture for r ≥ 3).
  2. [Section 4.2 / Theorem 3] The proof of Theorem 3 stops at r = 2 and says 'Proving the result for r = 2 should be sufficient, since the method is the same for any higher values of r'. The structure of (15) is additive, so the extension to r > 2 is plausible, but the proof is not written out, and Remark 3 points out that the single-leg case r = 1 is not a special case of the theorem. Because Theorem 5's recursive derivation uses (16), the generality of Theorem 3 has direct consequences for the main closed-form result. I ask that a complete induction argument for general r be included, or that the theorem be stated with its proof for r = 2 and the extension as a remark.
  3. [Section 6.1, Eq. (30)] The formula (30) for D_k^{(i)}(λ) is imported from the authors' previous paper [20] with the note that a different normalization of m and S is used there, and Remark 2.1 additionally flags a sign change relative to [20]. Since Eq. (30) is an essential input to Theorem 5, and since a normalization or sign mismatch in this imported result would propagate into the main formula, the manuscript should include a self-contained derivation of (30) under the present normalization, or at least a detailed and explicit statement of the conversion between the normalizations. As it stands, a reader cannot verify the correctness of (30) without consulting [20] and reconstructing the change of variables.
minor comments (4)
  1. [Section 3, proof of Proposition 1] In the sentence 'Restating the equation above with N = 2', the displayed formula contains a missing parenthesis: 'Ex (At(V1))n1At(V2))n2 )' should be 'E_x((A_t(V_1))^{n_1}(A_t(V_2))^{n_2})'.
  2. [Section 5, before Corollary 2] The phrase 'see also Theorem 4 in Barlow, Pitman and Yor [3], where where the formula is presented' contains a duplicated 'where'.
  3. [Corollary 2 and Eq. (27)] The notation switches between R (number of legs of the spider) and r (number of legs with positive z_i) without explicit definition of how legs with z_i = 0 are handled; a brief clarification would avoid ambiguity, especially since (27) is then claimed for any r ≤ R.
  4. [Remark 2.2] The remark states that D_k^{(i)}(λ) cannot depend on λ for self-similar spiders, and this is used repeatedly; making the short induction explicit in the remark would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the moment recursions are derived from Kac's formula and an independently stated Green kernel; the Bessel closed form is an induction from prior single-leg results. The r≥3 extension is asserted rather than proved, but that is a completeness issue, not circularity.

full rationale

The derivation chain is not circular. Theorem 3's recursion (15) follows from Proposition 1 (Kac's moment formula, proved in the paper), a strong-Markov decomposition before the first hitting time of 0, and the explicit Green kernel quoted from [17] (Theorem 1). No target moment formula is fed into the proof; the inputs are occupation-time indicators and the resolvent density. The self-similar version (16) is obtained by taking Laplace transforms and using scaling, not by assuming the moments being computed. Theorem 5 is proved by induction from recurrence (16), the single-leg formula (31), and the Bessel coefficients D_k^(i) in (30), both taken from the authors' earlier paper [20]. These are published, parameter-free results about one-dimensional (skew) Bessel occupation times, not restatements of the joint-moment formula being proved; under the review rules they count as independent evidence and do not make the argument circular. The paper itself flags the main caveat: the proof of Theorem 3 only writes out r=2 and says "the method is the same for any higher values of r", and Appendix A proves Theorem 5 only for r=2, saying the procedure "can then readily be repeated for a larger value of r". Thus the full r≥3 claims are asserted rather than demonstrated; this is a completeness/rigor risk, not a case of a prediction reducing to its inputs. No fitted parameters are relabelled as predictions, and no uniqueness theorem is invoked to force a choice.

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

The main formulas rest on the Green kernel of the spider and on prior Bessel-spider coefficient formulas; these are treated as inputs. No parameters are fitted to data, and no new entities are introduced.

assumptions (5)
  • domain assumption The diffusion spider has a Green kernel density gλ given by the explicit formula (7), quoted as Theorem 1 from Lempa-Mordecki-Salminen [17].
    Theorem 1 is cited, not proved; it underpins the definition of D_k in (13) and all subsequent recursions.
  • domain assumption For a Bessel spider, D_k^{(i)}(λ) = -β_i binom(ν+k-1,k), Eq. (30), and the single-leg moment formula (31), are taken from Salminen-Stenlund [20].
    The explicit joint-moment formula Theorem 5 is proved by inserting these prior expressions into the recurrence; they are not rederived.
  • standard math The Bessel process speed measure and eigenfunctions, including m(dx)=2x^{2ν+1}dx, φλ(x)=x^{-ν}Kν(x√(2λ)), and the cλ expression in Eq. (29), are taken from Borodin-Salminen [6].
    Used to compute gλ(0,(y,i)) and D_k in Section 6.1.
  • domain assumption Self-similarity of the spider implies A_t^{(i)} has the same law as t A_1^{(i)}, which is used to convert Laplace transforms into moments in (16) and (20).
    This scaling is stated, not proved, in Section 4 and used for Bessel and Brownian spiders.
  • standard math Combinatorial identities for Stirling numbers (Equations (6.15) in [13], Lemma 2 in [20]) used in Appendix A are taken as known.
    The proof of Theorem 5 relies on these identities without proof.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Occupation times on the legs of a diffusion spider." pith.science (2026). https://pith.science/paper/KXEHXAAY

@misc{pith2026241109976,
  author       = {Pith},
  title        = {Pith review of: Occupation times on the legs of a diffusion spider},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KXEHXAAY}},
  note         = {Machine review of arXiv:2411.09976}
}
read the original abstract

We study the joint moments of occupation times on the legs of a diffusion spider. Specifically, we give a recursive formula for the Laplace transform of the joint moments, which extends earlier results for a one-dimensional diffusion. For a Bessel spider, of which the Brownian spider is a special case, our approach yields an explicit formula for the joint moments of the occupation times.

Figures

Figures reproduced from arXiv: 2411.09976 by the authors.

Figure 1
Figure 1. The graph of a diffusion spider with five legs. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. General diffusions on metric graphs as limits of time-space Markov Chains

    math.PR 2025-07 conditional novelty 7.0 of 10

    A new space-time Markov chain approximation for general diffusions on metric graphs is shown to converge in p-Wasserstein distance at explicit rates governed by a thinness quantifier of the subdivision.

Reference graph

Works this paper leans on

26 extracted references · 26 canonical work pages · cited by 1 Pith paper

  1. [20]

    and Stenlund, D

    Salminen, P. and Stenlund, D. (2021). On occupation tim es of one-dimensional diffusions. J. Theoret. Probab. 34(2), 975–1011

    work page 2021

  2. [1]

    and Cohen, A

    Atar, R. and Cohen, A. (2019). Serve the shortest queue an d Walsh Brownian motion. Ann. Appl. Probab. 29(1), 613 – 651

    work page 2019

  3. [2]

    Barlow, M., Pitman, J., and Yor, M. (1989). On Walsh’s Bro wnian motions. In: Séminaire de Proba- bilités, XXIII , vol. 1372 of Lecture Notes in Math. , pp. 275–293. Springer, Berlin

    work page 1989

  4. [3]

    Barlow, M., Pitman, J., and Yor, M. (1989). Une extension multidimensionnelle de la loi de l’arc sinus. In: Séminaire de Probabilités, XXIII , vol. 1372 of Lecture Notes in Math. , pp. 294–314. Springer, Berlin

    work page 1989

  5. [4]

    and Zhang, X

    Bayraktar, E. and Zhang, X. (2021). Embedding of walsh br ownian motion. Stochastic Process. Appl. 134, 1–28

    work page 2021

  6. [5]

    A., and Osękowski, A

    Bednarz, E., Ernst, P. A., and Osękowski, A. (2024). On th e diameter of the stopped spider process. Math. Oper. Res. 49(1), 346–365

    work page 2024

  7. [6]

    Borodin, A. N. and Salminen, P. (2015). Handbook of Brownian motion—facts and formulae , corr. reprint, second edn. Birkhäuser Verlag, Basel

    work page 2015

  8. [7]

    Csáki, E., Csörgő, M., Földes, A., and Révész, P. (2016). Some limit theorems for heights of random walks on a spider. J. Theoret. Probab. 29(4), 1685–1709

    work page 2016

Show all 26 references
  1. [8]

    Csáki, E., Csörgő, M., Földes, A., and Révész, P. (2019). Limit theorems for local and occupation times of random walks and Brownian motion on a spider. J. Theoret. Probab. 32(1), 330–352

    2019

  2. [9]

    and Földes, A

    Csáki, E. and Földes, A. (2024). In memoriam pál révész (1 934–2022). Period. Math. Hungar. 89(1), 201–229

    2024

  3. [10]

    Ernst, P. (2016). Exercising control when confronted b y a (Brownian) spider. Oper. Res. Lett. 44(4), 487–490

    2016

  4. [11]

    Fitzsimmons, P. J. and Kuter, K. E. (2014). Harmonic fun ctions on Walsh’s Brownian motion. Stochastic Process. Appl. 124(6), 2228–2248

    2014

  5. [12]

    Freidlin, M. I. and Wentzell, A. D. (1993). Diffusion Pro cesses on Graphs and the A veraging Principle. Ann. Probab. 21(4), 2215 – 2245

    1993

  6. [13]

    L., Knuth, D

    Graham, R. L., Knuth, D. E., and Patashnik, O. (1994). Concrete mathematics , second edn. Addison- Wesley Publishing Company, Reading, MA

    1994

  7. [14]

    and McKean, H

    Itô, K. and McKean, H. P., Jr. (1974). Diffusion processes and their sample paths . Springer-Verlag, Berlin

    1974

  8. [15]

    Kac, M. (1951). On some connections between probabilit y theory and differential and integral equations. In: Proceedings of the Second Berkeley Symposium on Mathematic al Statistics and Probability, 1950 , pp. 189–215. University of California Press, Berkeley and Los A ngeles

    1951

  9. [16]

    and Yan, M

    Karatzas, I. and Yan, M. (2019). Semimartingales on ray s, walsh diffusions, and related problems of control and stopping. Stochastic Process. Appl. 129(6), 1921–1963

    2019

  10. [17]

    Lempa, J., Mordecki, E., and Salminen, P. (2024). Diffus ion spiders: Green kernel, excessive functions and optimal stopping. Stochastic Process. Appl. 167, 104229

    2024

  11. [18]

    Papanicolaou, V., Papageorgiou, E., and Lepipas, D. (2 012). Random motion on simple graphs. Methodol. Comput. Appl. Probab. 14, 285–297. 20

  12. [19]

    Salisbury, T. S. (1986). Construction of right process es from excursions. Probab. Theory Related Fields 73(3), 351–367

    1986

  13. [21]

    Stenlund, D. (2022). On the connection between Stirlin g numbers and Bessel numbers. Electron. J. Combin. 29(1), P1.40, 14

    2022

  14. [22]

    and Yor, M

    Vakeroudis, S. and Yor, M. (2012). A scaling proof for Wa lsh’s Brownian motion extended arc-sine law. Electron. Commun. Probab. 17, no. 63, 9

    2012

  15. [23]

    Walsh, J. B. (1978). A diffusion with a discontinuous loc al time. In: Temps locaux, vol. 52–53 of Astérisque, pp. 37–45. Société mathématique de France

    1978

  16. [24]

    Weber, M. (2001). On occupation time functionals for di ffusion processes and birth-and-death processes on graphs. Ann. Appl. Probab. 11(2), 544 – 567

    2001

  17. [25]

    Yang, S. L. and Qiao, Z. K. (2011). The bessel numbers and bessel matrices. J. Math. Res. Exposition 31(4), 627–636

    2011

  18. [26]

    Yano, Y. (2017). On the joint law of the occupation times for a diffusion process on multiray. J. Theoret. Probab. 30(2), 490–509. 21

    2017

Pith tools

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