{"id":"5d6d6092-79d9-4d91-9ed4-aaa0f2a7414e","arxiv_id":"2412.04334","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A Lévy-subordinated fractional counting process is defined and its distributional properties, compound variants, Bell-polynomial connections, and a shock model are derived.","lead":"This paper runs a fractional counting process, a Poisson-like count with memory, at random times given by a Lévy subordinator, and writes formulas for its distribution, waiting times, compound sums, and products. The work is a technical extension aimed at shock and reliability models for aging structures.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (10) and Prop. 4.1 presuppose finite fractional moments E[H(t)^{θk}] for all k; for the α-stable subordinator used in Remark 5.1 these moments are infinite, so the central pmf/LT formulas fail for a main example.","rationale":"The reader's weakest_assumption identifies exactly the same load-bearing concern: the derivation assumes finite fractional moments of the Lévy subordinator and interchange of sums/integrals, while the paper nowhere states these conditions and later uses the α-stable subordinator, whose fractional moments are infinite for large orders. This concern directly attacks the strongest_claim, since Eq. (10), Eq. (12), and Theorem 4.1 are the central distributional results. The issue is concrete: for α < 1, E[S_α(t)^p] = ∞ for p ≥ α, and θ(n+k) exceeds α for all large k, so the displayed series contains infinite terms. The TCFCP may still be a well-defined process for stable subordinators, but the explicit formulas in the paper are not valid as written. The reader's CONDITIONAL verdict already reflects that the flaws are repairable by adding moment assumptions, replacing invalid discrete conditioning by integrals, and correcting the Υ function. Since my read does not shift that conclusion, I set verdict_should_be to UNCHANGED rather than proposing a different verdict.","tokens_in":22074,"tokens_out":6300,"duration_ms":67110,"concrete_test":"Take H to be an α-stable subordinator with α = 1/2, set θ = 1, and choose any admissible (μ, ϑ, ζ) so that n = 1, k = 0 term of Eq. (10) contains E[S(t)] = ∞. Compute z(1,t) numerically by quadrature of ∫ P_{μ,ϑ}^{ζ,1}(1,y) h(y,t)dy against the known stable density, and also evaluate truncated partial sums of Eq. (10). If the partial sums do not converge to the quadrature value, Eq. (10) is false as a convergent series; if some regularization is needed, the paper must state it explicitly.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is the explicit pmf z(n,t) in Eq. (10), the LT in Eq. (12), and the mean/variance in Theorem 4.1, all obtained by conditioning on H(t) and interchanging expectation with an infinite series. This requires E[(H(t))^{θ(n+k)}] < ∞ for every n,k appearing in Eq. (10), E[(H(t))^{mθ}] for every m in Eq. (12), and E[(H(t))^{2θ}] in Theorem 4.1(ii). No such condition is stated. For the α-stable subordinator S_α(t), α ∈ (0,1), E[S_α(t)^p] is finite only for p < α; the formula quoted in Remark 5.1, Γ(1−p/α)/Γ(1−p) t^{p/α}, is valid only in that range. Since θ ≤ 1 and α < 1, for every n ≥ 1 and all sufficiently large k one has θ(n+k) ≥ α, so the individual terms of Eq. (10) are infinite. The actual probability P(Z(t)=n) = ∫ P_{μ,ϑ}^{ζ,θ}(n,y) h(y,t)dy can still be finite because the FCP pmf decays rapidly in y, but it is not given by the displayed series, and Eq. (12) likewise contains infinite terms. Thus the central formulas are not valid for the stable subordinator highlighted later in the paper; they hold only under an unstated all-fractional-moments assumption, e.g. a gamma subordinator. This is a substantive gap, not a stylistic one.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies Laskin's generalized fractional counting process (FCP), derives additional properties of it, and then introduces a time-changed fractional counting process (TCFCP) Z(t)=N(H(t)) obtained by subordinating the FCP by an independent Lévy subordinator H. The central results are the claimed explicit probability mass function (Eq. 10), Laplace transform (Eq. 12), probability generating function, mean and variance (Theorem 4.1), as well as waiting-time and first-passage-time formulas, compound variants at Lévy times, subordinated generalized fractional Bell polynomials (SGFBP), and a shock deterioration model. The paper's main contribution is the set of distributional formulas for the TCFCP expressed in terms of fractional moments of the subordinator, together with the SGFBP connection and the application section.","tokens_in":22373,"tokens_out":8555,"duration_ms":84421,"significance":"If the central formulas were valid under a clearly stated set of assumptions, the paper would provide a useful unifying framework for fractional counting processes under Lévy subordination. The basic conditioning idea is natural, and the mean and variance expressions follow from standard conditional expectation once the needed moments exist. The introduction of SGFBP and the shock deterioration model with a bounded series function are interesting extensions. However, as written, the central formulas require unstated finiteness of all fractional moments of the subordinator, and a main example (the α-stable subordinator) violates that condition. In addition, several proofs condition on discrete values of a continuous subordinator and contain ill-posed definitions. These are load-bearing mathematical issues rather than presentation problems.","major_comments":[{"comment":"The derivation of the pmf and Laplace transform interchanges an infinite series with the expectation over H(t), producing terms E[(H(t))^{θ(n+k)}] and E[(H(t))^{mθ}]. No condition ensuring finiteness of these fractional moments is stated. For the α-stable subordinator S_α(t) used later in Remark 5.1, E[S_α(t)^p] is finite only for p<α, so for every fixed n and all sufficiently large k the expectation in Eq. (10) is infinite, and the same happens in Eq. (12) for large m. The integral representation z(n,t)=∫ P_{μ,ϑ}^{ζ,θ}(n,y) h(y,t)dy may still define a finite probability, but it is not equal to the displayed series. The paper must either impose an explicit all-fractional-moments assumption (for example, a gamma subordinator) throughout Section 4 and adjust the examples accordingly, or provide a regularized/truncated statement for subordinators with only finitely many fractional moments.","section":"§4.1, Eqs. (10) and (12)"},{"comment":"Both proofs condition on P[H(t)=n] and sum over n as though the Lévy subordinator H(t) were a discrete random variable. For the stable, gamma, and tempered stable subordinators considered in the paper, H(t) is absolutely continuous, so P[H(t)=n]=0 and the sums over n of P[H(t)=n] are not meaningful. The final expressions involving E[(H(t))^{θ(...)}] are formally what one would obtain by integrating against the density h(y,t)dy and interchanging sums, but the derivations as written are invalid and need to be rewritten with integrals and explicit justification of the interchange.","section":"§4.2, Proposition 4.3 and §4.3, Proposition 4.4"},{"comment":"The first passage time is defined as T_w := inf{t ≥ 0 : z(n,t) ≥ w}, but z(n,t) is the deterministic probability mass function of Z(t), not the process Z(t) itself. The definition should be T_w := inf{t ≥ 0 : Z(t) ≥ w}. The subsequent computation P[T_w > t] = Σ_{n=0}^{w-1} z(n,t) is the correct survival probability for that process definition, so the final formula is salvageable, but the statement and proof currently use the undefined expression P[z(n,t) < w] and sum P[z(n,t)] over n.","section":"§4.3, Theorem 4.3"},{"comment":"The displayed equality E{[Z(t,λ_θ)]^p} = B_{SG}(λ_θ(H(t))^θ,m) equates a deterministic moment on the left with a random variable on the right. What is actually true, and consistent with the definition of the SGFBP, is the conditional moment E[Z(t)^p | H(t)] = B_{SG}(λ_θ(H(t))^θ,m), or alternatively the unconditional moment E[Z(t)^p] = E[B_{SG}(λ_θ(H(t))^θ,m)]. In addition, Eq. (11) defines z(n,x) with no x appearing on the right-hand side; the notation must be fixed so that the argument of B_{SG} is well-defined.","section":"§5.3, moments of TCFCP"},{"comment":"The formula E[S_α(t)^p] = Γ(1−p/α)/Γ(1−p) t^{p/α} is cited for p>0, but for an α-stable subordinator this identity is valid only for 0<p<α; for p≥α the moment is infinite. Since the SGFBP series in Eq. (23) includes powers θ(n+k) for arbitrarily large n+k, the α-stable case is not a valid example for the displayed series. The same problem affects the incomplete gamma subordinator in Remark 5.1(ii), where the asymptotic is cited only for p≤α. These remarks should be corrected or replaced with subordinators that have finite fractional moments of all orders.","section":"Remark 5.1"},{"comment":"The proof that the TCFCP does not have independent increments uses the equality E[(H(t_1+t_2))^{mθ}] = E[(H(t_1))^{mθ}] E[(H(t_2))^{mθ}]. This is not a valid identity for Lévy subordinators: independent increments factor the Laplace transform of the sum, not the moments of the sum. The conclusion may be true, but the displayed argument is incorrect and needs to be replaced by a direct computation involving the joint Laplace transform of the increments.","section":"Remark 4.2"}],"minor_comments":[{"comment":"The title contains a typo: 'A t L ´evy times' should read 'At Lévy times'.","section":"Title"},{"comment":"Several references contain corrupted author-name artifacts such as 'Wy/suppress loma´ nska' (refs. [15], [21]); these should be cleaned up.","section":"References"},{"comment":"The function H(s,t) is called the mgf but is defined with e^{-sn}; either use e^{sn} or call it the Laplace transform / generating function, and align the notation with Eq. (13).","section":"§4.1, Proposition 4.2"},{"comment":"In the beta-product example, after taking c=d_1=...=d_m=1, the density f_{R_m}(x) is written with B(1,m) in the denominator; please check the normalization and whether it should be B(1,m) or another expression.","section":"§4.2, Example 4.2"},{"comment":"The density h_{Z_π}(y,t) in Eq. (17) has support restrictions (y≠0 or y≠1 depending on the atom at 1) that are not fully spelled out; please state the domain of the density explicitly.","section":"§4.2, Corollary 4.1(i)"},{"comment":"In Example 4.4, the Poisson convolution formula b_s^{*m} = e^{-mρ}(mρ)^s/s! is stated for s ∈ N_0; it would be clearer to indicate that this is the standard compound Poisson convolution.","section":"§4.3, Corollary 4.4"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a promising construction, but the main formulas are only valid under an unstated all-fractional-moments condition, and the paper's own α-stable example violates it. The conditioning errors in Propositions 4.3 and 4.4 and the ill-posed first-passage-time definition are fixable by rewriting the proofs with integrals and correcting the definitions. I would not reject the paper outright, because the integral representations and the gamma-subordinator case appear sound; however, the authors must either restrict the scope or provide corrected statements and proofs before the paper can be accepted. The novelty is moderate: the time-changed FCP is a natural extension of existing subordinated counting processes, and the SGFBP and shock-model sections are largely computational adaptations of prior work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does a standard thing: take Laskin's fractional counting process and time-change it with an independent Lévy subordinator. The pmf, Laplace transform, mean, and variance follow from conditioning and are correct if the subordinator has finite fractional moments of all orders. That is the key condition, and it is never stated. For a gamma subordinator, fine; for the α-stable subordinator used in Remark 5.1, E[H(t)^p] is infinite for p ≥ α, so the series in Eq. (10) and Eq. (12) have infinitely many infinite terms. The paper quotes the standard formula for E[S_α(t)^p] and applies it outside its range of validity. This is a substantive gap, not a stylistic one.\n\nThe paper does have genuine content: the compound variants and the Bell-polynomial connection are natural but not previously written down for this process. The derivations are mostly routine, and the novelty is incremental, but it is a reasonable subfield contribution.\n\nThere are also two repairable proof errors. Propositions 4.3 and 4.4 condition on H(t)=n with a discrete sum, but H(t) is continuous; the sums should be integrals. The shock model in Section 5.4 mixes λθ and λσ, and the Υ function definition does not match the argument in Eq. (29). These are fixable with careful rewriting.\n\nThe SGFBP moments representation in Section 5.3 is essentially definitional, so the 'connection' with Bell polynomials is less deep than advertised, but it is not wrong.\n\nOverall, the paper deserves a serious referee, but it should not be accepted as is. The authors need to state the moment assumptions, replace discrete conditioning with integrals, fix the shock model notation, and correct or qualify Remark 5.1. The central idea is sound for subordinators with all fractional moments; the stable case needs a different treatment.","headline":"Standard subordination results with a missing moment condition; the paper's own stable-subordinator example breaks its central formulas.","tokens_in":22994,"tokens_out":2501,"would_cite":false,"duration_ms":24831,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60G22","60G55","11B73","60K10"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper constructs the time-changed fractional counting process by subordinating the recently introduced fractional counting process with an independent Lévy subordinator, and derives explicit formulas for its probability mass function…","keywords":["fractional counting process","Lévy subordinator","time-changed process","Mittag-Leffler function","subordinated Bell polynomials","waiting time distribution","first passage time","shock deterioration model"],"falsifier":"Choose an $\\alpha$-stable subordinator with $\\alpha=1/2$ and parameters $\\theta=1$. For this subordinator $\\mathbb{E}[H(t)^p]$ is finite only for $p<\\alpha$, so the terms of the pmf series in Eq. (10) with $n+k\\ge1$ are undefined. A concrete check: compute the truncated series for $z(1,t)$ and see whether it converges, or verify whether $\\sum_{n=0}^N z(n,t)$ tends to 1 as $N\\to\\infty$; if the moments are infinite, the series cannot define a probability mass function. The paper's Remark 5.1 applies the stable moment formula for all orders, so evaluating that formula for $n+k$ above $\\alpha$ should produce a contradiction.","tokens_in":21792,"feed_emoji":"🔢","tokens_out":7334,"duration_ms":63734,"temperature":0.7,"pith_summary":"The paper introduces the time-changed fractional counting process (TCFCP), obtained by running the recently introduced fractional counting process (FCP) on a random clock given by an independent Lévy subordinator. It derives an explicit probability mass function for the TCFCP as a double series in fractional moments of the subordinator, along with its Laplace transform, probability generating function, moment generating function, mean, and variance. Waiting-time and first-passage-time distributions are computed, and multiplicative and additive compound versions of both the FCP and the TCFCP are analyzed. The paper also defines subordinated generalized fractional Bell polynomials and shows that the moments of the TCFCP are exactly these polynomials. A shock deterioration model for aging structures is proposed, in which the FCP supplies the arrival process and gamma-distributed increments model the damage per shock.","feed_headline":"Time-changed fractional counting process gets explicit formulas","feed_subtitle":"Running a fractional counting process on a Lévy clock yields closed-form pmf, moments, and Bell connections.","key_machinery":"The engine of the argument is the probability mass function of the base FCP, built from the generalized three-parameter Mittag-Leffler function $E^{\\zeta}_{\\mu,\\vartheta}(z)$. Subordinating that pmf by a Lévy subordinator replaces every power $y^{\\theta(n+k)}$ with the fractional moment $\\mathbb{E}[(H(t))^{\\theta(n+k)}]$, and the same substitution is applied to the generating functions, so every distributional identity for the base process lifts directly to the time-changed process. The subordinated generalized fractional Bell polynomials are defined from the same series and serve as the moment representation of the TCFCP.","core_discovery":"The central claim is that subordinating the FCP by an independent Lévy subordinator $H(t)$ produces a process $Z(t)=N^{\\zeta,\\theta}_{\\mu,\\vartheta}(H(t))$ whose law is completely described by the series $z(n,t)=\\frac{(\\zeta)_n\\Gamma(\\vartheta)\\lambda_\\theta^n}{n!}\\sum_{k=0}^\\infty\\frac{(-\\lambda_\\theta)^k(\\zeta+n)_k}{k!\\Gamma(\\mu(n+k)+\\vartheta)}\\,\\mathbb{E}[(H(t))^{\\theta(n+k)}]$, with the same fractional moments $\\mathbb{E}[(H(t))^{m\\theta}]$ entering the Laplace transform, pgf, mgf, mean, and variance. The paper further claims that integer moments of the TCFCP are represented by the subordinated generalized fractional Bell polynomials, and that the multiplicative and additive compound variants inherit the same series form after replacing the subordinator moments. These formulas are intended as a unified description of fractional counting processes under Lévy subordination.","pith_inferences":["The derivation assumes all fractional moments of the subordinator are finite; for $\\alpha$-stable subordinators the standard moment formula is only valid for orders below $\\alpha$, so the pmf series may fail to define a probability distribution unless the parameter range is restricted. This is an editorial caveat, not a claim of the paper.","A direct numerical check for a gamma subordinator (all moments finite) would validate the pmf formula by comparing the series with Monte Carlo simulation of $N(H(t))$; no such simulation is reported in the paper.","The same substitution mechanics could be applied to other count processes, such as negative binomial or generalized fractional variants, to produce time-changed families with explicit moments.","The shock model could be extended to the time-changed FGCP at Lévy times; the paper stops at the FCP-driven shock model."],"forward_implications":["For any Lévy subordinator with the required fractional moments, the full distribution of the TCFCP is available in closed form, so fitting or simulating the process reduces to estimating the subordinator's moments.","The mean and variance formulas give an explicit overdispersion structure, allowing the four FCP parameters plus the subordinator to match a wide range of count-data behavior.","Waiting-time and first-passage-time distributions follow from the same series, giving direct tools for reliability, ruin, and shock-model calculations.","When the subordinator is an $\\alpha$-stable process with $\\alpha=1$ and $t^\\theta$ is replaced by $x$, the subordinated Bell polynomials reduce to the ordinary generalized fractional Bell polynomials, connecting the new process to the prior combinatorial framework.","The shock deterioration model with gamma-distributed increments extends the compound Poisson shock model; setting $\\mu=\\vartheta=\\zeta=\\theta=1$ recovers that model exactly."],"supporting_citations":[{"why":"Defines the base fractional counting process, its pmf via the three-parameter Mittag-Leffler function, and the generalized fractional Bell polynomials that the paper later subordinates.","marker":"[26]"},{"why":"Supplies the definition and Laplace-transform representation of a Lévy subordinator used throughout the construction.","marker":"[2]"},{"why":"Provides the generalized three-parameter Mittag-Leffler function that carries the FCP pmf and its derivatives.","marker":"[38]"},{"why":"Gives the fractional moments of the α-stable subordinator used in Remark 5.1 to specialize the subordinated Bell polynomials.","marker":"[4]"},{"why":"Gives the asymptotic fractional moments of the incomplete-gamma subordinator used in another specialization.","marker":"[5]"},{"why":"Provides the asymptotic moments of tempered stable subordinators used in the third specialization of Remark 5.1.","marker":"[21]"},{"why":"Establishes the time-change construction of the fractional Poisson process by an inverse stable subordinator, the motivating example for subordination.","marker":"[33]"},{"why":"Introduces the compound Poisson shock deterioration model that the paper generalizes to the FCP in its application section.","marker":"[45]"}],"fun_headline_variants":["Lévy-time fractional counting gets closed-form laws","Subordinated fractional counting: explicit moments and Bell polynomials","Fractional counting on a Lévy clock yields explicit formulas","Time-changed fractional counting process solved via Bell polynomials","Lévy subordination gives explicit laws for fractional counting"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Every formula in the paper assumes the Lévy subordinator has finite fractional moments $\\mathbb{E}[H(t)^{\\theta(n+k)}]$ for all $n,k$ appearing in the infinite series, and that the series and integrals can be interchanged; the paper never states this, and for stable subordinators the usual moment formula only holds below the stability index.","fun_headline_variants_meta":{"raw":{"variants":["Lévy-time fractional counting gets closed-form laws","Subordinated fractional counting: explicit moments and Bell polynomials","Fractional counting on a Lévy clock yields explicit formulas","Time-changed fractional counting process solved via Bell polynomials","Lévy subordination gives explicit laws for fractional counting"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000202,"raw_usage":{"total_tokens":1403,"prompt_tokens":985,"completion_tokens":418,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":601,"completion_tokens_details":{"reasoning_tokens":340}},"tokens_in":601,"tokens_out":418,"duration_ms":4485,"temperature":1.0,"reasoning_tokens":340,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:32:48.400832+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Choose an $\\alpha$-stable subordinator with $\\alpha=1/2$ and parameters $\\theta=1$. For this subordinator $\\mathbb{E}[H(t)^p]$ is finite only for $p<\\alpha$, so the terms of the pmf series in Eq. (10) with $n+k\\ge1$ are undefined. A concrete check: compute the truncated series for $z(1,t)$ and see whether it converges, or verify whether $\\sum_{n=0}^N z(n,t)$ tends to 1 as $N\\to\\infty$; if the moments are infinite, the series cannot define a probability mass function. The paper's Remark 5.1 applies the stable moment formula for all orders, so evaluating that formula for $n+k$ above $\\alpha$ should produce a contradiction.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the base fractional counting process, its pmf via the three-parameter Mittag-Leffler function, and the generalized fractional Bell polynomials that the paper later subordinates."},{"cited_title":"Applebaum","cited_arxiv_id":null,"evidence_quote":"Supplies the definition and Laplace-transform representation of a Lévy subordinator used throughout the construction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the generalized three-parameter Mittag-Leffler function that carries the FCP pmf and its derivatives."},{"cited_title":"Beghin and J","cited_arxiv_id":null,"evidence_quote":"Gives the fractional moments of the α-stable subordinator used in Remark 5.1 to specialize the subordinated Bell polynomials."},{"cited_title":"Beghin and C","cited_arxiv_id":null,"evidence_quote":"Gives the asymptotic fractional moments of the incomplete-gamma subordinator used in another specialization."},{"cited_title":"Kumar, J","cited_arxiv_id":null,"evidence_quote":"Provides the asymptotic moments of tempered stable subordinators used in the third specialization of Remark 5.1."},{"cited_title":"Meerschaert, E","cited_arxiv_id":null,"evidence_quote":"Establishes the time-change construction of the fractional Poisson process by an inverse stable subordinator, the motivating example for subordination."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the compound Poisson shock deterioration model that the paper generalizes to the FCP in its application section."}],"review_version":1}