pith. sign in

arxiv: 2606.11142 · v1 · pith:7D5RCKFAnew · submitted 2026-06-09 · 🧮 math.PR

Strong Approximations for Markov Chains Weakly Converging to Diffusions

V. Konakov , D. Kucher , E. Mammen This is my paper

Pith reviewed 2026-06-27 11:41 UTC · model grok-4.3

classification 🧮 math.PR
keywords strong approximationMarkov chainsdiffusionsweak convergencecouplingpathwise coincidencediscrete time grids
0
0 comments X

The pith

Markov chains weakly converging to diffusions can be coupled on one space to coincide exactly on time grids with maximal probability.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper constructs strong approximations by building versions of the discrete-time Markov chain and its limiting diffusion that live on the same probability space. Under bounded coefficients this coupling is arranged to maximize the probability that the two processes agree exactly at every point of a discrete time grid, and this holds whether the diffusion is non-degenerate or degenerate. The same construction also makes the probability that the linearly interpolated chain stays close to the continuous diffusion trajectory over the whole interval arbitrarily small. A reader would care because the result upgrades weak convergence to an explicit, pathwise control that applies directly to numerical simulation and error analysis.

Core claim

Under the assumption of bounded coefficients, we construct closely coupled versions of these processes on a shared probability space. In particular, for both non-degenerate and degenerate cases, we maximize the probability of their exact pathwise coincidence on discrete time grids. Moreover, we construct such probability space that the probability of a large deviation of the interpolated Markov chain from the continuous diffusion trajectory is small on the entire time interval.

What carries the argument

A joint coupling of the Markov chain and the diffusion on one probability space, constructed so that the probability of exact agreement on a discrete time grid is maximized.

If this is right

  • The same coupling works for perturbed versions of the Markov chain and diffusion.
  • Strong error bounds follow directly from the maximal coincidence probability on the grid.
  • The result covers both non-degenerate and degenerate limiting diffusions.
  • The probability of large deviation between the interpolated chain and the diffusion path can be made small uniformly over the time interval.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Numerical schemes that simulate the Markov chain can reuse the same random numbers as the diffusion to obtain pathwise comparisons without extra variance.
  • The construction supplies an explicit rate at which the grid-coincidence probability approaches one when the time step shrinks.
  • Extensions to unbounded coefficients would require localization arguments that keep the bounded-coefficient case as the core step.

Load-bearing premise

The coefficients of the Markov chain and the diffusion are bounded.

What would settle it

An explicit construction of a coupling on the same space that achieves a strictly higher probability of exact grid coincidence than the one given in the paper, while keeping the coefficients bounded.

read the original abstract

In this paper, we construct strong approximations for discrete-time Markov chains weakly converging to continuous diffusion processes, as well as for their perturbed counterparts. Under the assumption of bounded coefficients, we construct closely coupled versions of these processes on a shared probability space. In particular, for both non-degenerate and degenerate cases, we maximize the probability of their exact pathwise coincidence on discrete time grids. Moreover, we construct such probability space that the probability of a large deviation of the interpolated Markov chain from the continuous diffusion trajectory is small on the entire time interval.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The paper constructs strong approximations for discrete-time Markov chains that weakly converge to diffusions (and their perturbed versions). Under the assumption of bounded coefficients, it builds closely coupled versions of the processes on a common probability space. For both non-degenerate and degenerate cases, the construction maximizes the probability of exact pathwise coincidence on discrete time grids while ensuring that the probability of large deviations between the interpolated Markov chain and the diffusion trajectory remains small over the full continuous-time interval.

Significance. If the constructions are valid, the results strengthen standard weak-convergence statements by supplying explicit strong couplings. Such couplings are useful for obtaining pathwise error bounds, simulation schemes, and limit theorems that require joint realizations, and the treatment of both non-degenerate and degenerate cases broadens applicability in stochastic-process approximation theory.

minor comments (3)
  1. [Abstract] The abstract states that the constructions 'maximize' the coincidence probability; the main text should explicitly compare the achieved probability to the theoretical upper bound implied by the marginal laws.
  2. Notation for the discrete-time grid, the interpolation scheme, and the coupling measure should be introduced once in a dedicated notation subsection and used consistently thereafter.
  3. [Introduction] The manuscript would benefit from a short remark contrasting the present strong-coupling approach with existing Skorokhod-embedding or Komlós–Major–Tusnády-type results for diffusions.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the careful reading and positive assessment of the manuscript, including the accurate summary of the constructions for strong couplings in both non-degenerate and degenerate cases. The recommendation for minor revision is noted. No specific major comments were provided in the report.

Circularity Check

0 steps flagged

No significant circularity; direct constructive result

full rationale

The paper's central claim is an explicit construction of a joint probability space for coupling a discrete-time Markov chain to its limiting diffusion (under the stated bounded-coefficients assumption), maximizing exact coincidence on grids and controlling pathwise deviations. No equations, parameters, or uniqueness statements are shown to reduce to fitted inputs, self-citations, or ansatzes imported from prior work by the same authors. The boundedness hypothesis is used only to control moments and exit times and is not derived from the result itself. This is a standard coupling argument whose internal logic does not collapse to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the bounded-coefficients assumption stated in the abstract; no free parameters, additional axioms, or invented entities are identifiable from the given text.

axioms (1)
  • domain assumption Coefficients of the Markov chain and diffusion are bounded
    Explicitly invoked in the abstract as the setting under which the couplings are constructed.

pith-pipeline@v0.9.1-grok · 5615 in / 1081 out tokens · 25285 ms · 2026-06-27T11:41:53.724073+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

7 extracted references · 4 canonical work pages

  1. [1]

    Asymptotic version of the parametrix method for Markov chains converging to diffusions

    Ilya Bitter and Valentin Konakov. “Asymptotic version of the parametrix method for Markov chains converging to diffusions”. In:arXiv preprint arXiv:2505.24548(2025). eprint:2505.24548.url:https://arxiv.org/abs/2505.24548

    arXiv 2025

  2. [2]

    Stability estimates for singular SDEs and applications

    L. Galeati and C. Ling. “Stability estimates for singular SDEs and applications”. In: Electron. J. Probab.28.24 (2023), pp. 1–31

    2023

  3. [3]

    den Hollander.Probability Theory: The Coupling Method

    F. den Hollander.Probability Theory: The Coupling Method. Leiden: Leiden University, 2012, pp. 1–73

    2012

  4. [4]

    Stability of Densities for Perturbed Diffusions and Markov Chains

    Valentin Konakov, Anna Kozhina, and St´ ephane Menozzi. “Stability of Densities for Perturbed Diffusions and Markov Chains”. In:ESAIM: Probability and Statistics21 (2017), pp. 384–412.doi:10.1051/ps/2016028

    work page doi:10.1051/ps/2016028 2017

  5. [5]

    ParametrixMethodanditsApplicationsinProbabilityTheory

    AnnaKozhina.“ParametrixMethodanditsApplicationsinProbabilityTheory”.Disser- tation.Ruprecht-Karls-Universit¨ atHeidelberg,2018.doi:10.11588/heidok.00025129

    work page doi:10.11588/heidok.00025129 2018

  6. [6]

    Stability of Densities for Perturbed Degenerate Diffusions

    Anna Kozhina. “Stability of Densities for Perturbed Degenerate Diffusions”. In:Theory of Probability & Its Applications61.3(2017),pp.489–499.doi:10.1137/S0040585X97T988290

    work page doi:10.1137/s0040585x97t988290 2017

  7. [7]

    volume 293 of Grundlehren der mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences]

    DanielRevuzandMarcYor.Continuous Martingales and Brownian Motion.3rd.Vol.293. Grundlehren der mathematischen Wissenschaften. Springer-Verlag Berlin Heidelberg, 1999.doi:10.1007/978-3-662-06400-9. 17

    work page doi:10.1007/978-3-662-06400-9 1999