Paper: arXiv 2310.09181

Authors: Jim Gatheral, Radoš Radoičić

Abstract

Previously, in [“GR19”], we derived a rational approximation of the solution of the rough Heston fractional ODE in the special case λ= 0, which corresponds to a pure power-law kernel. In this paper we extend this solution to the general case of the Mittag-Leffler kernel with λ\geq 0. We provide numerical evidence of the convergence of the solution.

Complexity vs Empirical Score

  • Math Complexity: 9.5/10
  • Empirical Rigor: 3.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper is densely packed with advanced fractional calculus, asymptotic analysis, and complex integral equations, placing it firmly in high mathematical complexity. However, empirical evidence is limited to a few numerical convergence plots against a reference solver and lacks real-world data, backtests, or implementation details beyond theoretical approximations, indicating low empirical rigor.

Research Flowchart

  flowchart TD
  A["Research Goal: Generalize rational approximation of rough Heston solution"] --> B["Methodology: Derive analytical approximation for Mittag-Leffler kernel λ ≥ 0"]
  B --> C["Data/Inputs: Pure power-law kernel λ=0, Fractional ODE structure"]
  C --> D["Computational Process: Rational approximation scheme and numerical simulation"]
  D --> E["Key Finding: Validated convergence of the generalized solution"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper extends the rational approximation of the rough Heston model to the general case with a Mittag‑Leffler kernel (λ ≥ 0) and provides numerical evidence of convergence.

Source sentence
In this paper we extend this solution to the general case of the Mittag‑Leffler kernel with λ ≥0. We provide numerical evidence of the convergence of the solution.

Assumes 2

  • The kernel parameter λ is non‑negative (λ ≥ 0). model
    Source sentence
    In the case λ ≥0, the rough Heston model of [JR16] may be written in forward variance form
  • Realistic rough Heston parameters are H = 0.05, ν = 0.4, ρ = −0.65, λ = 1. model
    Source sentence
    In the following, we assume the following realistic rough Heston parameters: H = 0.05; ν = 0.4; ρ = −0.65; λ = 1.

Data

Synthetic data only.

Costs

Not applicable The paper reports no trading or execution result.

Finds 3

  • Rational approximations h(3,3), h(4,4) and h(5,5) are almost indistinguishable from the 1,000‑step Adams estimate and are significantly faster to compute than the classical Heston solution. theoretical
    Source sentence
    h(3,3), h(4,4) and h(5,5) are almost indistinguishable from the 1,000 step Adams estimate and significantly better than h(2,2). Moreover, all of these rational approximations are at least as fast to compute as the classical Heston solution.
  • Approximation quality decreases as H increases, but h(5,5) remains very accurate even in the classical case H = ½. theoretical
    Source sentence
    It turns out that the quality of the rational approximations decreases as H increases from 0 to 1/2, which corresponds to the classical Heston model
  • Numerical evidence shows exponential convergence of the diagonal approximants h(n,n) with respect to the order n. theoretical
    Source sentence
    we have provided numerical evidence of exponential convergence of the h(n,n) with respect to the approximation order n.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 13 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 1 paper in the archive

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.