Paper: arXiv 2511.19826

Authors: Yun-Feng Tu, Chuan-Hsiang Han

Abstract

This paper investigates asymptotically optimal importance sampling (IS) schemes for pricing European call options under the Heston stochastic volatility model. We focus on two distinct rare-event regimes where standard Monte Carlo methods suffer from significant variance deterioration: the limit as maturity approaches zero and the limit as the strike price tends to infinity. Leveraging the large deviation principle (LDP), we design a state-dependent change of measure derived from the asymptotic behavior of the log-price cumulant generating functions. In the short-maturity regime, we rigorously prove that our proposed IS drift, inspired by the variational characterization of the rate function, achieves logarithmic efficiency (asymptotic optimality) by minimizing the decay rate of the second moment of the estimator. In the deep OTM regime, we introduce a novel slow mean-reversion scaling for the variance process, where the mean-reversion speed scales as the inverse square of the small-noise parameter (defined as the reciprocal of the log-moneyness). We establish that under this specific scaling, the variance process contributes non-trivially to the large deviation rate function, requiring a specialized Riccati analysis to verify optimality. Numerical experiments demonstrate that the proposed method yields substantial variance reduction–characterized by factors exceeding several orders of magnitude–compared to standard estimators in both asymptotic regimes.

Complexity vs Empirical Score

  • Math Complexity: 9.0/10
  • Empirical Rigor: 3.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper is mathematically dense, featuring advanced concepts like large deviation principles, Riccati analysis, and asymptotic proofs, but its empirical implementation is limited to theoretical numerical experiments without actual backtest-ready data or code, placing it in the ‘Lab Rats’ quadrant.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Efficient Importance Sampling for Heston Model<br>Short Maturity & Deep OTM Regimes"] --> B["Methodology: Large Deviation Principle LDP<br>Analyze asymptotic behavior of log-price C.G.F."]
  
  B --> C{"Two Rare-Event Regimes"}
  
  C --> D["Regime 1: Short Maturity<br>State-dependent change of measure<br>Minimize 2nd moment decay rate"]
  C --> E["Regime 2: Deep OTM<br>Slow mean-reversion scaling<br>Specialized Riccati analysis"]
  
  D --> F["Computational Process<br>Monte Carlo Pricing<br>Variance Reduction Analysis"]
  E --> F
  
  F --> G["Outcomes<br>Logarithmic Efficiency achieved<br>Proven asymptotic optimality<br>Numerical variance reduction >10^3"]

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 proposes a state‑dependent importance sampling scheme for the Heston model, using a drift change that is affine in the square‑root of the variance.

Source sentence
In this paper, we propose a state-dependent importance sampling scheme for the Heston model constructed via a change of drift that is affine in the square root of the variance.

Assumes 5

  • Zero risk‑free rate (r = 0) is assumed. model
    Source sentence
    Assuming a zero risk‑free rate (r = 0) without loss of generality, the dynamics are governed by: ...
  • Feller condition 2κθ ≥ σ² holds to ensure strict positivity of Vₜ. model
    Source sentence
    We assume the Feller condition 2κθ ≥σ2 holds to ensure strict positivity of Vt.
  • Correlation ρ lies in (‑1, 1). model
    Source sentence
    The correlation between the asset price and its variance is captured by ρ ∈(−1, 1), and we define ¯ρ :=
  • Short‑maturity regime considers T → 0. other
    Source sentence
    Short‑Maturity Regime (T →0): We leverage the small‑time LDP results...
Show 1 more
  • Deep out‑of‑the‑money regime uses K → ∞ and defines ε = 1/ log(K/S₀). other
    Source sentence
    We introduce a small‑noise parameter ε = 1/ log(K/S0) and propose a slow mean‑reversion scaling where the speed of mean reversion scales as δ = ε−2.

Data

  • Synthetic Monte Carlo experiments with Heston parameters synthetic unknown
    Source sentence
    S0 = 2000, v0 = θ = 0.5, κ = 15, σ = 1, ρ = −0.1, r = 0. We vary the strike price K to span a range of moneyness K/S0 ∈[1.0, 2.0] and measure the VRR across three maturities...

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • The proposed IS scheme achieves logarithmic efficiency (asymptotic optimality) in the short‑maturity regime. theoretical
    Source sentence
    ... achieves logarithmic efficiency (asymptotic optimality) by minimizing the decay rate of the second moment of the estimator.
  • In the deep OTM regime with slow mean‑reversion scaling, the scheme is asymptotically optimal and captures non‑trivial contributions of the variance process. theoretical
    Source sentence
    we establish that under this specific scaling, the variance process contributes non‑trivially to the large deviation rate function, requiring a specialized Riccati analysis to verify optimality.
  • Variance reduction ratios (VRR) exceed 2500 for deep OTM strikes in the 1‑day maturity case. performance
    Source sentence
    VRR grows exponentially with moneyness, exceeding 2500 for deep OTM strikes.
  • VRR remains around 450 for a 1‑month maturity and between 100‑250 for a 1‑year maturity. performance
    Source sentence
    ... VRR saturates below 250, indicating the diminishing effectiveness of the specific drift as the time horizon allows for more complex path dynamics.
  • Overall numerical experiments show variance reduction spanning two to three orders of magnitude compared to standard Monte Carlo. performance
    Source sentence
    Numerical experiments confirmed our theoretical findings, showing substantial variance reduction ratios—spanning from two to three orders of magnitude—particularly in the regimes where standard estimators fail most severely.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 50 related papers). Coverage is partial, so this is neither replication nor refutation.

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.

Cited References (13)

  1. A closed-form solution for options with stochastic volatility with applications to bond and currency options (1993)
  2. The Volatility Surface: A Practitioner’s Guide (2006)
  3. Derivatives in Financial Markets with Stochastic Volatility (2000)
  4. Asymptotically optimal importance sampling and stratification for pricing path-dependent options (1999)
  5. Optimal importance sampling with explicit formulas in continuous time (2008)
  6. Sample path large deviations and optimal importance sampling for stochastic volatility models (2010)
  7. Small-time asymptotics for implied volatility under the Heston model (2011)
  8. Smile asymptotics II : Models with known moment generating functions (2009)
and 5 more
  1. Short-maturity asymptotics for a fast mean-reverting Heston stochastic volatility model (2012)
  2. The moment formula for implied volatility at extreme strikes (2004)
  3. Asymptotic formulas for implied volatility in the Heston model (2010)
  4. Large deviations in finance (2015)
  5. Large Deviations Techniques and Applications (1998)

0 of 13 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.