Paper: arXiv 2505.22836

Authors: Pierre Brugière, Gabriel Turinici

Abstract

Option pricing theory, such as the Black and Scholes (1973) model, provides an explicit solution to construct a strategy that perfectly hedges an option in a continuous-time setting. In practice, however, trading occurs in discrete time and often involves transaction costs, making the direct application of continuous-time solutions potentially suboptimal. Previous studies, such as those by Buehler et al. (2018), Buehler et al. (2019) and Cao et al. (2019), have shown that deep learning or reinforcement learning can be used to derive better hedging strategies than those based on continuous-time models. However, these approaches typically rely on a large number of trajectories (of the order of \(10^5\) or \(10^6\)) to train the model. In this work, we show that using as few as 256 trajectories is sufficient to train a neural network that significantly outperforms, in the Geometric Brownian Motion framework, both the classical Black & Scholes formula and the Leland model, which is arguably one of the most effective explicit alternatives for incorporating transaction costs. The ability to train neural networks with such a small number of trajectories suggests the potential for more practical and simple implementation on real-time financial series.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 7.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced stochastic calculus and deep learning optimization with dense mathematical formulations, while demonstrating practical performance with limited data trajectories suitable for real-time implementation.

Research Flowchart

  flowchart TD
  Start["Research Goal: Develop a deep hedging model<br>with minimal data and transaction costs"] --> Methodology["Methodology: Model-Free Deep Reinforcement Learning"]
  
  Inputs["Data Inputs: 256 Trajectories<br>Geometric Brownian Motion (GBM)"] --> Methodology
  
  Methodology --> Process["Computational Process:<br>Neural Network Optimization<br>Minimize Risk + Transaction Costs"]
  
  Process --> Compare["Benchmark Comparison:<br>Black & Scholes vs Leland Model"]
  
  Compare --> Outcomes["Key Findings:<br>1. Significantly outperforms B&S & Leland<br>2. Efficient with only 256 trajectories<br>3. Practical for real-time financial data"]
  
  style Start fill:#e1f5fe
  style Outcomes fill:#e8f5e8

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 training a neural network on as few as 256 simulated trajectories to obtain a hedging strategy that outperforms the Black‑Scholes delta hedge and the Leland model under transaction costs.

Source sentence
In this work, we show that using as few as 256 trajectories is sufficient to train a neural network that significantly outperforms, in the Geometric Brownian Motion framework, both the classical Black & Scholes formula and the Leland model

Assumes 6

  • The underlying asset follows a geometric Brownian motion (GBM) process. model
    Source sentence
    In the first case, the asset follows a geometric Brownian motion (GBM) process, and the instantaneous interest rate rs is stochastic.
  • Transaction costs are proportional to the dollar amount traded. market
    Source sentence
    We assume that the trading costs are proportional to the value traded.
  • Hedging is performed at discrete time points with no costs at inception or maturity. execution
    Source sentence
    Trading costs are assumed to apply, neither at inception, when the trader exchanges the delta with the option buyer, nor at maturity for the stocks which are delivered against the strike
  • The risk‑free rate is taken as constant (4.5% in the S&P 500 experiment). market
    Source sentence
    We assume a constant instantaneous interest rate of 4.5%, which is the average Secured Overnight
Show 2 more
  • Neural network is trained on data simulated under the real‑world probability measure (drift not zero). data
    Source sentence
    In our experiments, it was not necessary to simulate trajectories under a risk‑neutral probability; instead, we simulated under a probability measure where the drift does not ne…
  • The hedging objective is to minimize the standard deviation of discounted terminal wealth. statistical
    Source sentence
    we choose to minimize the standard deviation of discounted terminal wealth ZT

Data

  • Simulated GBM trajectories synthetic unknown
    Source sentence
    In our sample of 256 trajectories simulated with σ = 20%, the estimated values of σ computed with 30 observation points vary between 12.51% and 26.89%.
  • Overlapping sequences from a single simulated series synthetic unknown
    Source sentence
    a single sequence of length 285 is simulated, and 256 overlapping sequences of length 30 derived from it are used to train the neural network for hedging
  • S&P 500 Total Return Index · S&P Dow Jones Indices market-prices public
    01/01/2024 to 25/02/2025 · S&P 500 Total Return Index
    Source sentence
    Here, a single sequence of length 286 is constructed from the closing prices of the S&P 500 Total Return Index over the period from 01/01/2024 to 25/02/2025.

Costs

Modeled Costs are part of the reported headline results.

Proportional cost α per dollar traded, applied only when deltas are readjusted at discrete times.

Source sentence
the transaction costs of α per dollar traded arises only when the deltas are readjusted at times ti+1

Finds 5

  • A neural network trained on 256 GBM trajectories outperforms both the Black‑Scholes delta hedge and the Leland model under proportional transaction costs. performance
    Source sentence
    using as few as 256 trajectories is sufficient to train a neural network that significantly outperforms, in the Geometric Brownian Motion framework, both the classical Black & Scholes formula and the Leland model
  • The network further improves on the Leland approach for transaction costs at 0.5% and above despite using only 256×90 observation points. performance
    Source sentence
    network further improves the Leland approach for transaction costs at 0.5% and above, despite the small number of observation points (256 × 90) used to train it.
  • Training on overlapping sequences from a single simulated series yields hedging accuracy comparable to training on independent sequences. performance
    Source sentence
    The calibration speed of the neural network is comparable to that achieved when training on independent sequences of length 30
  • Overlapping sequences extracted from real S&P 500 data can be used to train the network, suggesting feasibility on real‑world datasets. performance
    Source sentence
    256 sequences of 31 data points are extracted from this dataset and used to train the neural network to price options with a maturity of T =
  • The method does not require simulation under a risk‑neutral measure, simplifying data requirements. theoretical
    Source sentence
    In our experiments, it was not necessary to simulate trajectories under a risk‑neutral probability; instead, we simulated under a probability measure where the drift does not ne…

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 45 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.