MLP-Initialized Reinforcement Learning for Exit Timing in SPX Short-Straddle Portfolios

This paper studies exit timing for SPX short-straddle portfolios using a two-stage supervised and reinforcement-learning framework. Historical episodes first generate ex-post optimal hold/exit labels for training a multilayer perceptron (MLP) using only contemporaneously available state variables at deployment. The trained MLP then initializes reinforcement-learning actors that are fine-tuned with a risk-aware trading reward. Out-of-sample evaluation over 2024-2026 shows that the supervised MLP substantially outperforms passive holding and representative fixed-time and profit-and-loss-based exit rules. Across seven independent seeds, MLP-initialized Actor-Critic averages 37.54% cumulative return and a Sharpe ratio of 1.705, compared with 35.01% and 1.651 for the supervised MLP. Actor-Critic trained from scratch averages only 0.18% return. These findings show that oracle-guided supervised learning provides a stable foundation for reward-based policy fine-tuning in a sparse and path-dependent option-exit problem.

1. 29 Aug 2025

   ARL-Based Multi-Action Market Making with Hawkes Processes and Variable Volatility

   The study combines Adversarial Reinforcement Learning, Hawkes Processes, and variable volatility to enhance market-making strategies, showing improved adaptability in high-volatility conditions and better market simulations.

   arXivIn Proceedings of the 5th ACM International Conference on AI in Finance

   3cites
2. 21 May 2025

   Foundation Time-Series AI Model for Realized Volatility Forecasting

   The study finds that the TimesFM model, with incremental fine-tuning, is effective for volatility forecasting in financial risk management, outperforming traditional models.

   arXiv

   6cites
3. 9 Apr 2025

   Deep Reinforcement Learning Algorithms for Option Hedging

   A comparison of eight Deep Reinforcement Learning algorithms for dynamic hedging found that Monte Carlo Policy Gradient and Proximal Policy Optimization performed best, with the former outperforming the Black-Scholes delta hedge baseline.

   arXiv

   7cites
4. 5 Mar 2025

   To Hedge or Not to Hedge: Optimal Strategies for Stochastic Trade Flow Management

   The paper proposes using reinforcement learning methods to manage stochastic trade flows, offering an alternative to traditional grid-based numerical PDE techniques.

   arXiv

   2cites
5. 20 Nov 2024

   A Risk Sensitive Contract-unified Reinforcement Learning Approach for Option Hedging

   The paper proposes a risk-sensitive reinforcement learning approach for dynamic hedging of options, reducing tail risk using historical market data.

   arXiv

   3cites
6. 23 Oct 2024

   Solving The Dynamic Volatility Fitting Problem: A Deep Reinforcement Learning Approach

   The article discusses the use of Deep Reinforcement Learning in solving volatility issues in equity derivatives, showing its effectiveness and adaptability in handling complex functions and online learning.

   arXivFeatured 3×

   4cites