Return Predictability in Equity Options - A Supervised Machine Learning Analysis of Delta-Hedged Call Option Returns

Aug 5, 2026

This thesis investigates whether supervised machine learning can predict cross-sectional variation in next-day delta-hedged returns for at-the-money call options and whether the predictions provide economically exploitable information. A snapshot dataset is built from two consecutive trading days for actively traded U.S. stocks, combining option, stock, and market-level variables. Models include OLS, Ridge, and LASSO, as well as nonlinear XGBoost and a multilayer perceptron with an 80/20 split and nested five-fold cross-validation. Nonlinear models, especially XGBoost, deliver the highest out-of-sample accuracy and stable liquidity robustness. Economic usefulness is assessed via quantile ranking and SHAP interpretability, highlighting gamma, realized volatility, option price, and implied volatility as key drivers.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages96

Reading Time4 h 2 min

AccessFree to download

1. Introduction

2. Literature Review

2.1. Empirical Research on Delta-Hedged Option Returns

2.2. Machine Learning in Empirical Asset Pricing

2.3. Machine Learning in Option Return Prediction

2.4. Economic Evaluation of Predictive Models

2.5. Interpretability in Machine Learning Models

2.6. Summary of Key Literature

3. Research Objective

4. Methodology

4.1. Machine Learning and Option Return Predictability

4.2. Supervised Machine learning Methods and Option Return Predictability

4.3. Classes of Models Used

4.4. Economic Usefulness

4.5. Model Evaluation & Interpretability

5. Data and Model Development

5.1. Data Collection

RETURN PREDICTABILITY IN EQUITY OPTIONS: A SUPERVISED MACHINE LEARNING ANALYSIS OF DELTA-HEDGED CALL OPTION RETURNS Lappeenranta-Lahti University of Technology LUT Master´s programme in Business Analytics, Master's thesis 2025 Emaan Shahid Examiners: Associate Professor, Azzurra Morreale Professor, Eero Pätäri ABSTRACT Lappeenranta–Lahti University of Technology LUTLUT Business School Master’s in Business Analytics Emaan Shahid Return Predictability in Equity Options: A Supervised Machine Learning Analysis of DeltaHedged Call Option Returns Master’s thesis 2025 96 pages, 19 figures, 11 tables Examiner: Associate Professor, Azzurra Morreale; Professor, Eero Pätäri Keywords: Supervised machine learning, Delta-hedged option returns, XGBoost, MLP, Economic usefulness, SHAP This thesis aims to explore whether machine learning models can predict cross-sectional variation in next-day delta-hedged returns for at-the-money (ATM) call options and whether these predictions translate into economically exploitable information. The empirical work depends on a snapshot dataset constructed from two consecutive trading days for a group of actively traded U.S. stocks. The analysis integrates option, stock, and market-level variables. The study evaluates three linear models (OLS, Ridge, LASSO) and two nonlinear models (Extreme Gradient Boosting and a feed-forward Multilayer Perceptron using an 80/20 train-test split with nested five-fold cross-validation. The findings indicate that nonlinear models, particularly XGBoost, outperforms linear models, achieves highest out-of-sample accuracy and shows stable performance under liquidity-based robustness checks for this study sample. To assess the economic usefulness of the XGBoost model, predicted returns were sorted into quantiles to examine whether the model establishes a monotonic relationship between predicted and realised delta-hedged returns. The monotonic ordering test confirmed that higher predicted returns are systematically associated with higher realised returns, reflecting meaningful cross-sectional ranking ability. Interpretability for the model is examined using Shapley Additive Explanations (SHAP), which identify the most influential predictors and reveal that gamma, realised volatility, option price, and implied volatility are primary drivers of predicted return, while lower order Greeks and market variables contribute minimally. Overall, the thesis demonstrates that machine learning can uncover short-horizon predictability in delta-hedged call option returns and that these predictions contain economically exploitable information within the limitations of the snapshot-based setting. ACKNOWLEDGEMENTS I am truly grateful to my supervisor, Azzurra Morreale, for her guidance and constructive feedback from time to time. I owe heartfelt thanks to my mom for her unwavering support and to my sister for always being by my side. 4 TABLE OF CONTENTS Abstract (Acknowledgements) Declarations 1. Introduction............................................................................................................................ 9 2. Literature Review................................................................................................................. 12 2.1. Empirical Research on Delta-Hedged Option Returns................................................12 2.2. Machine Learning in Empirical Asset Pricing............................................................ 15 2.3. Machine Learning in Option Return Prediction.......................................................... 16 2.4. Economic Evaluation of Predictive Models................................................................ 18 2.5. Interpretability in Machine Learning Models............................................................. 21 2.6. Summary of Key Literature.........................................................................................22 3. Research

FAQs

What prediction task does the thesis focus on?

It studies whether supervised machine learning can forecast the next-day cross-sectional variation in delta-hedged returns for at-the-money call options.

Which models are compared for return prediction?

The thesis evaluates OLS, Ridge, and LASSO, and also nonlinear models including XGBoost and a feed-forward multilayer perceptron with an 80/20 train-test split and nested five-fold cross-validation.

How is economic usefulness of the best model tested?

Predicted returns are sorted into quantiles to check whether higher predictions correspond to higher realized delta-hedged returns, confirming a monotonic ranking relationship.

Aug 5, 2026

Upload by Theodora