The Impact of NBA-Related Signals and Market Factors on the Short-Term Price Movement of DKNG Stock

Abstract

This study investigates whether NBA-related signals and basic market factors can help predict the short-term behavior of DraftKings (DKNG), a major U.S. sports betting company whose business activity is closely tied to NBA games. Using daily price data, market benchmarks, technical indicators, and selected game-intensity variables, the study builds predictive models for two tasks: next-day return direction classification and next-day volatility regression. Three machine learning models—Ridge, Random Forest, and XGBoost—were trained and evaluated using time-ordered rolling validation. Results show that non-linear models significantly outperform linear baselines. Direction prediction improves only modestly, but calibration scores become more reliable. Volatility prediction shows stronger improvement, especially with technical indicators. The findings suggest that combining market information with NBA-related signals provides meaningful predictive power, especially during high-attention game periods. This work offers a foundation for future research incorporating higher-frequency betting data, injury reports, and sentiment indicators. Such extensions could enhance understanding of how sports-event dynamics interact with financial market behavior in real time.