Trading on Emotion: Behavioral and Predictive Determinants of Algorithmic Trading Adoption

Abstract

Algorithmic trading systems have become increasingly accessible to retail investors, yet the psychological and emotional determinants of their adoption remain insufficiently understood. This study examines how emotions, specifically anxiety and joy, in conjunction with trading experience, influence traders’ willingness to adopt algorithmic trading bots. Drawing on survey data from retail traders, we compare the predictive performance of linear regression models with that of flexible machine learning approaches. Nonlinear ensemble models achieved superior predictive accuracy, indicating that emotional variables convey meaningful information beyond demographic and experiential factors. Feature-importance analyses further reveal that emotion-related measures provided the strongest predictive signal. These findings suggest that incorporating emotional dimensions substantially improves models of technology adoption in finance and underscore the need for experimental validation before real-world implementation.