Predicting Mental Performance Drop-Offs in Tennis Using Data Science and Cognitive Modeling

Abstract

Mental lapses or “choking” under pressure are common challenges in competitive tennis. This study proposes a data science approach combined with cognitive science insights to predict when a tennis player’s performance might significantly drop after high-pressure moments. A realistic dataset of tennis performance metrics was simulated, including situational factors (pressure of the moment, prior errors) and cognitive-physiological factors (fatigue, focus level, experience). A machine learning model (random forest classifier) was trained to predict performance drop-offs, achieving around 78–80% accuracy in distinguishing drop-off instances from normal performance. Key predictive features included focus level and fatigue, underscoring the role of cognitive and emotional factors alongside situational pressure. The results align with psychological theories: high pressure and anxiety can impair attention and motor execution, and accumulated errors can trigger a downward performance spiral. Cognitive frameworks, such as attentional control theory and emotional regulation strategies, are used to explain these findings. This hybrid research offers a model for integrating data-driven predictions with cognitive psychology to not only forecast performance drop-offs but also inform interventions (e.g., mental resilience training) to help athletes maintain peak performance under pressure.