Revolutionizing Responsible Gaming: A Data-Driven Solution for Managing Player Risk and Enhancing Protection
Session Title
Player Protection, Gambling Literacy, & Risk Reduction
Presentation Type
Paper Presentation
Start Date
28-5-2026 12:00 AM
Abstract
Responsible gaming is essential for sustainable betting ecosystems, yet most approaches remain reactive. This work presents a scalable, data‑driven solution that achieves highly accurate predictions of high‑risk gambling behavior from past activity patterns, supporting operator interventions before harm develops. Using behavioral and financial data, we engineered features capturing frequency, intensity, variability, and trajectory of player activity. A logistic regression model trained on a dataset of 145,545 players achieved strong performance (ROC AUC 0.982, accuracy 96.8%, recall 92.2%) in identifying behavioral patterns indicative of emerging risk leading to operator‑imposed exclusions. Feature importance analysis highlights the predictive value of betting trajectories, impulsive session patterns, and deposit behaviors, offering actionable insights for targeted harm‑prevention strategies. Beyond detection, this solution enables personalized interventions, allowing operators to tailor harm-prevention strategies for maximum effectiveness. By operationalizing predictive analytics, responsible gaming moves beyond compliance into proactive harm prevention, supporting sustainable, player-centric practices. This work was carried out in collaboration with Zurich University of Applied Sciences (ZHAW) and Sportradar, as part of an Innovation project supported by Innosuisse, combining academic expertise and industry experience to ensure rigor and practical applicability.
Revolutionizing Responsible Gaming: A Data-Driven Solution for Managing Player Risk and Enhancing Protection
Responsible gaming is essential for sustainable betting ecosystems, yet most approaches remain reactive. This work presents a scalable, data‑driven solution that achieves highly accurate predictions of high‑risk gambling behavior from past activity patterns, supporting operator interventions before harm develops. Using behavioral and financial data, we engineered features capturing frequency, intensity, variability, and trajectory of player activity. A logistic regression model trained on a dataset of 145,545 players achieved strong performance (ROC AUC 0.982, accuracy 96.8%, recall 92.2%) in identifying behavioral patterns indicative of emerging risk leading to operator‑imposed exclusions. Feature importance analysis highlights the predictive value of betting trajectories, impulsive session patterns, and deposit behaviors, offering actionable insights for targeted harm‑prevention strategies. Beyond detection, this solution enables personalized interventions, allowing operators to tailor harm-prevention strategies for maximum effectiveness. By operationalizing predictive analytics, responsible gaming moves beyond compliance into proactive harm prevention, supporting sustainable, player-centric practices. This work was carried out in collaboration with Zurich University of Applied Sciences (ZHAW) and Sportradar, as part of an Innovation project supported by Innosuisse, combining academic expertise and industry experience to ensure rigor and practical applicability.