コーチング学（偶数演題）／ポスター発表

[09コ-ポ-10] Performance deviations as early warning signals: A monitoring framework for detecting match-fixing in professional basketballA monitoring framework for detecting match-fixing in professional basketball

*WenBin Lin¹ (1. Taipei National University of the Arts)

Sport integrity has become an increasingly important governance concern in professional sport, particularly in competitions operating within regulated betting environments. However, most match-fixing studies focus on betting-market irregularities or post-hoc judicial investigations, offering limited guidance for deployable monitoring systems. This study proposes a performance-based anomaly-monitoring framework to detect potential match-fixing behaviours in professional basketball. Using the 2023 season of Taiwan’s Super Basketball League (SBL) as an empirical case, the framework integrates three player-level performance deviation signals—efficiency deviation (eFF), game score deviation (GmSc), and on-court team impact deviation (+/−)—within an output-oriented BCC Data Envelopment Analysis (DEA) benchmark and player-specific exponentially weighted moving average (EWMA) monitoring. Court-confirmed match-fixing fixtures are used exclusively for validation to avoid outcome leakage. An artificial neural network forecasting module is further incorporated to generate expected performance trajectories, enabling forecast-residual deviations to serve as an additional anomaly signal. The results show that multi-signal monitoring improves detection coverage compared with single-indicator approaches, while severity-based aggregation enables tiered anomaly classification suitable for governance triage under limited investigative capacity. The proposed framework demonstrates how performance analytics can support transparent and deployable early-warning monitoring systems for sport integrity governance in professional leagues.