This chapter examines the evolution and application of predictive approaches to police misconduct, tracing the shift from reactive accountability systems to proactive risk assessment. It reviews foundational work linking demographic, psychological, and career variables to officer deviance, before highlighting methodological advances in the late 2000s and 2010s. The contribution of important empirical studies is discussed, such as Harris’s work on complaint pathways and Kane and White’s “Bad Cops” research. Building on this foundation, the chapter introduces machine learning models as a powerful alternative to traditional statistical approaches, emphasising their capacity to capture non-linear relationships and complex interactions. This chapter provides an analysis of administrative complaints data, to compare traditional and machine learning approaches to predicting misconduct. Findings underscore the promise of predictive analytics for refining early intervention systems while highlighting limitations related to sample size, data quality, and the risk of bias.