40 minutes for talk & 20 minutes for discussion In the summer of 2020, a large subset of the mathematical community agreed to boycott collaboration with the police. But how exactly have mathematicians collaborated with the police? In this talk, I will present a selective overview of predictive policing and predicting police, with a focus on network approaches, ranging from social network analysis of crime to network models of police misconduct, with a view towards a critical mathematical theory of policing. In doing so, I will make the case that there remains much interesting and urgent work to be done in this space.