Public policy problems are increasingly being characterised as wicked or tame problems, assuming that this classification is also meaningful for attempts to effective problem-solving. But do distinct ‘wicked’ or ‘tame’ problems empirically exist? We investigate 37 water-related problems in Germany, based on interview-based data on problem wickedness and official data on policy delivery. Our analysis clearly reveals four clusters of water governance problems (system complexity, uncertainty, tame and wicked problems), based on variations of three factors of wickedness (goals, uncertainty and system complexity). These clusters of problems vary in their effects on different dimensions of policy delivery (goal formulation, stages and degrees of implementation of measures), with significant effects on goal formulation and the number of measures ‘in construction’. These empirical insights may contribute to a more systematic design of governance strategies for addressing water governance problems in practice.