When do more severe punishments not deter crimes and when are stronger incentives for control personnel ineffective to motive control? This contribution shows when incentives for crime control have paradoxical effects. It follows from game theoretical reasoning that the opposed interests between criminals and inspectors lead to strategies, where both seek to outsmart each other. If both do so, ego's incentives only affect alter's behavior. This "strategic incentive effect" implies that higher punishments do not cause less crimes, but less inspections. This was experimentally corroborated by Rauhut (2009). The second implication is that stronger inspection incentives do not cause more inspections, but less crimes. This paper studies the second implication under farsighted strategic actors, shows its robustness by agent-based simulations of backward-looking learning and provides experimental evidence. In the laboratory experiment, 200 subjects were partitioned into "citizens" who could steal money from each other and "inspectors" who could invest in inspection to detect criminal citizens. Results confirm that stronger inspection incentives cause less crime. However, actors are less strategic than predicted: stronger inspection incentives also cause more inspections. This is discussed by differentiating between the "strategic incentive-effect", where ego's incentives only affect alter's behavior and the "own incentive-effect", where ego's incentives also affect ego's behavior. Conclusions discuss alternative models of rationality and heuristics and how the presented findings may be used for constructing novel theories on crime and social norms. Also policy implications are discussed.