Deterrence research has evolved considerably since the 1970s when a favourite offence for study was the emerging phenomenon of widespread cannabis use among mainstream populations. The deterrent model of crime prevention has expanded far beyond the study of objective and subjective indicators of certainty and severity to encompass social support, moral evaluations, peer involvement and, most recently, risk sensitivity and situational factors. Most earlier research found no evidence of deterrence of cannabis use, a fi nding attributed to its low actual and perceived risk of arrest and changes in public attitudes. This interview study with adult experienced cannabis users drawn from a representative survey base in Toronto, Canada, where possession is still treated as a criminal offence, provides a more nuanced interpretation of deterrence. The authors found that users had generally inaccurate knowledge about the current law and penalties and believed they would avoid arrest in the future. However, they were not oblivious to the possibility of police intervention, and took precautions such as carrying small amounts and avoiding public use. Thus, users were not unaffected by the law, but rather these discreet practices illustrate the operation of restrictive deterrence, shaping their choices of when, where and how to commit the crime of cannabis use. Further research on deterrence should examine situated choices by risk-sensitive offenders, and should also include cross-national patterns of arrest and user risk perceptions. While cannabis continues to be prohibited by international treaties, the reality of widespread use presents opportunities for innovative deterrence studies into this normalised yet illegal crime. The variation in policies, penalty structures and enforcement across European nations, compared to those in other Western jurisdictions, can foster relevant research for a transatlantic discussion about global drug policy transformation.