This is why we use statistics and probability when analyzing data. No model is going to perfectly describe everybody. A couple of individual counterexamples is

This is why we use statistics and probability when analyzing data. No model is going to perfectly describe everybody. A couple of individual counterexamples is not necessarily strong evidence against a trend or association. Also there are plenty of other athletes who swear by marijuana usage, you aren't even unique in that! My point is that the stereotype of a heavy marijuana user is very different from you, and that said stereotype could provide an alternative causal mechanism that we need to be accounted for in research.