I am looking for the name of a logical fallacy that denies validity of induction based on an observed and repeated pattern (AKA statistics) by citing exceptions to the pattern which are outliers of small percentage. Just to clarify, I am well familiar with the problem of induction and am not discussing the validity of using induction. I'm just looking for the name of something that, to me, appears as a commonly (ab)used logical fallacy.
E.g. insurance companies use statistics not to establish universal truths but to guide their investment risk. So, based on observed patterns, they established that drivers below 25 years old are more risky. So if Joe Denialist comes and says that the criteria is invalid because he's 24.5 years old and has been driving for 7 years with not even a speeding ticket, that would be "anecdotal evidence". But if he goes a step further and perhaps collects some data that presents still an overwhelming minority of cases relative to the documented stats to invalidate the risk strategy, is there a name for that logical fallacy?