Basically it is possible to produce objective evidence that is also a statistical anomaly. Its still evidence something can happen evem if it's not good evidence for or against a hypothesis.
Or, in gambling, such as poker, players can make informed and good decisions and produce the evidence to support their decisions, but the outcome can still be unfavorable to them.
Odd things where we exeptions instead of the rules. It can all still produce objective evidence, but it has to be further evaluated to know of it's good evidence or not.