Earlier quoted context omitted.
> statistical defensibility Requesting an arbitrarily high standard doesn’t create any obligation. Evidence of a high standard does.
There's no obligation in either direction in this context (idle chitchat) unless of course you care to convince someone of something. He objected to what was provided and you accused him of ignoring evidence. I'm voicing agreement with his objection. The original claim was one of a statistical nature. Thus any purported evidence should be expected to match.
If a group is more likely to be X than the average population then being a member of that group is statistical evidence you are X. Really when referring to statistical evidence here it’s an indication the evidence is of a very low standard not a high one.
He provided evidence that cops are more likely to be white supremacists which doesn’t actually mean much which is kind of a point on its own. That being it’s the same low standard as used by actual racists, but as you said there’s no actual obligation to go beyond such wordplay. Personally I was very amused by the whole thing but obviously it’s quite offensive to some people.
Taking the other side of this one, you could say something like “sure the odds at least one of them are white supremacists is non trivial, but suggesting all seven are is unlikely.” Again not a strong argument, but it’s at minimum an actual argument.