But the thrust of "doing the math" is that although we know that rape stats are difficult, they're not that difficult.
OK, we won't get exact figures, we can't; but we can get a steer. And the steer is, that the overwhelming majority of men who are accused of rape and are then do not end up in a court of law being found guilty, are actually rapists.
Yes once you get to court your chance of conviction versus other crimes is actually higher for rape; but seeing as how most rape cases don't make it to court, so what? Only the most horrendous and clear cut cases get there - the ones most likely to secure a conviction - and still a large number, though a minority, get off, even though the probability is that most of them are guilty. As to the cases that don't get to court, mostly those men are rapists. That's without taking into account the 85-90% of rape cases not reported in the first place.
Don't do the maths then, because the maths are uncertain. But what isn't uncertain, is the general gist. That when a man walks free from a court or a police station not having been found guilty or charged with rape, the statistical probability is that he is in fact a rapist.
It's one of the reasons we ought to be improving rape conviction rates; because for the tiny minority of men who are in fact falsely accused, we need to ensure that the probability of them being guilty is far far lower than it is now so that we can all assume the same level of probability of their guilt, as we would a man accused of burglary or fraud or whatever.