I think in most of these cases, the accusations are correct. But some of them aren't correct and we must protect the innocent as well.

And me, as an outsider on the internet, has no business trying to pick a side or work out who should be punished. That's the job of the legal system.

Why do you think most accusations are correct?

The majority of people are not malicious and would not make this stuff up. Some obviously are and will but my feeling is that these people are the minority.

Not sure the math makes sense on that.

Assuming a two-party interaction, for every person that makes a true accusation of abuse, there is one malicious actor (the abuser). For every person who makes a false accusation, there is one malicious actor (the accuser).

Regardless of the ratio of true to false accusations, the number of malicious people in the world remains unchanged.

This was an interesting thought but I think the difference is that the defender has a much greater incentive to defend than an accuser has to accuse. There is no downside to not falsely accusing someone but there is a major downside to not falsely denying an accusation.

So it would be expected that almost every accused person would defend themselves when they would not lie if the sides were reversed.

There is an assumption you're making that is missing from your analysis. Yes, it does seem likely that an abuser or what have you will be highly incentivized to defend themselves, more so than you'd expect people to falsely accuse people. What is in question is how likely you think someone being an actual abuser/creep/etc is in the first place. You're not comparing apples to apples.

Because effectively you are saying that this kind of bad behavior is more likely than the bad behavior where one person would lie about another. And it is not clear why you think one type of antisocial behavior is more common. People lie every day, oftentimes to their own detriment. Likewise people are creeps every day, again often to their detriment. I think it is far from clear at this point which behavior is more common.

This is still bad reasoning.

In a pool of one hundred people, where one is an abuser and nine are false accusers, you'll have one false denial and nine real denials.

If it's nine abusers and one false accuser, you'll have the opposite ratio.

Without demonstrating which toy world applies, you can't know what's going on.