For a write up of the general, unavoidable problems of proof of stake, see Andrew Poelstra’s proof of stake FAQ.
As far as Poelstra, that's a rather old document. A great deal of research has been done in the field since.
Cardano's paper: https://eprint.iacr.org/2016/889.pdf
What has happened instead is that the basic / easy to analyze systems were broken, and so complications were added that defeated the simple attacks, but really just make attacking it more complicated. But there's no asymmetric advantage against attackers here -- a similarly complex attack can be made for any proof of stake scheme.
There are a whole lot of very smart people who disagree with you.
I'm done with this discussion as you're simply repeating the same things over and over.
I don't believe either (a) or (b) applies to either xtz or to ada, though ada hasn't yet deployed POS.
I can well believe that there are broken POS implementations, there have been a lot of broken implementations of a lot of things in the space -- but claiming every system is known to be broken or cannot be analyzed is a very bold claim to make.