Does POW provide a superior "proof" than POS?

Recently, I needed to explain to a friend the difference and pros/cons between POW and POS. It just crossed my mind that "proof" in POW is self-contained in the sense that once you see it you know that work (and thus energy) was expended. Thus, it's easily verifiable and cannot be faked.

In contrast, "proof" in POS does not carry the same properties since "stake" is dynamic. It can't have a self-contained, stand-alone proof of the leader selection. I take it to mean that, at least theoretically, possible to fake it. I could be wrong here.

I wonder if there's been some scholarly discussion on this. I looked at Cardano's Ouroborous paper (supposedly the most scholarly-studied blockchain) but there is no hint that this is being addressed.