Ethereum PoS has finality. It doesn't matter if someone later comes out with a different version: full nodes accept the first valid block it sees as final and immutable, after a short period of time.
Is this perfect? No, but maybe its worth all the energy savings?
If you run a node, by definition you can never end up on the "wrong" chain, because your node,s chain is immutably the chain, and so is everyone else running a full node, thus making it the canonical chain.
As a newcomer - how can I tell the "right" chain apart? In Bitcoin I simply choose the one that burned the most work - it's trivial. In PoS I do what? The longest? Its trivial to produce a long chain. The longest and that have many different public signatures?
The one that everyone else is using. Look at websites, ask friends, etc. After you do that once, you've got the right one from then on. Kinda trust-on-first-use, which most of us use for SSH anyway, and you can verify that you and the people you know are on the same chain beforehand.
I feel like the "I would just choose the chain with the most hashpower" idea kinda falls down in that you still have to choose which crypto - Bitcoin, any of the bitcoin forks, Ethereum, Ethereum Classic, etc. You're relying on the internet etc for information to help you decide which cryptocurrency you're going to use when you start out, you can do the same thing for which Proof-of-Stake chain. Most of the time, there should be one obvious choice, more obvious than which crypto to use, I imagine.
How it's better than WebMoney then? About other cryptos - well, yeah, there can be only one. If you don't control 51% of overall humanity hashrate it's not secure (and not better than WebMoney).
I don't know much about how WebMoney works other than what I just read on Wikipedia, but it looks like WebMoney might be best compared to a permissioned blockchain where a group of companies mine the blocks.
Proof of Stake has the advantages that anyone (with sufficient money, or by gathering into pools) can become a validator, and the system itself provides extremely strong incentives not to cheat by slashing their stakes if they do. This slashing weeds out bad actors over time, as well - once their money's gone, they can't stake anymore.
I could imagine all sorts of shenanigans that could cause a small group of companies operating a money system to cheat, either of their own accord, or forced to by a government, organized crime, hackers, etc. Much harder to attack a huge self-healing group of globally distributed validators with extremely strong incentive not to cheat.
Well I too could imagine some sorts of shenanigans that will mislead the majority of validators especially in PoS that doesn't lock any of validators computation resources. It's even worse in the sense that a validator can be anonymous and doesn't hold any legal responsibility (in contrast with WebMoney). Thousands of validators could be one entity now or in future and you'll never know that (in contrast with PoW). Some infuencer (just like Vitalik) can come and convince validators to mine incompatible fork of Ethereum - you can do this for free and won't be slashed on main chain - and why not, maybe this fork will prove better and people switch (yeah this can happen with any crypto that's not Bitcoin)?
All I'm saying that Satoshi really produced the one undeniable cryptocurrency. His idea is sound, you don't need any friends or websites, or "common" knowledge besides his simple idea of the most work burned chain. Anything other is really questionable.
Is this perfect? No, but maybe its worth all the energy savings?
If you run a node, by definition you can never end up on the "wrong" chain, because your node,s chain is immutably the chain, and so is everyone else running a full node, thus making it the canonical chain.