While you’re right about the liveness/halting point, I’d caution you from depending on the data sets you cite for Ethereum. I don’t believe either are highly maintained in the last few years, have large ‘unknown’ blocks that are mostly the same big operators as the other buckets, and Lido is mostly the same operators (with better distribution) too. If we don’t have a finality failure from the two biggest operators going offline we miss it by a very small margin and 3 is quite likely enough. Ethereum from a node perspective is very centralised vs the rest of the PoS chains where foundations weigh in more actively to decentralise their markets, (and that’s why we should CFI 7716 :) )