Hello Killari,
Algorand’s secret non-single leader election is indeed another plausible route to achieving validator privacy. However, while it might seem simple in theory, in practice it complicates both the fork-choice and the networking subsystems.
In particular, see Vitalik’s recent secret non-single leader election proposal on how it complicates the fork-choice and also opens it up to potential MEV time-buying attacks. Furthermore, by making the fork-choice more susceptible to forks and reorgs it makes it harder to apply potential improvements to it.
With regards to networking, a non-single leader election considerably increases the communication within the P2P network, especially in a sharded world where blocks are considerably bigger. For this purpose, Algorand uses a mix of smart gossiping and timeouts which would need to be adapted to Ethereum’s P2P logic (see section 6 of Algorand’s paper).
All in all, I agree that Algorand’s approach is still worth examining further and seeing how its tradeoffs apply to our use case.