If I follow correctly, it seems like you’re proposing a generalization of the “proposal mechanism” that Casper FFG lives on top of.
This seems reasonable in some ways (see proposals for this in full PoS here and here) but also not totally in others. For example, we could switch the proposal mechanism out for Tendermint as the “secondary consensus” you describe, but in this case, Casper FFG doesn’t make much sense (as all blocks in the main chain are already finalized).
So I guess my pushback is: we don’t necessarily need a secondary layer of consensus - we really just need a well-defined forkchoice for some proposal mechanism (where many blocks at many heights can exist without needing to be finalized by a “secondary consensus”).
The advantage are those that come with specifying block forger in advance and would presumably be an overall more coordinated flow of network traffic.
I’m not totally sure how this makes sense. For example, in the PoW-as-secondary-consensus example, you give in the previous example, it doesn’t seem like there is really a way to optimize network traffic (beyond what PoW already does).