Given the dependence on the aggregator, I wonder how much you gain by having a committee of size greater than 1?
Any analysis depends on assumed aggregator behaviour of course. If the aggregator reward is not sufficiently large we can ignore economics and imagine a binary world of byzantine and honest nodes. In this world byzantine aggregators simply won’t produce an IS if it includes “blacklisted” transactions and honest nodes include everything. Does having a committee of greater than 1 add anything in this world? It doesn’t seem like it on first pass.
Naturally economics do play a role some role. One question is how expensive do you need to make censorship to change aggregator behaviour? Currently some staking pools take a decently sized fee which suggests that it might need to be quite large for them to start losing customers.