Proposal weights are being discussed. A more thorough security analysis is underway.
Maybe you could give some more details on how you see this solve the problem?
Since the adversary again participates in these votes, but might show its votes only selectively to some but not to other participants (as in the balancing attack and its variant above), different honest validators might “know” different sets of votes at the second point in time, and again vote inconsistently and hence perpetuate the tie. So it’s not clear to me at this point how two votes would solve the problem here.