Ah, I think I see what the insight was - you're not actually trying to solve logical uncertainty, you're trying to solve a specific problem of UDT with logical uncertainty, as outlined in your third link, and you're doing this by replacing a search through proofs with an iteration over "logical worlds" where the proof would be true or false.

Am I close?

EDIT: well, I guess that's not totally true - you are trying to solve logical uncertainty, in the sense that logical uncertainty is going to be defined relative to some bounded model of decision-making, so you're proposing a model.