I think the absurd old spectre of “Classical Knowledge” is rearing its head here.
Everything in the universe is “unknowable” in classical terms, even the most trivially predictable things.
This is also why a priori argument is such a waste of time.
Hence, I think likely outcomes isn’t just a good framework, it’s the only possible one.
That we are sometimes very bad at predicting things isn’t really an indictment of the framework itself.