The first criteria to identify science was Aristotle’s: science is certain and all else is doxa. Science gave up on conclusions transcending doxa but held out hope that it could be identified by some method. So science becomes the method to find what is most probable.
1.) Probability is secondary or instrumental. Probability is a development of background assumptions, which you can say are probable but you can’t treat that way. Far from being a modest, humble approach where we settle for the probable, science becomes a way of systematically forgetting what we’re treating as certain.
2.) So far as probability is real, it is unintelligible. Probability is an index of uncertainty and so of our minds. Something like this is in the world too, but it is chance, which might make for an interesting story but not one where there is any intelligible connection between the terms. And no, a Parmenidean-Einsteinian block universe does not make all narratives necessary. To say that there is a unification of all physical events is to say infinitely less than that there is a unification of all physical events under all possible descriptions. A chance event might have a narrative (after it happens) but it needn’t have a (human) logos.