I would not put it that way - of course we want «replicable» Science, as we (and in particular Lakatos) want it to provide predictions, and we value Science pri

I would not put it that way - of course we want «replicable» Science, as we (and in particular Lakatos) want it to provide predictions, and we value Science primarily for that capacity. Nonetheless, disappointing experiments must be treated reasonably (still Lakatos). We do not throw away theories - complex beasts that are best considered as "programmes" also to underline that they involve components - because an outcome was problematic: we investigate till we frame the problem in a sub-theory of "what could have gone wrong". Going towards the case of OP's statements which I countered, to corroborate some theory you conceive experiments to see if it is fact-resistant (Popper - but also fact compliant, which remains the main basic thing), but then you would also refine it according to the complications given by facts. Maybe you need more sophisticated ways to assess your results. > I put that word in quotes cuz I'm not sure what [science] means anymore Well, since we are on Lakatos, let us rephrase him: Science would be something that allows predictions.