Well, sure, but the standard model also says it does. If string theory plus whatever additions are needed to make it "save the phenomena" (predict the observed

Well, sure, but the standard model also says it does. If string theory plus whatever additions are needed to make it "save the phenomena" (predict the observed world in quantitative detail -- which hasn't happened so far) turns out to be simpler than the standard model, then at that point it'd make sense to move from talking about new predictions. Otherwise we're talking about predictions that aren't really predictions at all, in the sense of separating string theory from competing theories such as the standard model. I think I'm using a well-accepted understanding of science which I'd taken for granted, instead of moving any goalposts. A theory retrodicts some of the same phenomena as its parent theory, as an assumption: that's not evidence that the new theory is more true than the old theory. What I started the thread with was an operationalization of "what if string theory is not true?" It's not a perfect one: string theory could also become accepted as true by finding a way to make it retrodict only things already observed, but with a less complex theory than the standard model. But come on, we're very far from there.