And how many "holes" proved to be incorrect? I think that using the history of past true positives is a weak argument for dark matter.

I'm not sure it's as much an argument for dark matter and more an argument against the idea that if you must create a force or type of matter (or matter itself) in order to make a model work, then the flaw is in the model. By showing examples of this being a useful method in the past, it helps discredit the idea that the model has to be wrong, as opposed to proving idea that the model has to be right which are two different arguments entirely.

Obviously it's not sufficient evidence for dark matter. But I'm pretty sure that more often than not predictions of unobserved phenomena based on theories with great predictive power and mathematical elegance, turned out to be correct. Although sometimes, like with the neutrino, things turn out to be slightly different than expected.

In my opinion Verlinde's theory is a better candidate as a replacement of dark matter, since it doesn't require changing the theory of gravity as drastically as MOND does. That said, his theory does require a combination of general relativity with quantum physics, which is tricky. Compared to both just adding a bit of dark (not necessarily invisible) matter is a lot simpler.

Black holes were discovered in theory before any there was any evidence too.

Although interestingly we're not necessarily sure they're exactly what we think they are - you can come up with objects which would give the same observations we have but be quite different : https://en.m.wikipedia.org/wiki/Gravastar