argument from performance

One of the things for which i am looking in my current reading of ancient material is what i call the "argument from performance". The contemporary version of this within my field is roughly the claim that sciences/engineering does know something about the real world and we can tell that it does because of things like the Deep Impact type missions. That is ... the amount of things our species must have right in order to have successfully completed the Deep Impact probe missions is astounding.
What i am looking for in the literature is, then, when did this form of argumentation begin, and are there cases where it has been wrong?

I am reading A.A. Long's book on Hellenistic philosophy after Alexander and i did come across a related argument tonight. The Stoics claim that their concept of determinism must be right partially because divination and astrology work. Since, obviously, we do not think that they do (but make great allowances for confirmation bias making it seem as it does), perhaps this counts as a strike against. And yet, if a Stoic were to list the things that would have to have been wrong if astrology was not correct would be fairly short and highly theoretical ... not so for Deep Imapct.

So ...