No need to reinvent the wheel. Re-examination and changes should, and do, occur to all scientific theories as more knowledge is gathered. This is why the Newtonian view of gravity - while still valid as an explanatory model in most cases - was extrapolated on in the last century.
As for black holes, I can't see them or touch them. They are unmeasurable. To date, no empirical information can exit the event horizon.
You make it sound like scientists first dreamed up a fantasy concept out of thin air, then started theorizing around that fantasy concept. I would think it rather started out with observations of interactions in the universe that needed an explanation, like any other scientific discovery.
Despite its interior being invisible, a black hole may reveal its presence through an interaction with matter that lies in orbit outside its event horizon. For example, a black hole may be perceived by tracking the movement of a group of stars that orbit its center. Alternatively, one may observe gas (from a nearby star, for instance) that has been drawn into the black hole. The gas spirals inward, heating up to very high temperatures and emitting large amounts of radiation that can be detected from earthbound and earth-orbiting telescopes. [2] [3] [4] Such observations have resulted in the general scientific consensus that — barring a breakdown of our understanding nature— black holes do exist in our universe. [5]
It seems it needs to be said again and again; science doesn't have all the answers, and that's how it has to be. But we can further our understanding by meticulously following the scientific method as we collect more data about the world. This black and white "science doesn't know everything, so I'll go with whatever hunch I feel is better" thinking doesn't hold water. We have to rely on our current best understanding, based on our current set of data. Sometimes scientists have to put in 'place holders' in areas we don't yet have enough data, like is the case with dark energy and dark matter. But it's not taken completely out of thin air, but are based on observed effects.