I do think it's important to draw a distinction between different kinds of predictions.
In this thread, folks have talked about predictions that are made by a scientific theory, for example, the nested heirarchical pattern of species. This was a classification system that predated Darwinian evolution. Since it is "predicted" by the theory, so the argument goes, it is considered validation of the theory. (I think it's more accurate to describe this kind of a prediction as an explanation, but for present purposes will accept it
arguendo.) The problem with this type of argument is that it's possible to construct a theory is such a way that it appears to make this kind of "prediction." And creos have gotten a fair amount of mileage out of doing just that.
For example, creos will say that special creation predicts that most mutations will be harmful and then argue that, since most mutations are harmful, that validates their theory. We have to carefully evaluate any post hoc "prediction" of this sort.
Another kind of prediction, called a "retrodiction" by some here, is a claim that certain evidence will be discovered that is, as of the time of the prediction, unknown. Einstein's prediction about the bending of light is a wonderful example. These are considerably more forceful, because they are predictions whose accuracy is not yet known. The more detailed such a prediction is, or the more unlikely it would appear that the prediction should be accurate, the more power it has toward validating the theory when the prediction is borne out.
Here is a partial list of some of the hundred of predictions of this sort that evolution has made that have been fulfilled.
Page Not Found | Department of Chemistry
Those who would sacrifice an essential liberty for a temporary security will lose both, and deserve neither. -- Benjamin Franklin