Gravitational theory is a theory about how things work today can be experimentally tested. Likewise atomic theory and germ theory. These theories are useful for (1) predicting what can happen, and (2) predicting what one may be able to do to cause desired things to happen. If evolutionary theory were purely applied to present and forward-going studies, there would be no objection to it.
The problem is that evolutionary theory is being applied to the past, when there are severe limits as to how well any scientific theory can be applied in that direction. To be sure, sometimes things are pretty obvious (e.g. if the floor of a buildng contains some materials, and the ceiling seems to have holes in the approximate shape of those materials, one could usually conclude that gravity caused the materials to fall from the ceiling onto the floor). In many cases, though, trying to ascertain the past by examining the present is fraught with uncertainty. For example, one might try to guess how the earth came to have its moon where it is (meteor strike, meteor capture, or whatever) but there's no way of absolutely proving any particular theory to be correct.
Can you offer any non-trivial examples where "retrograde science" (trying to predict the past) is considered infallible?