Interesting thoughts from Steve Hays about Miracles and reconstructing the past:

In the case of miracles, induction hits a wall. When the subsequent course of events is the result of a miracle, inductive inference can’t go further back than the miracle. It can’t reconstruct the past before the miracle occurred, because the post-miraculous state is not a product of the pre-miraculous state. Induction can only take you from the present to as far back in time as the precipitating miracle. It can’t jump over that to the other side, because the chain of events prior to the miracle is a dead-end. The prior chain of events terminated with the miracle, which represents a new beginning.

This raises a potential problem regarding past-oriented sciences (e.g. cosmology, historical geology, paleontology, evolution). If miracles occur in the past, are they even detectable? What’s the scope of any particular miracle to reset the status quo? That limits our ability to reconstruct the past.

~http://triablogue.blogspot.com/2017/07/miracles-induction-and-retrodiction.html