I don't know why anyone thinks this could be a reasonable argument. It allows anyone to argue, "Well, that may be what the evidence says, but that's not what actually happened because the evidence was changed, and no trace was left behind of the change."
It has other even more practical implications.
I believe that my car will stop because the last time I stepped on the brake pedal that was what happened. Unless the evidence of that was just inserted and it was not the break pedal but the accelerator that I stepped on last time.
The medicine that the doctor is prescribing for me was shown to be three times as effective as a placebo. Unless of course that wasn't really the results of the test and the placebo was actually far safer.
The assumption that the past changed (the reality of the past, not what we know of the past) is one of those ideas that must simply be rejected and thrown away if there is to be an progress whatsoever.
Aslan is not a Tame Lion