Lawrence Shapiro (University of Wisconsin) is a talented philosopher. He has recently written a piece for Aeon Magazine which defends in a somewhat more sophisticated form an argument first made by David Hume (1711-1776). The argument is for the conclusion that it is never rational to believe that a miracle (such as the resurrection of Jesus) has occurred on the basis of testimony. The article is well-written and I recommend a careful reading.
According to him we can treat witnesses to miracles as we do tests for rare diseases. If someone says that a miracle has occurred, then in order to evaluate whether this witness gives us good evidence to believe it or not we have to know the base rate of such miracles. Because even if witnesses only make very rare mistakes if there are much more mistakes than events being witnessed, then it is always far more likely that any report is a mistake. (Even if the report is correct it would not be reasonable to believe it.) To return to the medical case, consider if the disease being tested for has only one known case in the history of the world. But grant also that the test is pretty accurate making mistakes only one in a million times. You get a positive test. Is it reasonable to think you are the one person with the illness? No. Because there will be many more false positives the test probably erred.
The reasoning here is pretty compelling and straightforward. The relevance to belief in Jesus resurrection is obvious. This event (if it occurred) is unique in history. A resurrection from the dead has only happened once so far. While groups of generally honest people coming to strongly believe that someone has risen from the dead is also pretty rare, these sorts of mistakes happen far more often than people actually coming back to life (which if the resurrection didn’t happen is zero.) One who didn’t have a high opinion of the early Christian witnesses may think that such craziness happens far more often than actual dead people coming back to life.