2. A miracle is an event which violates a law of nature.
3. We ought to believe propositions which have a high(er) probability of being true.
4. Either we believe John's testimony is proof of a violation of a law of nature, or we don't.
5. If we believe John's testimony is proof, then we believe that a law of nature was violated.
6. But the probability that a law of nature is violated is lower than the probability that John's testimony is false.
7. Therefore, we ought to believe that John's testimony that a miracle occurred is false.
What if, instead of denying premise 1 as a result of the reductio, we deny premise 2? This seems to make more sense; for why would God set in place laws of nature that He would have to break in order to do what He wills? It seems plausible that the laws of nature are more flexible when a supernatural being is involved. For instance, maybe there is a law of nature that every human dies a bodily death; but does that mean that it is not possible that there is another natural law that if the spirit is then raised to life, the body is also raised to life? This would entail that eventually every human whose spirit is raised to life is also resurrected in body. That this has not yet happened is not conclusive proof that it will not happen. In fact, the Christian doctrine is that every human dies but is also raised to life at the second coming of Jesus. Can we not count this as a law of nature which, having not been experienced yet by us, appears implausible? Then a miracle, like the resurrection of Jesus, would be defined as "an event which occurs so rarely, though according to a law of nature, that it almost merits disbelief"?