Previosly, I've defended Hume's argument against miracles. As I see it, the problem is this: we must judge the plausibility of historical claims by experience. There is no question that miracles fall quite far outside of our experiences. Therefore, we must judge them extraordinarily implausible, requiring an extraordinary amount of evidence to establish ("the testimony be of such a kind, that its falsehood would be more miraculous, than the fact, which it endeavours to establish," in Hume's words).
Now, it is clear that miracles in fact lay outside of our experience, but need this be the case in principle? Anthony Flew has answered "yes." He does not deny that seemingly miraculous things could be established by repeated experiment. In fact, he was at one time involved with the Society for Psychical Research, though he believed, on Humean grounds, that parapsychologists needed a repeatable experiment to establish their claims. Nevertheless, Flew says, if such a state of affairs ever came about, the commonness of the events would mean they would cease to be evidence of anything theological. For example, if resurrections happened every day, the resurrection of Jesus would not be evidence of his divinity.
A little thought should show that this is clearly wrong. Consider the following hypothetical state of affairs: there is certain condition which strikes without warning, causes victims become violent, as well as developing extraordinary strength and the ability to levitate. There are hundreds of documented cases each year in America alone. The existence of the condition is undisputed to the point that police officers are trained to deal with superstrong, levitating perpetrators. Careful investigations show that Catholic priests have a high sucess rate in dispelling the condition, while Protestant clergy, non-Christian clergy, and scientists are totally unsucessful in stopping it. Clearly, such a state of affairs would have important theological implications, and the fact that the events in question would be frequent would not change this in the least.