Supernatural miracles which defy reality and then Miracles such as miraculous healing on someones deathbed, where Docs say its unexplainable. Of course this last one can be easily called "goodluck".
As mentioned earlier, statistically improbable things happen all the time, and the sheer number of people in the world dictate that sometimes it happens to someone.
People are notoriously bad at judging statistical chance and they tend to put way too much weight on anecdotes.
Ill try to illustrate with an example.
Weve all heard the story about someone who dreamt of their uncle/grandma/friend/whatever the very night they died. Usually people take this as a sign of some kind of psychic connection or something equally silly. And they invariably forget to do the math.
In the US there are in the excess of 300 million people. Lets be generous and say that about half of those people are adults, about 150 million. Now, how many of those people do you think dreams about someone close to them every night? One percent? Heck, lets go all out and call it one tenth of a percent. That still leaves about 150.000 people dreaming about loved ones every night of the year. Now, in any given day, how many people die in the US? According to the CIA Factbook that number is 6,744. So lets quickly calculate the chances that someone who died had someone dreaming about them the very night they kicked the bucket.
Each person in the US has a statistical chance of someone dreaming about them each night of 0,05% (150000 / 300000000 x 100). Multiply that with the number of people who on average die each night and you get in the excess of 300%. In other words, this should happen to around three people in the US every single night!!! Even given leeway for the fact that people often doesnt remember their dreams this phenomenon should crop up pretty damn often. But these are only the hits. When doing science we have to keep track of the misses as well, and in this case we still have at least 6741 people who died without anyone at all dreaming about them. The problem is that we never hear about the misses.
The same calculation could be done with 'miraculous' recoveries and suchlike.
There is little in the way of consistency, and the fact of the matter is that sometimes, by pure chance, people just get lucky.