Wes apparently liked my use of "random" in my description of Creationism, and I got to thinking in the shower. Imagine you roll a die several times and it consistently comes up six. Three conjectures come to mind:
1) The die is heavily loaded so that it comes up six all the time.
2) The die just randomly came up six all those times.
3) The unknowable magic man in the sky wanted it to be six all those times.
The first idea is a testable hypothesis. It predicts that you'll find the die to be unbalanced if you examine it carefully. It predicts that the most likely result of further rolls will be more sixes.
The second idea is a null hypothesis. It predicts that there's a 5/6 chance you won't roll a six again. Technically, this one can explain any result, but each roll of a six makes it less and less likely to be true.
The third idea is unfalsifiable, since we can't make predictions from it. If you roll a six, you can say the magic man's in the same mood. If you roll a five, you could just as easily say he changed into a five-ish mood for that moment. Same with the other numbers.
I guess it shouldn't be hard to guess that #1 is pretty much every scientific theory ever and #3 is what Creationism boils down to: It's actually less predictive than randomness.