The 'god of the gaps' argument, which is pushed very hard by the new atheists[0], is question-begging and therefore worthless. The assumption is that if something is going to be explained at all, it must be via scientific methods. Therefore (they say) the appeal to God should be seen as a poor hypothesis, comparable in kind to scientific hypotheses, but vastly less useful in terms of prediction.
The whole debate is about whether it's rational to think that the entirety of an explanation for something is scientific. Therefore, the 'god of the gaps' argument is circular and question-begging.
The good arguments for God's existence and attributes, which the new atheists ignore[1] or grossly straw-man, involve the most basic of observations (such as 'there is change' or 'there are objects of the same type'), followed by deductions from said observations. They're rational arguments, assuming they're successful, but they're not scientific hypotheses. They do not appeal to things we can't explain, nor to complexity, nor to any alleged design (the last phrase is misleading anyway).
[0] It's in Dawkins, Hitchens, Harris, I think Krauss too
[1] Dennett gives them one paragraph out of 478 pages in Breaking the Spell, and is guilty in this paragraph of a very common straw-man, which shows how little he's read. Dawkins also uses the same straw-man in God Delusion IIRC.