The argument may be valid, but only as far as it argues an irrelevant case: Hardly any religious person would suggest that God came into existence sometime after the creation of the Universe. Am I missing something?

You're suggesting half of the Bayesian argument: Given a universe without God, it is quite probable that religions exist. But what I would like to see is the other half: Given a universe in which (some) God exists, the current universe is nearly impossible.

"It's a valid strawman argument" isn't my conception of a valid argument. Validity includes relevance, not just internal consistency. The point of religion is that there is no such thing as disproving the existance of god, just as there is no such thing as a valid proof of the existance of god. Religion is built on faith, which is the rejection of proof. Dawkin's faith in science is a religious view, not a scientific one.

An interesting point.

I think ultimately he has the same problem of first causes that the argument for God does. In both arguments you have to draw a box around the universe/timespan/laws you will consider and then make an assumption about the initial state. In Christianity, the box is drawn outside the universe itself by considering God to exist outside of time and space and predating the universe itself. Dawkins just draws a slightly smaller box, but neither argument is without assumptions.

Besides, these arguments all sort of start on shaky ground anyway because the nature of faith is that it's not based totally on things that are empirically provable. By it's nature it isn't something that claims to be rooted in empiricism. No matter what side of the argument you fall on, at a certain point after weighing all the evidence there is a chasm you have to cross between what is actually provable, and what you believe to be true.