Sean Carroll at Cosmic Variance:
Scientifically speaking, the existence of God is an untenable hypothesis. It’s not well-defined, it’s completely unnecessary to fit the data, and it adds unhelpful layers of complexity without any corresponding increase in understanding. Again, this is not an a priori result; the God hypothesis could have fit the data better than the alternatives, and indeed there are still respected religious people who argue that it does. Those people are just wrong, in precisely analogous ways to how people who cling to the Steady State theory are wrong. Fifty years ago, the Steady State model was a reasonable hypothesis; likewise, a couple of millennia ago God was a reasonable hypothesis. But our understanding (and our data) has improved greatly since then, and these are no longer viable models. The same kind of reasoning would hold for belief in miracles, various creation stories, and so on.
More here.
