In his book, Gregory Dawes looks at whether the idea of God works as a successful explanation.
Dawes begins by carefully defending his controversial stance that there are no in principle objections to positing God as an explanation. In doing so, he directly confronts the idea that supernatural explanations in general, and theistic explanations specifically, are outside of the scope of science. Instead, Dawes argues, if we take it that God will act rationally to create an optimally good world, we have a hypothesis that is testable and falsifiable in principle, well within the scope of science, and thus worthy of consideration as a possible explanation.
Now it seems at first that Dawes, in his defense of God as a potential explanation, would be very sympathetic to God as a successful explanation. Instead, Dawes points out that even if God might potentially explain things, in every criteria we use to evaluate a successful explanation, God does very poorly.
When examined in light of "testability", whenever theists do posit testable accounts of God, theism tends to fail those tests. In light of "background knowledge," a disembodied, timeless being goes against our established knowledge of humans as embodied, finite, and temporally limited. In light of "past success", naturalistic explanations have been wildly successful, with supernatural explanations never reaching any success whatsoever. The theistic hypothesis similarly fails in light of explanatory virtues like simplicity, ontological economy, and informativeness.
Unlike many philosophy books, Theism and Explanation is very readable and not prohibitively long, with only 166 pages of text outside of the index, bibliography, and notes. Dawes shows his fair minded nature by avoiding any of the easy criticisms of God that many atheists take. Instead, he builds the strongest, most defensible view of God as an explanation, and then proceeds to demolish that very view in terms of his explanatory virtues.