Paul Draper is one of my absolute favorite philosophers who also happens to be an atheist; more precisely he is a Naturalist, which is to say that he believes that God does not exist, and nothing relevantly similar to ‘God’ exists. He has argued that while some arguments for God’s existence are good, and thus increase the plausibility that Theism is true, still on the whole Naturalism is more probably true than theism all things considered.
In a debate with William Lane Craig, when Draper was confronted with Craig’s Cosmological argument (among others) he used an analogy. He said suppose you have two jars of jellybeans of colors blue or red. Now, in Jar A, 9/10 of the beans are blue, and the remaining are red, while the converse is true of Jar B; Jar B contains 9/10 red beans and 1/10 blue. Now, supposing that the jars were not transparent, and you were given one of these Jars (you do not know which one) – if you were to pull out a blue bean randomly, you would have good reason to think that you were dealing with Jar A. If the next 5 you pulled out were red, however, you would have good reason to think that you were actually dealing with Jar B. In neither case does this probabilistic reasoning prove which jar you have, but the more red beans (relative to blue) one randomly manages to choose from the jar, the more the likelihood increases that it is Jar B you are picking from. So, Draper argues, although there are good arguments for Theism which are very good, and qualify as ‘blue beans’ there are still more ‘red beans’ than not (where Jar A represents the ‘Theistic’ hypothesis, and Jar B represents the hypothesis of Naturalism). So, the likelihood of Naturalism is still, all things considered, greater than the likelihood of theism – so Draper argues.
Now, of course, this stands or falls quite obviously with whether Naturalism is even logically possible, and since I have become convinced that one might make a stronger case for being properly a non-cognitivist with regard to atheism than an atheist could make for being a non-cognitivist with regard to theism, in principle, I think that deserves to be questioned. However, provisionally entertaining Naturalism for the sake of argument, it seems to me there is a much more immediate retort to Draper available; namely, given the Fine Tuning arguments it can be demonstrated that it is more probable, on Naturalism, that the universe really is just a Boltzman Brain imagining a larger universe. This means that, using Draper’s rhetoric, and on the supposition of Naturalism, he ought to believe in solipsism. Moreover one can do the same for something like the “Matrix hypothesis”.
Let’s imagine that the “Naturalism Matrix Hypothesis” has just as much explanatory value (scope, power, simplicity etc.) as a competing hypothesis which we’ll call “Naturalism Realism Hypothesis”. In that case, there would be no way to privilege one to another (though presumably the NRH would be simpler than the NMH). However, let us suppose that NRH is less explanatorily powerful than the NMH – that would mean that on Draper’s suggestion, he ought to believe in the NMH over the NRH. The NMH explains all of the data which the NRH explains just as well as the NRH explains it, but it has an additional advantage of also explaining apparently paranormal events which the NRH cannot explain (Note: think of all Deja-Vue type events, for example, as explained in the movie). I think such a demonstration is in principle possible for something like the Matrix Hypothesis (NMH), and an analogous demonstration is available of the Naturalism ‘Boltzman Brain‘ Hypothesis – NBBH. This provides one with a good argument to undermine a Naturalist who is a ‘realist’ in this sense, and who argues in the same fashion Draper does about the likelihood principle. If they are consistent, they will either find that ‘Theism Realism Hypothesis” is more likely than the NMH, while the NMH (or the NBBH) is more likely than the NRH.