Keller’s second “clue” that god exists is the anthropic principal. (read the link and it argues against Keller…) As he describes it there are many cosmological constants that must be tuned within a small margin of error for the universe to be as we see it. Keller says it is as if some supernatural force created the universe by tuning several dials very carefully so that we could exist. Keller grants that this argument is not a slam dunk and gives the multiverse hypothesis as an alternative explanation, that is, that there are many many universes so one is bound to be as this one is. He dismisses this explanation, though by relaying two analogies. One is that a man deals himself four aces in twenty straight games of poker. The other is a man sentenced to death by a firing squad of fifty expert marksmen and is not hit. In both cases you could say that there are infinite universes, so one in which this rare thing occurs is probable, or certain, but in both cases Keller says nobody would be convinced by this argument.
This completely misunderstands that anthropic principal, however. Both analogies are false. The point of the anthropic principal is that no one could exist to observe any universe that wasn’t capable of supporting life. In both analogies there is no accounting for this, which is why they are not applicable. So ignore those, they’re just distractions, straw men.
Lets get back to the actual universe. It is true that the cosmological constants, if not as they are, would not allow a universe as we know it to form, and many possible combinations exist where it is difficult to imagine life existing at all since stars and planets would not form. However, this doesn’t mean anything. Although there are infinite possible values for the cosmological constants (assuming there aren’t some unknown constraints somehow), there are really only two categories. The universe could exist in such a way that life exists, or not. In the first case the life that exists will evolve to fit whatever conditions exist as best they can, so the universe will seem “fine tuned” to them. If not, then nothing will exist to observe the universe and it will go unremarked. Because of these two possibilities it is impossible to calculate the true odds of a universe supporting life, except that we know it is non-zero. (Unless we can somehow detect, count, and determine the cosmological constants of other universes) Since observers can only exist in, and observe, one part of the fraction, the other part is indeterminable. From what we can observe the odds of a universe supporting life are just as likely to be one as one in a trillion trillion.
Another way to think of this is from the perspective of a null hypothesis. Lets assume there is no god, that’s the null hypothesis. What is predicted? We would expect the universe to appear to be fined tuned for any life within it, for the reasons explained above. Now lets compare to an alternate hypothesis, that there is a god. What is predicted? We would expect the universe to be fine tuned for life within it. Since these two predictions are identical the fact that the universe is fined tuned is not useful in determining if there is a god or not.
Keller, in his Intermission chapter, wrote some things that indicated to me that he knows how science works. He should be able to see that this argument is useless.I don’t know why he chose to keep it in his book.