Fine-tuning is the term used to describe the incredibly precise calibration of dozens of cosmic laws, parameters, and initial conditions whose values have to be fixed within unimaginably small ranges in order for life to be possible. That these amazingly fine tolerances are both necessary for life and actually exist is something of which scientists have only become aware in the last twenty to forty years. Examples of some of these may be found here.
Roberts' argument goes like this: Life (L) exists in the universe. Life could be the result of the work of a Designer (D) or it could be the result solely of chance (C). We can represent the alternatives as (L + D) and (L + C).
With the discovery of fine-tuning and its critical importance to life, a conjunction he designates (R), we can ask whether it'd be more probable or likely that R would be found if there is life and a designer or whether it'd be more probable that R would be the case given the existence of life and chance. Philosophers express such comparisons of probability (Pr) like this:
Pr (R / L + D)which is read: the probability of R on L and D which, translated, is "what is the likelihood that fine-tuning would exist and be necessary for life given that there is life and there is a designer?"
The question, then, is: between the two alternatives which has the higher probability? Is Pr (R / L + D) higher than Pr (R / L + C). The answer, of course, is that the exquisite precision of the cosmic fine-tuning being needed for life is clearly more probable on the assumption that it's the product of an intentional agent than that it's the product of blind chance. Thus:
Pr (R / L + D) > Pr (R / L + C)Since we should always believe what is more probable than what is less probable fine-tuning counts as strong evidence that there is a Designer. Of course, as Roberts observes, other factors might count against the claim that there is a Designer and must also be considered.
Parenthetically, I'd add that the factors that are often cited as counting against the existence of a Designer, e.g. evil and suffering, do not count against the existence of a Designer but rather count against the claim that the Designer is omnibenevolent. After all, a malevolent designer is still a designer.
Roberts goes on to explain how this version of the argument differs from earlier versions and how it avoids objections to earlier versions. Torley summarizes these nicely.
