Another alternative was proposed by Hawking & Hertog (Phys Rev D, 2006, see below): walk the present values surface back to a superposition of possible universe

Another alternative was proposed by Hawking & Hertog (Phys Rev D, 2006, see below): walk the present values surface back to a superposition of possible universes and study the mechanisms that generate probabilities near one. That then is their way of asking about the degree of fine-tuning, with the anthropic principle being recasta as a constraint on the set of possible consistent histories. H&H spend a lot of time comparing themselves to the approach of selecting an initial values surface in the early universe and marching it forward; they more or less do the same but flip the direction of the arrow of time and arrive at a superposed state. Re-flipping the arrow of time leads to relic fields like the CMB carrying evidence for the superposition. Essentially this is just taking time-reversibility of local microscopic physics seriously (great), and absorbing the difference in degrees of freedom in macroscopic physics into a superposition (uh, ok). The problem of setting up a values surface remains pretty much the same. "Fine tuning" is just a way of saying that the system is sensitive to the values it evolves from (or to). I'm not really convinced that that a present-day spacelike hypersurface is easier to write down than an early-time initial values surface. ArXiv version: http://arxiv.org/abs/hep-th/0602091 http://arxiv.org/abs/hep-th/0602091