Deterministic within the stochastic...selection that is

The answers keep coming in for my query about "what is evolution?" RPM took me to task for basically answering "what is selection" with my initial response. This is a good criticism...honestly, I wanted to focus on selection because I think random genetic drift confuses many people, and it quickly turns into a black box incantation that explains everything and nothing. But here's another point of interest: selection is stochastic as well! That is, selection favors fitness, and fitness tends to exhibit particular characteristics in particular environments (e.g., extremely fast vertebrate aquatic creatures seem to converge upon similar forms because of the constraints of physics), but at least in the short term there are often many solutions to the same problem. You want an example? High altitude peoples, different genetic and physiological responses. Now, I did say "short term," because I am to understand that the Tibetans, for example, seem to have "better" strategies in regards to optimizing fitness than the peoples of the Andes. Why might this be? Well, Tibet and (or its environs) have been inhabited by modern humans for at least 50,000 years. In contrast, the Andes probably hasn't seem human habitation for more than 12,000 years. Selection takes time. And, selection has to work with the variation it has on hand, and Tibet is demographically much more "hooked in" than the Andes, which had to replenish selection after the Beringian bottlneck. Of course, the Andeans have hit up a local fitness optimum, but given enough time it seems possible that would they keep ascending toward the Tibetan strategies (which are less stressful on the body). Or would they? Perhaps they would hit upon modifier genes and wander down an alternative developmental path from the Tibetans altogether...cycles within cycles, and so it never ends. My overall point is that sometimes selection simply reduces the sample space from nearly infinite opportunies to explore genetic variants (e.g., neutrality) to only having recourse to dozens of strategies. How does selection select from those strategies? God knows....