For item #1 anyway...
It is not unlikely that the "really good" Pastels are actually expressing more than one gene. That would explain pretty easily why some pastels (heterozygotes) look better or worse than others, but also why even the "bad" pastels can still combine with other pastels to make the Super pastel (homozygote).
What I am theorizing is that there is one gene that is most easily recognizable as "pastel," and that there may be another gene, or a few other genes that contribute to enhancing that "pastel" gene in the overall phenotype. This could even be something of a linkage issue, where depending upon how closely any of these genes are linked, they are mostly inherited together, but can occasionally be inherited seperately.
I love genetics - the snakes got me hooked on genetics at a young age... and I am [an older] college student taking a few genetics-specific courses, so I know I am glossing over some technical and complicated stuff. Let me know if I should explain this more - or if I should just keep my silly theories to myself. 
~Rebecca
-----
1.1 Ball Pythons (1.0 '05 Ghost, 0.1 '03 Normal)
0.1 Dumeril's Boa '04
0.2 American Pit Bull Terriers (40 lb darling lap dogs)