Super form. It's too early to know now if one "kind" of pastel makes a better super, but if one does, then I'd think it would be worth more dough. I have definately seen some mediocre, as well as some knockout supers!! So when I think of the money-making abilities of pastels, I think of combos and especially the supers. Pastels are pretty in their own right, but being co-dom and relatively cheap, enterprising folks are thinking about the combos and the supers, and the super combos instead of what "quality" pastels they can make.
I don't think that right now there has been enough breeding to know if there are different flavors of pastel that should be worth more. What about the partner they are paired with?? What about variability within clutches?? We need more time to see what produces what with more consistency.
I have a pastel male who definately isn't the best pastel. I've heard that using a clean-looking normal female makes a difference in the way the offspring will look. We've all seen the pastel ghost, the bumble bee and the killer bee. Will it matter much in those combos wether you used a browned-out pastel, a lemon pastel, or a blonde pastel??? I don't think we can know that. . . yet. One or two or ten animals that show a trait (or traits) is just not enough to make that judgement IMHO!
Theoretically, couldn't you take the ugliest pastel in the world and breed out the "ugly" by being picky with it's mates and it's offspring's mates and it's offspring's offspring's mates? We shall see. Right now, I would not pick a lemon or a blonde over a regular pastel, just as I think it's crazy to pick a high-white pied over a low white when it's known to be a veriable trait. This is only my opinion of course. Maybe getting yourself a blonde or a lemon is only saving you the trouble of breeding out the "ugly?" It seems to me ball breeding is really just in it's infancy. Look at cornsnakes as a comparison. Their selective breeding has gone from breeding for simple recessive traits to breeding certain individuals for subgroups within those traits (eg. amel: sunglow, reverse okeetee, candy cane, etc.) Will balls get there some day? With smaller clutch sizes it will be longer before we find out, but I believe it's a good model to look at.
For example, I have noticed the mustache trait carries in several SK line axanthics and axanthic hets. Then there is the "belly line" suspected to link pied hets. My 03 normal pastel sibling has some pretty damn light eyes. There are a huge number of poissibilities in what subtle traits can be inherited and what can not- "blushing", black backs, "alien heads", bands, eye stripes that stop or eye stripes that fade onto the chin, etc. Some day hopefully we'll have the ability to know consistently what is reproduceable and what is not. Maybe then I'll hold lemons or blondes over normal pastels, but for now to me they are all pastels with a hundred thousand possibilities in store for them.