Thanks to some feedback from Queen Quon, I just realized some things I did wrong.  Here's what was wrong with that analysis:
1) Insufficient data.
2) Wildly variant data.
3) Unknown optimization algorithm.
4) Model and constraints to limit the effects of #1 and #2.  (limited X1-X4 and C to being between 0 and 1.)

I think that: (1) I should have allowed X1 through X4 to exceed 1.  (2) C should have been allowed to be negative, and certainly greater in magnitude than 1.

(The first time I ran it, some of the coefficients came back *negative* whichs makes no sense---yes, I like the texture less, so I'll rate the cookie higher.  So that's the only constraint I should have had: X1-X4 >=0 )

I'll run some calc's again.  But I think the irrational #2 will result in wacky results.

Anyway, I did just make up a new cookie, last night..."asteroids".  Yummm...

I'm not sure the "averaging" method here is a logical idea.  In some ways, I think it's the antithesis of "geeky"...a scientifically minded person wouldn't just average numbers and think that will work out well.

But still, I would like to try to make these "mean chocolate chip cookies."