I think that this is a bit too oversimplified to take seriously, but some
interesting points are raised.  In any case my vote is:

   historical 3%
   musical value  75 %
   composer 21 %
   performer 1 %

The problem is- what to make of Bill's  hypothetical high school student (
or computer program for that matter) who writes in a Mozartian style and
does it very well? This came up about a year ago, and my stance was and
remains that such efforts are to music what forgeries are to art. I assume
that a previously unknown Van Gogh would be subject to the sternest
scrutiny and would be rejected if it was determined to be a forgery, even
if it was a GREAT forgery. So musical value and composer are categories
which are hard to separate out. At the time that this was discussed others
energetically defended an opposing point of view.

(Who will drop the "professor" bit from now on because he is not a
professor of anything which gives him expertise in music.)

Professor Bernard Chasan
Physics Department, Boston University