Turns out that the revised criteria does not make it much easier to determine books to be represented in a shorter list. The fundamental issues that arise are:
1) Greek tradition to be so early that it could trump later texts to the extent that it is hardly balanced or representative
2) Specialization - there are not many texts or author that has broad "coverage" (across genres), with the clear exception of the Bible, and then Voltaire (he is both historian, thinker and literary author - though arguably he influenced his time more than his texts ended up being canonical), and maybe Cicero, Machiavelli and possibly Marx.
3) Polycentric nature of the "tradition" - you have these pairs like Herodotus vs. Thucydides, Plato vs. Aristotle, Augustine vs. Origen, Kant vs. Marx, Shakespeare vs. Cervantes -- that in a short list one has to just pick one and ignore the other.
The process I ended up trying to apply my criteria are convoluted, first do it by Greco-Roman, Christian, and modern Western separately; then do it by genres separately, and then list out the top ones no matter how one cuts it.
1. Bible - very easily stand out
2, 3. Plato, Herodotus - easy to pick out if one can accept that Thucydides and Aristotle can't be all be represented in a short-list
4. Augustine - also quite clear if there are more than just the Bible representing Christianity. Also the only clear Latin texts.
All these apparently does not cut into the modern Western tradition. In the "balance" of the scheme, in this go-around, I am still going for a 3-3-3 scheme for Greco-Roman, Christian and Modern Western "periods" What I end up having is:
5. Homer (instead of Virgil)
6. Dante (the 3rd texts of Christianity)
7. Shakespeare (English, 17th c.))
8. Marx (instead of Kant, German, 19th c.)
9. Voltaire (instead of Cervantes; it is in a way a choice between French or Spanish; and Voltaire is the only cross-genre polymath around - though any of his texts are likely to be less canonical than Cervantes, or Virgil, or Kant, or Thucydides or Aristotle, for that matter. This is a very hard choice; 18th c.)
I am sure the next time I play with this I will have a different list.
As I read through this pass of text of 25, I feel that many of the criticism I have for the list of 36 continues - for example the focus on Epics - and a main reason is the reputation of excellence that is built in to the current criteria - that makes a truly spreading out to more minor traditions and other types of literature more challenging.