>No new techniques, bare minimum changes to require compiling a new crack, and just enough to make it difficult to automate.

From a practical perspective you also have to have a steady stream of features for the newer versions to be worth cracking. Otherwise why use v1.09 when v1.01 works fine? Moreover spending less effort into improving the DRM is still playing at the cat and mouse game, albeit with less time investment. If you're making minimal changes, the cracker also has to spend minimal time updating the crack.