The key bit for me is the comments on writing good GM advice. Without it, a "logic-based" game is nothing but a designer cop out.
That advice has to do at least two things:
- Talk about all the possible bases of logic for the game in question (e.g. the realism, cinematic, particular story, whatever this game is intended to support) - and then guide the GM and players in how to pick a mix and stick with it.
- Now having a particular base or mix of bases, give good advice on how to apply the more flexible rules to this logic.
Then on top of that, I don't see how you really teach that in a book without many, extensive examples. (It's also possible the some video presentations or other media outside the books might also be helpful--as an alternate to having someone experienced teach it directly.)
Finally, the designers and developers of the game must attempt to mitigate their own illogical biases when producing all this. No small order, that. Sometimes they don't see the need for more advice or examples or instructions--because they've got some assumed "logic" of how it must work enshrined in the rules, and it really doesn't make much sense. At the very least, such embedded logic needs to be ruthlessly deconstructed enough to explain it in the design notes.
This is why I keep echoing a famed VB reviewer: "More examples, damn it!"
Having the examples is important to convey how the game works.
Writing the examples is important to make sure the game does what the designers thinks it does.