hi, yes i agree that we need to be practical: choose a way to pronounce (whether devine and stephens, modern grk style, etc), work out how to do it (if you need to: i needed to; for a modern grk using the modern grk way, this would be unnecessary), and tell people which system you use when you talk about/use grk pronunciation.
the devine and stephens book is not adapted for "using" though (and reading it is like trying to push your forehead through a brick without a run-up), so i tried to turn it into a useable technique, which is definitely not perfect but it is how i do it.
following my pitch model doc, your e.g. would have a drop to the second syll as you noted, then would rise steadily to the end. if that continuous rise sounds strange, there are at least four possible reasons:
(a) devine and stephens' book doesn't get grk pronunciation right,
(b) i didn't get devine and stephens' book right when making my model,
(c) the author of your e.g. did not write natural-sounding grk in stacking so many grave-accented words together; in fact from my reading of Attic prose I am not used to hearing so many graves together except in e.g. definitions given by Aristotle in his logical texts, e.g. the definition of á½„Î½Î¿Î¼Î±
http://www.tlg.uci.edu/demo/browser?uid ... nicode_All
or (d) the continuous rise is normal in grk but it sounds strange if you are not used to this in your native language (incidentally i have had to study french phonetics books on intonation, as well as absorbing what i hear around me, to start to change my anglo-saxon intonation patterns when speaking french).