The arithmetic is quite clear - the most accurate calibration numbers will be obtained by calibrating always at the meridian and the equator and then using DEC compensation.
I've read Chris's book and the only thing that I recall is that he seems to think that if you decide to image something at DEC 80 or so, you might want to turn off DEC guiding altogether. I don't recall anything in the book that says it's inherntly better to calibrate at DEC 50 rather than DEC 0.
Thinking about it, though, if you were imaging from the Congo, DEC 0 is right on the horizon and that would make me vary my workflow. That's a good edge case! Some people are imaging these days from near the equator.
I may be dense here, but at the north pole polaris is directly overhead isn't it? And Polaris is at dec 90? Then wouldn't you have to be at the pole (N or S) to have dec zero at the horizon?
Wouldn't dec 0 pass from east to west through the zenith at the equator? And wouldn't Dec 90 then be at the horizon?