I tried a similar experiment - but I was getting near perfect sine waves every time. In my case the magnets were touching...
also though - in my case I tested the "whole" coil...never thought to simplify it like you did.
My guess, from your experiments - is ideally (with space between the magnets) you'd want the hole the same size, or slightly larger than the magnet. But - since coils normally have many windings, I don't think this is so much determining the size of the hole in the center of the coil, but rather the "average" diameter. If you use that as your "average" diameter, then the hole should obviously be slightly smaller than the magnet and the outer dia of the coil slightly larger.
My newest alternator (this weeks project) has holes somewhat smaller than the magnets diameter, and the OD of the coils is larger than the magnets - so that one coil occupies the space of 1 and 1/3 magnets (so 3 coils fit over 4 magnets). I've tested this with a couple rotors, one where the magnets (1" dia) were touching, and then one where the magnets (1.8" diameter) have about 1/2" between them. In both cases it yeilded the most power to actually run coils larger than the magnets and have fewer coils than magnets.
I think it should be pointed out that ... the rules change with different types of rotor/stator setups. With slotted laminates and very thin airgaps things would probably change a bit over using flat laminates and a fairly thick airgap. Dual rotors might change things some too.