but because the secret of overlapping coils has eluded you i can't give you the gold star.
all things being equal if i can spread the number of turns required per phase to reach a charging voltage over 8 coils rather than 2 coils, i can use a thicker conductor.
the issue Adriaan is talking about has nothing to do with what you're talking about.
if you want to solve which alternator is most efficient, i have explained this before:
make your coils, and short them out.
apply a constant torque to the machine and measure the rpm it spins at. slower is better. you have three variables to figure out:
air gap width, inside diameter of the coils, number of coils. (the number of coils sets the outer dimensions of the coil, because if you have any gap between the coils, if that were really optima then you should push the magnets closer to each other to reduce the outer diameter of the machine. so its assumed the coils are pushed together as close as you can get them, or you wind as many turns as you can fit, or you increase the diameter of the copper so the number of turns you needs pushes the outer diameter of the coil as large as you can get it. )
the actual voltage and current is totally arbitrary, depends on the number of turns, and has nothing to do with efficiency. "i can use thicker wire" yes you should use the thickest wire you can get for the least number of turns (voltage) you need. you're fundamentally limited by the mass of copper you can fit between the two magnet plates and that has only to do with (for single layer windings) the inside diameter of the coils. if you make the id too small, yes you get more volts but your resistance increased disproportionately more. same goes for the thickness of the coil. it is of no advantage to make the coils thicker than is optimal, because the stiffness of the machine is proportional to volume of effective copper multiplied by flux squared. as the air gap increases sure you get more copper, but the reduction in flux matters more.
in the case of a 9 coil 8 pole generator, it is less efficient than 6 coils because the coils are no longer optimal, electrically they aren't as wide and each coil doesn't even produce a sine wave, as Adriaan said they are out of phase, so their voltages don't even add up in phase, you loose some voltage.
for example, my 24 slot, 10 pitch 2 pole generator. each of the 8 coils generates 145 volts. connected wye, they sum to 240, not 1.73*145. it does however produce a very nice looking sine wave. this nice sine wave which comes at a cost of 10 or 15 missing volts is of no use if i'm directly rectifying the voltage to dc.
you want the messiest, square or trapezoidal wave you can get out of a machine if you're directly rectifying it into a battery, (unless the harmonics cause eddy current losses in the magnets, not a concern for anyone on this board) and it may in fact be true that for single layer windings the 8 pole 6 coil is the most efficient.
once you go to a two layer winding the end turns get longer because you have to fit the copper around each other. but you can approximately double the amount of copper (by weight). you probably won't get it electrically twice as stiff, because the end turns are longer.
note that because air core alternators have no core loss, if we ignore eddy current losses we can only specify the stiffness of the machine. the efficiency drops from 100 percent (ignoring bearing and windage losses) to about 50% at reasonable loads, perhaps even lower.