I'm not sure of the reasons, exactly, but I do know that there are certain relationships between the numbers of magnets and the numbers of stator poles in a motor (brushless, at least) that are more effective than other combinations, and IIRC it depends on the application as to whcih ones might be more effective. (torque vs speed, for instance).
In a brushed motor, it is going to always be an even number, AFAIK, so that you might have say, 16 windings on the armature with 32 commutator segments (one pair for each winding, with one end of each winding on each segment of a pair), and either two or four magnets, depending on whether you've wired the windings as two-pole or four-pole. (I also think there are even higher numbers of poles used in certain applications, but I've never actually seen one in person; I suspect they are not common).
The higher number of poles in a brushed motor, the slower it spins, compared to the same size motor with same armature but wound with fewer poles. But it has higher direct torque, in a direct linear relationship IIRC. So a two-pole motor will spin at say, 3200RPM, while an otherwise identical four-pole will spin at only 1600RPM but it will have twice the direct torque capacity of the two-pole. This is something I saw in experiment on CB2, when I changed from the 2-pole powerchair motor to the 4-pole. Now, I'm not totally sure about the exact relationship of the two, as it's been a while since I did any of those numbers (and I don't get along with numbers too well
).
I probably haven't answered your question very well (if at all) though.
