The windings do determine alot... efficiency, speed, torque... and should be matched up as close as possible with the impeller. What you say about the impeller being the critical component, and the only thing that matters is PARTLY true.
Im afraid its a huge balancing act, and much too much to try to explain here 100% though (its a whole different thread). Lets just put it this way though... the reason I dont like the Mag is the same reason I dont like The Octopus pumps, and other cheaper makes. While you are somewhat correct that you can 'customize' the impeller to overcome the drawbacks of a pump, the truth is that you have to start out with a good pump in the first place. If the windings and pump design are not done properly, there is nothing you can do to overcome this with the impeller. There are limits that a poor motor will have. The problem with some cheaper motors isnt in how they match up to the impellers, but how they match up to the current coming from the wall.
Most European made pumps are of higher quality because underneath it all, they are 220v. But some of these american makes and chinese makes have skimped on things like copper to make the production costs cheaper. This is what causes the pumps to heat up more. See, eheims, hagens, aquabees... they have to be built to 220v tolerances, so running on 110v is a piece of cake even if designed cheaply by EU standards. This is part of the reason why aquabees are so great to us, yet to many euro-reefers, they have a reputation of a mag-drive in the US.
The idea of using a Mag is a good one, but the reason I wont do it is because I know that there is only so much that can be done with them before you hit their operational limit. Smaller windings = voltage not carried as well = more current used = more heat on smaller windings = even less efficiency (electricity wasted on heat rather than movement). From an engineers standpoint, the efficiency improvement between a Mag to an eheim may be going from 60 to 65%... as a good amount of the energy used on pumps is wasted as heat in the first place.... to an engineering team, that 5% loss in efficiency may be worth the 10% greater profit margin in the US given price of copper. In the US, the value more efficient motor hasnt been a huge issue until recent, and still isnt compared to say... Germany... so a US mfg might sacrifice that 5% efficiency, and an EU mfg might gain more from using the windings, making a better, more efficient motor because the double cost of electricity. The only way to make a Mag as efficient as an eheim or Laguna would be to limit its incoming voltage really (among a couple other things like pf correction circuits)... so it wont heat up as much, it will run more efficient then, etc.
My suggestion to ANYONE modding ANY pumps for needlewheels is to use a kill-a-watt or other device to measure power factor. You really want to match up the power factor of the modded pump with the original as best as possible... a realistic PF for a good pump is .80 < and so if you do your needlewheel mod and you get .40 (not uncommon), you want to increase (or you can increase) the diameter/thickness until the power factor is really as high as you can get it without stalling the pump on startup. This may be .80, or this may be .6 ... The power factor is important because it determines efficiency, and to a certain extent, will determine the longevity of the pump. If the power factor is poor (say, .25) the pump may run hotter than normal, can crud up your electrical lines with interference, etc. pf is after all, real power/apparent power. Apparent power is what you pay for at the meter, but real power is what the pump will actually be using. So even though the RMS reading may be 40 watts, the pump may really be using over 100 easily (the VA reading). This of course, will burn out the motor faster.
For example, on my crappy Aqua Euro 590 pumps, the stock pump itself has a PF of about .3 This is simply dismal. The worst pumps I had tested until this point were about a .75. This number is used to calculate pump efficiency in a rather big way...
Power Out of Motor = Sqrt(3)*V*I *pf *efficiency.
As you can see, the powerfactor is a major number here. A pump with a power factor of .3 will have half the efficiency of a pump that is .6... and most pumps with great power factors are still only at about 80% efficiency... the rest goes to heat. So that pump with a pf of .3 will run very very warm. Using a pump with a good power factor to begin with is an indication that it is a cooler running pump.
Now, if you have a pump that runs normally at .3, there is little you can do to get above that number no matter what you do with the impeller. Its all downhill from there...