Wind generators have strange requirements from their alternator that are somewhat difficult to accept. Some of the conventional machines as designed for other uses are not ideally suited to this application. In fact I am not aware of any machine design that even remotely suits this application of working at constant speed into an effectively fixed voltage.

Firstly radial machines are more common and easier to visualise and in most common radials the opposite sides of a coil do lie on alternate poles ( the coil span is one pole pitch). Now large commercial alternators have characteristics that are mostly not dominated by the winding resistance ( except that it determines the efficiency). The inductive reactance of the winding mainly determines the current it can produce, so having long end connections to the coils is of small consequence.

Now when we try to build the axial equivalent we start out with coils having radial sides again centred over one pole pitch ( sides of coils over adjacent magnets).

This ideally needs sector shaped magnets and this is the best way to build a conventional slotted cored single phase axial for normal use. Adapting to 3 phase just means 3 identical windings displaced equally.

Now put it on a windmill and get rid of the iron core and slots ( for simplicity of construction and also a side benefit of removing iron loss) and you have a different situation. There is no noticeable leakage reactance within the working range, the output is determined by the difference between the generated voltage( emf) and battery volts divided by the resistance.  For the maximum output you need the highest emf and the lowest resistance.

We are rectifying the output so waveform is not confined to a sine wave as it would be for a grid alternator ( I won't go into that because special steps have to be taken to get a sine wave in an iron cored alternator that will confuse this argument beyond recognition). As long as each magnet links all the turns of each coil the mean emf will be the same, the waveform will change drastically with geometry but let's not worry about that. We then reach the condition where reduction of winding resistance is far more important than worrying about the ideal coil placement for the highest possible square wave voltage which we should get with the iron core and coil sides over adjacent magnets. ( removing the iron changes the whole voltage wave shape).

We finally reach an empirical design where we use all the available magnet flux reasonably effectively with a coil shape that gives us the lowest resistance.

We further simplify the winding to keep a 3 phase scheme without overlapping coils for ease of construction. This requirement alone means that with the 12/9 magnet coil arrangement at best it would only be possible for the outside turns of a coil to be placed over the centres of adjacent magnets. This air gap type alternator always comes out as a distributed winding as it is impossible for all the turns to occupy the ideal position. This is mechanically true of iron cores but the flux snaps from tooth to tooth and links all the turns of a coil at the same time.

That is the theory out of the way, now forget it ans see what works best as an empirical design and we come up with something like this ( may be room for minor tweaking).

Space magnets so that gap is about magnet width, not precise as the gap changes with radius. Either go for centre line spacing or spacing at inner radius if you want a less compact machine with a bit more output. Make coil hole the shape of the magnet and the same size or a bit smaller ( with round magnets you can squeeze the width, with rectangular you can squeeze the bit near the centre egg shaped  if you want to get more wire in )

Wind the outer to use all available space ( coils will touch at one point).

Choose turns for correct cut in. Wind space with the thickest wire that will go in.

Make distance between magnet poles about 80% of two magnet thickness and make stator as thick as you can and still leave safe mechanical clearance.

If you are over enthusiastic with materials and the alternator is too powerful for the blades then add line resistance Reduces stall and keeps heat out of the stator).

I hope this makes sense, it is based more on practical experience than absolute theory.