I've considered the question about low wind startup and high wind efficiency.
The losses in the windings are I^2 * R, so if I is minimized, the efficiency will be maximized.
To minimize I, run at a high voltage output (100V or higher) and use a buck-type switching converter (after rectification and filtering) to deliver the voltage to the batteries. At 100V, 10A=1 kW and 50A=5 kW. Using polyimide (220 deg celsius) rated [magnet] wire, 50A could easily be run with 10 AWG windings.
If the switching converter is operated with an appropriatelly programmed microcontroller sensing input voltage and current at the converter input, maximum power could be extracted from the wind generator in any wind condition, as the generator would see its optimum load at all times. This is identical to maximum power point tracking for solar arrays (e.g. Solar Boost(tm).) The 100V at 10A (for example) would result in ~14V at 64A into the batteries (with 100W dissipated in the converter, assuming 90% converter efficiency.) That same microcontroller could also switch the output (or input) to a diversion load when the batteries were full.
It could also have a charging algorithm for the batteries (bulk, absorption, float [and equalize if required] with output voltage sensing.) Since it would know the input power and converter efficiency, it would know the output current given the output voltage. In this case the diversion load handling must be thought out carefully, and the system load current monitored. Probably switching a diversion load onto the battery bank would be the best way to control the battery charging in this case (and a second converter could be used to vary the diversion load too.)