I'd also like to add that shorting a winding on a transformer turns it into a very small value inductor- on the order of say 10 microhenries, with some resistance dependent on winding resistance, say 0.1 to 0.5 ohm. This inductance is too small to have any effect on a 50/60 Hz waveform. The resistance will not help as Listerflicker is not a load related phenomina. This will NOT help with Listerflicker, Starfire was mistaken.
Listerflicker is caused by the variation in speed of the engine during compression stroke and power stroke; it results in voltage and frequency drop during the former and rise during the latter. A fast responding AVR with plenty of source power can partially compensate the voltage change, but can't do a thing about the frequency variation.
I'm quite familiar with the effects of shorting of a winding per Starfire's suggestion since this is required in my 2 transformer, 5 step sine inverter. This uses the two transformer's secondary windings in series per the old Trace SW series inverter. Without shorting the primary, the huge inductance of the transformer secondary (half a Henry) will just filter the secondary output of the other transformer secondary, effectively eating it.