As long as we don't really know what causes reversals, that's still okay. It has been recognised that variations in the current period have a periodicity of 10 and 100k years. Apparently the field did not flip even during the variations. Yet is the internal motor of the magnetic field actually stable - who knows?

I -think- (danger Will Robinson danger!!!) it could work like flipping a coin: every time the dynamic destabilises it has a 50% chance of actually uprighting itself again (non-reversal) or the field turns over completely. If so, in the current period the field has destabilised previously but just never reversed.