Then there is the case of low power on an engine and the pilot nursing the engine to get it home is actually creating more drag that it would when its feathered which thus increases the Vmca .
My turbine twin the simulated torque for feathered is 10-12% with the flight idle 4%. I should imagine a petrol engine would be quite a bit more for simulated feathered?
There have been loads of crashes due to people nursing engines below 10% not shutting them down.