On a four-engined regional airliner I was involved with, we measured VMCG like this:
Usually light weight and aft CG.
Shut down (cut to fuel off) at progressively lower speeds until the deviation just exceeded 30ft, sometimes a cliff edge would be reached and the deviation would suddenly be very large. The aircraft had a VMCG that went up as the weight came down, due to tyre footprint effects we believed.
We did not disengage the nosewheel steering in case it was needed in a hurry. The rudder was not connected to the nosewheel. We allowed “instinctive” use of aileron. We always briefed that in the case of a large deviation we would throttle back the opposite outboard (the benefit of a four-engined aeroplane)
We had the luxury of lines painted the runway 30ft either side of the centreline and would offset the aircraft one side or the other, depending on the crosswind as we knew which engine was going to be cut. The camber wasn’t significant.
The deviation was determined by use of the painted lines and the fact that on this particular type the distance from the centre to the outboard edge of the flaps was 28ft. Latterly differential GPS was used.
Hope this helps.