Stability
Crab
I can't publish a drawing right now, but the point I have trying to make (difficult apparently without a drawing) is that dynamic (un)stability depends mostly on geometry.
Picture a heli on the ground. Look at the lateral angle the rotor can make and draw a line between the hub and the pivotal skid. If the rotor can tilt enough to get an intersection outside the skid, then the cyclic would be able to redress the heli. In the case of commercial heli's due to the static stability requirement this will not be the case (perhaps a sky-crane?)
I think any normal pilot will try to avoid tipping over by applying cyclic, due to surprise perhaps not full cyclic, but at least quite a bit. I would start analyzing at this point.
Keeping the cyclic were it is, allows you to draw the trust vector. The perpendicular to this trust vector from the skid gives the arm (if cyclic is applied this will be less than half width)
Now you will see that exactly for the same reasons width gives static stability, it creates dynamic unstability, because the rotor trust also gets a greater arm to tip the heli. So smaller retaining forces on the skid will be able to make the make heli flip.
A wider heli has however the advantage that it will need to lift higher before tipping over, mathematically this is equivalent to stating that the inertia tensor for roll around the skid will be much larger than the tensor around the COG. This will slow down angular acceleration and give the pilot more time to lower the collective, so at this point I should nuanciate my previous statement.
d3