It's the same thing with practicing landings. We still have accidents  

Predictable outcome in an airplane is easier because you are working with a stable platform.  Divergence will exist after a control input but if the condition is neutralized the tendency is to return to last known point of stability.  In a helicopter you are always in a state of  divergence.  The pilot for the most part is the known point of stability without autopilots or stability augmentation systems. 

I don't know if the rotor blades could get to the tailboom in a true state of VRS because the rotor system would be in a positive G configuration.  Extreme aft cyclic and unloading the rotor system at the same time might cause it but that would a totally out of character maneuver for VRS recovery.