wasn’t the worry about Vs that if derived from the performance stalling manoeuvre then the load factor i.e. ratio of lift to weight could be significantly less than unity. In which case the multiplying factor would give illusory rather than desired margins. So the Vsr technique could be said to give “proper” V2 speeds rather than “lower”. [1-g stall speed as the basis for compliance… refers]
The desire to use 1'g' stall speeds instead of minimum stall speeds arises from a desire for repeatable speeds, not a desire to make V2 (or any other stall speed related operational speed 'correct')
In fact, the intent of the reduced reference speeds is to achieve a situation of 'equivalent safety'.
Broadly, V2min=1.2Vs has been shown, empirically, to provide a sound basis for a minimum takeoff safety speed.
Vsr (1g stall speed) is going to come out higher than Vs (since the load factor at Vs is <1.0). In order to avoid penalising more modern designs for taking a more scientific approach to the determination of stall speed, the V2min requirement for a 'Vsr' aircraft changes to 1.13Vsr, precisely so that the same V2min will be arrived at whether one uses the Vs or Vsr methodology. In other words, a 'Vsr' aircraft has precisely the same margin between V2min and the 'real stall' as a 'Vsmin' aircraft.
(Which implies that for a typical aircraft, one would expect that Vsr=1.13/1.2=0.94Vs)