R-corr AMPT comparison and problems in R3 definition

 Roy's modification is just a constraint to Psi3 range (0<Psi3<2pi/3) and phi range (Psi3<phi<Psi3+2pi).

 
The constraint to Psi3 by definition won't change anything,
because it at most makes DS into -DS for the whole event while DS distribution is already forced to be even.
 
The constraint to phi is just an arbitrary choice, so physical observables should not be affected at all.
However, DS3 has dependence on this nonphysical choice of phi range,
indicating DS3 as well as R3 are ill-defined and not physical!.
Then, I think one cannot draw any physical insights from R3, and the so-called R3 scaling is nonsense!
 
Details can be found on page 5 of the attached file.