Factor of 1/2 mismatch (moller cross-section) between Griffiths and CalcHEP

All coefficients in CalcHEP are right. But for processes with identical
outgoing particles CalcHEP does not present diagrams
which appears after permutation of outgoing particles. In numerical
session symmetritization is performed at last step numerically.
For symbolic session we also need symmetrization at last step. If you
use MATHEMATICA, then you have to load first
  sum_22.m and after that load symbolic results in math format.
Finally variable "sum" should contain answer.

Write me again if you will not have agreement. I'll try to explain better.

Best
    Alexander Pukhov

On 07/18/2017 05:14 PM, N_ wrote:
> New question #649696 on CalcHEP:
> https://answers.launchpad.net/calchep/+question/649696
>
> I just started using CalcHEP, and to test it I was evaluating cross-sections (in the relativistic limit m_i --> 0) for standard QED processes. For the Moller process, the CalcHEP answer is half that of Griffiths Elementary particles (Prob. 7.34):
> CalcHEP answer: 2 EE^4 (s^2 + st +t^2)^2 /( t^2 (s+t)^2)
> Griffiths answer (after substituting in the Mandelstam variables): 4 EE^4 (s^2 + st +t^2)^2 /( t^2 (s+t)^2)
>
> As the CalcHEP code seemed to work fine for a Bhabha scattering, I suspect some issue with spin averaging (over 2 fermions vs. over 4 fermions). What do you recommend?
>