ee -> jjjj (RS) broken

[Enrico Bothmann]

Running Examples/Jets_at_LeptonColliders/LEP_Jets/Run.NLO.dat on rel-2-2-4 produces a giant cross section for ee -> jjjj (RS). Also the distributions are very awkward.

One can use the threshold_initialisation.patch for #375, as a workaround for the #375 crash.

The bug is quite old. If first occurs on trunk with r25928 ("check against kinematic and flavor mismatch in r jetcrit"), i.e. between the 2.1.1 and the 2.2.0 release. Note that the r25928 revision will only start integrating after applying r26025, which fixes a bug in the internal DY virtual ME.

Here are the integration results, note the 4.50949e+06 pb for the 2_4__e-__e+__j__j__j__j__QCD(RS) channel:

Read in channels from directory : Results/Amegic/MC_2_2__e-__e+__j__j__QCD(BVI)
Process_Group::CalculateTotalXSec(): Calculate xs for '2_2__e-__e+__j__j__QCD(BVI)' (Amegic)
2_2__e-__e+__j__j__QCD(BVI) : 31552.3 pb +- ( 232.279 pb = 0.736173 % )  exp. eff: 9.42397 %
Read in channels from directory : Results/Comix/MC_2_3__e-__e+__j__j__j__QCD(RS)
Process_Group::CalculateTotalXSec(): Calculate xs for '2_3__e-__e+__j__j__j__QCD(RS)' (Comix)
2_3__e-__e+__j__j__j__QCD(RS) : -373.379 pb +- ( 9.09491 pb = 2.43584 % )  exp. eff: 1.1781 %
  reduce max for 2_2__e-__e+__j__j__QCD(BVI) to 0.179983 ( eps = 0.001 )
  reduce max for 2_3__e-__e+__j__j__j__QCD(RS) to 0.515557 ( eps = 0.001 )
Read in channels from directory : Results/Amegic/MC_2_3__e-__e+__j__j__j__QCD(BVI)
Process_Group::CalculateTotalXSec(): Calculate xs for '2_3__e-__e+__j__j__j__QCD(BVI)' (Amegic)
2_3__e-__e+__j__j__j__QCD(BVI) : 11399.2 pb +- ( 518.311 pb = 4.54692 % )  exp. eff: 1.59119 %
Read in channels from directory : Results/Comix/MC_2_4__e-__e+__j__j__j__j__QCD(RS)
Process_Group::CalculateTotalXSec(): Calculate xs for '2_4__e-__e+__j__j__j__j__QCD(RS)' (Comix)
2_4__e-__e+__j__j__j__j__QCD(RS) : 4.50949e+06 pb +- ( 438975 pb = 9.73448 % )  exp. eff: 0.01594 %
  reduce max for 2_3__e-__e+__j__j__j__QCD(BVI) to 0.322027 ( eps = 0.001 )
  reduce max for 2_4__e-__e+__j__j__j__j__QCD(RS) to 0.894462 ( eps = 0.001 )
Read in channels from directory : Results/Comix/MC_2_4__e-__e+__j__j__j__j
Process_Group::CalculateTotalXSec(): Calculate xs for '2_4__e-__e+__j__j__j__j' (Comix)
2_4__e-__e+__j__j__j__j : 1704.02 pb +- ( 37.0925 pb = 2.17676 % )  exp. eff: 0.561434 %
  reduce max for 2_4__e-__e+__j__j__j__j to 0.618604 ( eps = 0.001 )

[Enrico Bothmann]

Reverting r25928 on rel-2-2-4 allows to generate reasonable results also for distributions (checked for ALEPH_2004_S5765862).

[Enrico Bothmann]

This is fixed on rel-2-2-4 with r31010.

Note that this might not only have affected lepton colliders, but also hadronic set-ups and the resulting physics results, see the change diff of r31010 and its log message.

Note that on trunk the affected logic had been rewritten in the meantime, such that #377 is not an issue there.

Thanks to Stefan for helping me fix this!