#166 closed defect (fixed)
Gauge invariance violation in Comix
| Reported by: | Frank Krauss | Owned by: | Stefan Hoeche |
|---|---|---|---|
| Priority: | minor | Milestone: | |
| Component: | Unknown | Version: | 0.trunk |
| Keywords: | Cc: |
Description
In ttbar production with decays, Comix complains about a violation of gauge invariance in the sub-permille region. I suspect this is due to the fixed width scheme.
Amplitude::GaugeTest(): Large deviation {
1.07183021878e-11
vs 1.07217719645e-11 => -0.000323619702526
}
.
I would suggest to suppress this output, to keep users quite.
Alternatively one could think about using the CM scheme.
Attachments (1)
Change History (4)
Changed 14 years ago by
| Attachment: | Run.LHCB_T_Asym.tt.HL.dat added |
|---|
comment:1 Changed 14 years ago by
comment:2 Changed 14 years ago by
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
Just tried to initialize the process without explicit decay specification. No gauge test violation there, neither for fixed width scheme, nor for complex mass scheme. The case with decays also works if the W-width is set to zero. So, isn't this simply due to the fact that W's/Z's are not allowed as on-shell final states once they have a width (and are therefore considered unstable)? The narrow-width approximation (i.e. our decay treatment) does not respect this fact. Therefore, I can see no reason to avoid the error message, users should rather be aware of the problem.
Comix is actually using the CM scheme, but of course the widths are not set to their tree-level values, which might be the source of the problem. I would like to keep this warning to be alert about possible issues, otherwise what would be the point of the gauge test? Note that Amegic produces error message too.