MadGraph5_aMC@NLO

Trouble generating correct # of events for p p > h > ta+ ta-, ta+ > l+ vl vt~, ta- > l- vl~ vt

Asked by Roman Marcarelli

I have recently been trying to generate these events after importing the model "heft":

p p > h > ta+ ta-, ta+ > l+ vl vt~, ta- > l- vl~ vt

If I try to generate, say, 1000 events, Madgraph says "Need to Improve [n] channels," (n varies, usually 1-2) and the number of events generated ends up becoming 1. I have tried to set the event generation up in various different ways, i.e.:

p p > ta+ ta- /a z, ta+ > l+ vl vt~, ta- > l- vl~ vt
p p > ta+ ta- > l+ l- vl vl~ vt vt~ /a z
p p > h, h > ta+ ta- > l+ l- vl vl~ vt vt~

etc. I have also tried narrowing down the issue. Both

p p > h > ta+ ta-

and

p p > h > ta+ ta-, ta+ > l+ vl vt~

generate the requested number of events, but once I try to decay both taus, I run into a problem. I suspect it may have to do with a collinear divergence from the neutrinos, but I am unsure how to deal with this (I've tried placing cuts on missetmax to no avail). Please let me know if you have any recommendations on how to generate this process successfully.

Best,

Roman

Question information

Language:
English Edit question
Status:
Answered
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Last query:
Last reply:
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#1

Hi,

What is the width for the tau?
If it is zero then your matrix-element is divergent and it is impossible to generate events,

Cheers,

Olivier

Revision history for this message
Roman Marcarelli (rmarcarelli) said : 		#2

Hi Oliver,

The width of the tau in the param card is 1.777e+00.

Best,

Roman

Revision history for this message
Roman Marcarelli (rmarcarelli) said : 		#3

Hi Oliver,

Actually, that is the mass, but I see that there's also "INFORMATION FOR DECAY," and there it looks like the tau width is set to zero next to "DECAY 15." I tried setting this value to something other than zero and received an error:

WARNING: Particle 15 with small width detected (1.426e-11): See https://answers.launchpad.net/mg5amcnlo/+faq/3053 to learn the special handling of that case
INFO: Update the dependent parameter of the param_card.dat
WARNING: Failed to update dependent parameter. This might create trouble for external program (like MadSpin/shower/...)
Generating 1000 events with run name run_01
survey run_01
INFO: compile directory
 ____________________________
( Really, this sounds fishy. )
 ----------------------------
  o [Art by K. Gotanda]
    o
       o

     ><(((('> ><(((('> ><(((('> ><(((('> ><(((('>

Command "generate_events run_01" interrupted with error:
InvalidParamCard : parameter decay: 15 is not at zero
Please report this bug on https://bugs.launchpad.net/mg5amcnlo

Revision history for this message
Roman Marcarelli (rmarcarelli) said : 		#4

Also, I don't know if it's worth noting, but the process

p p > z > ta+ ta-, ta+ > l+ vl vt~, ta- > l- vl~ vt

does generate the requested number of events. Sorry for the barrage of comments.

Best,

Roman

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#5

Hi,

If the width of the tau is zero, then the tau is stable and cannot decay.
The associate matrix-element
p p > ta+ ta-, ta+ > l+ vl vt~, ta- > l- vl~ vt
is actually divergent and therefore in top of the cross-section to be infinite for such process, it is impossible to generate events.

Long story short, you have to use a model which does not assume the width of the tau to be zero (most UFO model does)
This is why you computation crashes since your UFO model assumes zero width for that particle.

The second issue with tau decay is that the width of the tau is so small that you have issue with numerical accuracy of the code.
This is why we have some special handling for very small width. The special handling is explained here: https://answers.launchpad.net/mg5amcnlo/+faq/3053

The third issue of the tau decay is off course all the hadronic decay where perturbative QCD is not valid.
Some collaborator have written an model extension for that (add model tau_decay). You can see the paper here: 1212.6247

Cheers,

Olivier

Can you help with this problem?

Provide an answer of your own, or ask Roman Marcarelli for more information if necessary.

To post a message you must log in.