MadAnalysis 5

Invisible particles & pythia

Asked by Markus Ebert

Hi,

I have some problems implementing a model with a new invisible particle.
On parton level, everything works fine, but after running pythia/pgs, there was no missing energy in my events.
I found some workarounds, but I am not very convinced about them, so please tell me if my procedure is ok.

The first problem I encountered was that pythia set the particle ID always to zero, which I fixed as described in https://answers.launchpad.net/madgraph5/+question/222285.
Afterwards, however, the pythia.lhe-file contained every particle twice. An example:
<event>
 11 100 0.10000000E+01 0.00000000E+00 0.00000000E+00 0.00000000E+00
      21 -1 0 0 0 0 -0.10714570E+03 -0.76306745E+02 0.20287612E+04 0.20330211E+04 0.00000000E+00 0. 0.
      21 -1 0 0 0 0 -0.26772322E+02 -0.24043248E+01 -0.18396642E+04 0.18398606E+04 0.00000000E+00 0. 0.
 9000006 2 1 1 0 0 -0.25757382E+03 -0.87222907E+03 -0.59084362E+03 0.11942462E+04 0.50000000E+03 0. 0.
-9000006 2 1 1 0 0 0.38392038E+03 0.50567446E+03 -0.74860778E+03 0.11015966E+04 0.50000000E+03 0. 0.
      21 2 2 2 0 0 -0.26026458E+03 0.28784354E+03 0.15285483E+04 0.15770389E+04 0.00000000E+00 0. 0.
 9000006 1 1 1 0 0 -0.25757382E+03 -0.87222907E+03 -0.59084362E+03 0.11942462E+04 0.50000000E+03 0. 0.
-9000006 1 1 1 0 0 0.38392038E+03 0.50567446E+03 -0.74860778E+03 0.11015966E+04 0.50000000E+03 0. 0
.....
</event>
(Note that the momenta are absolutely identical among the copies)

I then tried to set my particle ID to that of a neutralino. This works for the particle, but for the antiparticle there are again two entries, e.g.:
<event>
 10 100 0.10000000E+01 0.00000000E+00 0.00000000E+00 0.00000000E+00
       2 -1 0 0 0 0 0.43628158E+01 0.80184769E-01 0.17795183E+04 0.17795237E+04 0.00000000E+00 0. 0.
       1 -1 0 0 0 0 0.10181761E-01 0.12094626E+01 -0.12769449E+04 0.12769455E+04 0.00000000E+00 0. 0.
 1000022 2 1 1 0 0 0.21850158E+03 -0.18960081E+03 0.48798450E+03 0.57290512E+03 0.80000000E+02 0. 0.
-1000022 2 1 1 0 0 0.90491675E+02 0.72339625E+03 0.30341707E+03 0.79369566E+03 0.80000000E+02 0. 0.
       2 2 2 2 0 0 0.28503649E+03 -0.60110842E+03 0.40335444E+03 0.77799225E+03 0.00000000E+00 0. 0.
       1 2 3 3 0 0 -0.58965674E+03 0.68602634E+02 -0.69218257E+03 0.91187615E+03 0.00000000E+00 0. 0.
-1000022 1 1 1 0 0 0.90491675E+02 0.72339625E+03 0.30341707E+03 0.79369566E+03 0.80000000E+02 0. 0.
...
</event>

I am not sure if this problem also occurs in the HEP-files, as I don't know how to read them.

When analyzing the files (using MadAnalysis 5), I explicitely defined the particles using their IDs.
In the HEP-case, this seems to work, but for the LHE-file it doesn't. To be more precise:
- Using PIDs 900006, LHE-plots are completely empty.
- Using Neutralino-ID, the antiparticle is ignored
(Both are consistent in the sense that non-SM-IDs are ignored).

So to me it looks as if there is some bug in converting the HEP to the LHE-file.
However, I am anyway not very sure about these methods to include invisible particles, so is there an "official" guideline to implement them? I have read various answers here, but all of them seem to be some kind of hacks.

Another question: The pythia-log says that my invisible particles are actually changed during showering (their momenta shift slightly), although I would have expected that they do not participate at all during showering,\ and should not be merged with other particles to form jets (which would explain the changed momenta). Is this some misunderstanding of mine or is there another problem here?

Cheers,
Markus

Question information

Language:
English Edit question
Status:
Solved
For:
MadAnalysis 5 Edit question
Assignee:
MadAnalysisTeam Edit question
Solved by:
Markus Ebert
Solved:
Last query:
Last reply:
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#1

Hi Markus,

Afterwards, however, the pythia.lhe-file contained every particle twice. An example:

This is not the exact same line. the status of the particles changed (the first line is status 2 and the "copy" is status 1.
So they are technically not a copy of each other.

However, I am anyway not very sure about these methods to include invisible particles, so is there an "official" guideline to implement them? I have read various answers here, but all of them seem to be some kind of hacks.

Another question: The pythia-log says that my invisible particles are actually changed during showering (their momenta shift slightly), although I would have expected that they do not participate at all during showering,\ and should not be merged with other particles to form jets (which would explain the changed momenta). Is this some misunderstanding of mine or is there another problem here?

For those two question, I would suggest that you contact the pythia author.

Cheers,

Olivier

Revision history for this message
Markus Ebert (markus-ebert) said : 		#2

Hi Olivier,

thanks for the answer!

What do these status flags say? Anyway, it is strange to observe different behavior depending on which particle ID I assign.

How about my second problem, that analyzing the LHE-file with MA5 gives unreasonable results?
This makes using this file kind of useless in this case.

I will contact the pythia authors about the rest and post the response here, as it might be quite useful for other people as well.

Cheers,
Markus

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#3

The status indicates if a particle is an input/intermediate/final particle (look at LHE convention paper for more details).
The fact of in some case you have two lines and sometimes one depends of how pythia handle the particle. I will not worry about that. (But you can ask Pythia author about that.)

Concerning MA5, I will assign this question to the MA5 team such that they can help you on this.

Cheers,

Olivier

Revision history for this message
Benjamin Fuks (fuks) said : 		#4

Hi Markus,

Could you please provide me the script you are using as well as the event file?

Cheers,

Benjamin

Revision history for this message
Markus Ebert (markus-ebert) said : 		#5

Hi Benjamin,

this is the script I use:

> #The HEP file from pythia output (hadronized)
> import rm100_pythia_events.hep.gz as HEP
> #The converted LHE file (clustered)
> import rm100_pythia_events.lhe.gz as LHE
> #For comparison: parton level
> import events.lhe.gz as PARTON
>
> set main.stacking_method = normalize2one
>
> #new invisible particles
> define wimp = 9000006
> define wimp~ = -9000006
> define invisible = invisible wimp wimp~
>
> plot MET 100 5 1000
> plot PT(wimp)
> plot PT(wimp~)
>
> submit

I attached sample files with only 100 events as well as the output of the same script for a sample with 10^6 events.

There are 2 things I observe in the output:
1) Pythias HEP file plots MET correctly (compared to the parton-level output), but does not show PT(wimp). This is reasonable on hadronized level and I checked that it is consistent with the output after running PGS.
2) The LHE file plots PT(wimp) correctly (compared to the parton-level output).
   However, in MET they are ignored. There is only some peak around 0 which also occurs when not adding the particles to invisible.
  This also looks like a problem, since it does not appear in pyhias HEP-file.

Maybe I am using the LHE file for something I should not, but my intention was to compare parton level results with pythia-results. Therefore it would be great if MET would be correctly interpreted in order to check that that my invisible particles are not influenced due to showering.

Cheers,
Markus

Revision history for this message
Markus Ebert (markus-ebert) said : 		#6

Ok, stupid question: How can I append a file to my question? :-)

Revision history for this message
Benjamin Fuks (fuks) said : 		#7

Hi Markus,

Please send it to me via e-mail if it is not too large.

Cheers,

Benjamin

On Nov 21, 2013, at 16:21 , Markus Ebert <email address hidden> wrote:

> Question #239543 on MadAnalysis 5 changed:
> https://answers.launchpad.net/madanalysis5/+question/239543
>
> Markus Ebert posted a new comment:
> Ok, stupid question: How can I append a file to my question? :-)
>
> --
> You received this question notification because you are an answer
> contact for MadAnalysis 5.

Revision history for this message
Benjamin Fuks (fuks) said : 		#8

Hi Markus,

> 2) The LHE file plots PT(wimp) correctly (compared to the parton-level output).
> However, in MET they are ignored. There is only some peak around 0 which also occurs when not adding the particles to invisible.
> This also looks like a problem, since it does not appear in pyhias HEP-file.
This is fixed in v1.1.10. Now MadAnalysis account for what is in invisible to calculate the MET. Now both samples give identical results for PT(wimp), PT(wimp~) and MET.

Cheers,

Benj

PS: the new version of the program can be obtained from https://code.launchpad.net/~ma5/madanalysis5/v1.1.10beta

Revision history for this message
Markus Ebert (markus-ebert) said : 		#9

Thanks Benjamin!