unable to include stable particle (tau) in MET spectrum ...

[Shankha Banerjee]

Dear Experts,

I have a hepmc output from Herwig++ and I want to analyse it with Delphes (version 3.3.0). In my hepmc file, I have ascertained my tau particles to be stable (I want to use these as dummies for MET).

In the tcl card in Delphes I add the following line :

add EnergyFraction {15} {0.0 0.0}

However, I see that by adding this line, the MET peak and RMS values do not change as expected (by expected, I mean that I have run a Hadron level analysis in Madanalysis 5 and find that the peak changes from ~ 14 GeV to ~ 38 GeV and the RMS changes from 18 to 60 GeV).

However, here in Delphes I find that without using the line of code (mentioned above), the mean and the RMS of the MET distribution are ~28 GeV and ~22 GeV. Whereas, after adding this line these respective values change to ~ 31 GeV and ~ 27 GeV.

I am unable to understand this mismatch. Am I missing something ? Are the taus somehow not properly treated as MET ? Is the tau module in the tcl file somehow creating any trouble ?

I will be very grateful if you can help me out regarding this.

Thank you very much.

P.S. My root version is 5.34/32 and I am using Ubuntu 14.04.

Best regards,

Shankha.

[Michele Selvaggi]

Hi,

do I understand correctly that you want to treat tau's as stable particle and undetectable particles?
I don't suggest you do this, but if you really know what you are doing you'll need to hack the HepMC reader class in: classes/DelphesHepMCreader.cc

Replace the condition (l.390):

if(fStatus == 1 && pdgParticle->Stable())

by:

if(fStatus == 1)

The problem is that we check in pdf look up table from ROOT whether each particle is really stable or not. In this case the tau is not, so it probably not even properly passed as input Delphes propagation.

In addition you'll have to set the Charge by hand, replace l.350 by:

candidate->Charge = (fPID==15) ? -1.0 : 1.0;

Then you'll have to recompile from the main Delphes dir by typing make.

Cheers,
Michele

[Shankha Banerjee]

Hi Michele,

Thank you very much for your reply.

If I understood you correctly do you mean to change this line :

candidate->Charge = pdgParticle ? int(pdgParticle->Charge()/3.0) : -999;

(which is line 352 in this version.)

by

candidate->Charge = (fPID==15) ? -1.0 : 1.0;

or should I just put this line in the empty line (350) and keep both ?

Do I need to change anything in the tcl card as well ?

Thank you very much for your help.

Best regards,

Shankha.

[Michele Selvaggi]

Hi,

yes you have to replace that line by what I wrote.
The only thing you want to change in the card are the energy fraction as you stated above.

Michele

[Shankha Banerjee]

Hi Michele,

Thank you very much for your kind reply.

I followed the suggestions. However, now I am getting an even pessimistic
result for the MET distribution on including the taus.

Mean : ~ 25 GeV, RMS ~ 23 GeV

For the one without including tau it is now :

Mean : ~ 22 GeV, RMS ~ 19 GeV

Probably, I am still missing something.

Thanks a lot in advance.

Best regards,

Shankha.

[Michele Selvaggi]

Could you attach a small sample (1k event) of the input event file you are using?

Michele

[Shankha Banerjee]

Hi Michele,

I generated a sample for 1k events. The file was ~ 75 MB even after compressing
so I am attaching the dropbox link :

​https://www.dropbox.com/s/cgtgexr3ncqxl0h/HH-bbWW-B3.hepmc.tar.bz2?dl=0

Please let me know if I should send something more.

Thanks a lot.

Best regards,

Shankha.

[Michele Selvaggi]

Hi,

actually ROOT considers tau's stable and in Delphes we are safe because we apply the Status == 1 condition. So replacing that line had no effect.

Now, forget the above and start from a fresh Delphes version (preferably 3.3.2 so we have the same results)

Set the tau as neutral by adding the following line:

if(fStatus == 1 && fabs(fPID) == 15)candidate->Charge = 0;

after candidate->Charge = pdgParticle …

This will avoid confusion with the particle-flow.

Now you can add or remove the line:

add EnergyFraction? {15} {0.0 0.0}

in the calorimeter section and you should see additional MET materializing by adding that line.

For the GenMET I have similar numbers as you. For the reconstructed MET I have now mean 33 GeV vs 52 GeV.

Of course let me repeat that you should not use stable taus (and even less treat them as MET) for a physics analysis, but you should instead provide to Delphes tau that have already decayed into long lived particles.

Michele

[Shankha Banerjee]

Hi Michele,

Thanks a lot. I am now getting the mean and RMS at 48 GeV and 57 GeV respectively.

I am just trying it as a toy model.

I will keep your suggestions fully in mind.

I just have one more small query, when I am converting my .hepmc file to the root format
using the .tcl card, I am getting this message at the start :

WARNING: too small value, return 0-------------------------------] (0.00%)

Should this be of any concern ? I am not getting any warning message in madanalysis though.

Thank you very much once again.

Best regards,

Shankha.

[Michele Selvaggi]

Hi,

no that is no concern, you can ignore it.

Cheers,
Michele

[Shankha Banerjee]

Hi Michele,

That indeed solved my problem.

Thank you so much.

Yours sincerely,

Shankha.