Jet Charge and b-tagging

[Amin Abou Ibrahim]

Dear Delphes team,

I am using the SNOWMASS background files in my analysis (14 TeV with no pile-up) and I have come across some unusual things. I have contacted the SNOWMASS team but they are all scattered and no one seems to know the answer (as it's been 4 years now and for some reason their website is down).
I realized they generated those files using Delphes 3.0.9.1 so I figured I'd ask here:

1) The jet charge (corresponding to the hadronic tau) amounts to 0 and 1 in the Jet branch. It seems there are no events with negative charge jets. Do you know why is this the case? On the other hand, my ROOT file corresponding to a point I simulated has a histogram nearly symmetric around zero which makes sense.

2) The b-tagged jets are binned using some 0, 1, 2 and 3 bits. I understand this points to no b-jets, loose, tight and both loose and tight b-tagging, respectively. How can I extract this information in a simple code?

Thanks in advance.

Best,
Amin

[Michele Selvaggi]

Hi,

1) the tau charge was probably bugged at that time. There are two ways you could recover it:

• you match the tau -tagged jet with a DR criterion to a gen level tau and give it the charge of the gen level tau.
• reproduce the samples with a more recent Delphes version.

2) see here:

https://cp3.irmp.ucl.ac.be/projects/delphes/wiki/WorkBook/Modules#BTagging

Michele

[Amin Abou Ibrahim]

Hi Michele,

Thank you for the reply!
I will try to implement the first choice as reproducing all the samples will take forever. However, I cannot find resources online that explain how this matching is done using delta R.
Can you elaborate more on this point?

Thanks again.
Amin

[Michele Selvaggi]

It simply consists in finding the parton that is most likely to have originated the jet.
We usually do this by looping at the event record on generated partons and look for the one that is closest to the reconstructed jet.
See here for an example:

​https://github.com/delphes/delphes/blob/master/examples/Example4.C#L187-L214

[Amin Abou Ibrahim]

Thanks a lot. I will implement this method.
I have one more general question about the tree description:
In the Jet branch of a ROOT file, there is the "Jet.Charge" leaf. The Delphes manual says that the Charge leaf corresponds to the charge of the tau. But the way I am extracting the charge of the tau in the ROOT file which properly displays it (not bugged) is by looping over all jets and asking whether it is tau tagged or not. If tau tagged I accept and store the data in a list. Then I access that list and loop over tau jets now and then access the charge (jet->Charge). Now, when I directly access the leaf named Charge in the Jet branch, the output I get differs from what I'm getting using the method I described. I am confused on which is the right way. My intuition says that what I'm doing is correct and Delphes probably meant the charge of all jets (tau tagged as well as others).
What do you think?
Thank you again and sorry for bombarding you with so many questions.

Best,
Amin

[Michele Selvaggi]

Not sure I understand exactly your problem, but what is meant in the documentation by "tau charge" is simply that when the jet is tau-tagged the Charge variable gives the charge of the associated tau. If it not tagged then it is simply giving the sum of charges of the charged tracks constituents of the jets (if any). Hope this clarifies.

[Amin Abou Ibrahim]

Yes it does. That's what I was implying: the "Charge" here means that of all jets and one needs to check whether it is tau tagged or not to be able to extract the right charge of the hadronic tau.
Thanks a lot for all the help.

All the best,
Amin