PHYSICS->Isol->eflow->relIsolation returns -1

Asked by Jakub Salko on 2019-11-13


I am working on a recast of the CMS search for top quark partners with charge 5/3 where they define the isolation variable for leptons as "the scalar pT sum of all PF candidates within a cone of varying size around the particle, divided by its pT". I defined this variable in my analysis as

double DR = 10./std::min(std::max(pt,50.),200.); // cone radius
double imini = (PHYSICS->Isol->eflow->relIsolation(Lep,event.rec(),DR,0.5,IsolationEFlow::ALL_COMPONENTS)); // isolation variable
cout << imini << endl; // control output.

The code compiles with no errors and runs with no errors or warnings. The control output shows that imini variable returns -1 for all the leptons. I found the same behaviour for all values of IsolationEFlow::*_component as well as for PHYSICS->Isol->*.
After checking how the relIsolation function is defined in MA I found out it takes the information from Track and Tower branches of the Delphes output which my Delphes output does not contain.

The problem why I find this finding surprising is that I constructed my analysis by using CMS-TOP-17-009 (already implemented in PAD) as a template. The output of my Delphes card mimics the output of CMS-TOP-17-009 delphes card and the imini definitions are practically identical in both analyses.

I would like to ask you if my observation on why the isolation variable defined above returns -1 is correct and if there is a another way of finding the isolation variable besides either hard coding it myself or making some major changes to the Delphes card.

Thanks a lot!

Question information

English Edit question
MadAnalysis 5 Edit question
No assignee Edit question
Solved by:
Benjamin Fuks
Last query:
Last reply:
Benjamin Fuks (fuks) said : #1

Hi Jakub,

Some lines in the public delphes card were commented, and should not have been commented. I am fixing it right away. No change is needed in the code (only remove 5 hashes at the end of the delphes card).



Jakub Salko (jasalko) said : #2

Hi Benjamin,

which Delphes card are you referring to exactly? This fix works for me only with the public CMS delphes card without the pile up but not with the delphes card which includes the pile up nor the CMS-TOP-17-009 one.


Best Benjamin Fuks (fuks) said : #3

Hi Jakub,

I have modified the CMS-TOP-009 delphes card by adding the track and calo information in the output. This should allow for the correct calculation of the isolation variables.



Jakub Salko (jasalko) said : #4

Thanks Benjamin Fuks, that solved my question.