include first version of the 2TagCount code#127
Conversation
| // N.B. for GetBinContent(X) 'X' must be weight index + 1 due to way histogram is filled (0th bin always underflow). | ||
| systWeight["qcdScaleLo"]=ev.ttbar_w[9]*(xsecWgt->GetBinContent(10)/xsecWgt->GetBinContent(1)); | ||
| systWeight["qcdScaleHi"]=ev.ttbar_w[5]*(xsecWgt->GetBinContent(6)/xsecWgt->GetBinContent(1)); | ||
| systWeight["hdampLo"]=ev.ttbar_w[ev.ttbar_nw-29]*(xsecWgt->GetBinContent(ev.ttbar_nw-29+1)/xsecWgt->GetBinContent(1)); |
There was a problem hiding this comment.
@Wilsker These weights don't seem correct. From data/mc it is clear mc normalization is much too large. Is there another factor?
|
@capalmer85 it's fine to merge it here, but you should also forward port to all the following branches (10.2 etc..). Given the analysis code is quite independent from the BTA release cycle you could/should open a new repository under cms-btv-pog to keep your code there. |
|
@mverzett Thanks for having a look. I've integrated this code with the same code for the kin fit. In fact the only part that is really independent is the part that measures efficiency from the histograms (although kin fit has a separate script like that too). Maybe everything in test/ttbar should go into a new repo? @Wilsker what do you think? |
|
The code is essentially a standalone package run on the BTA output. It has a different release cycle and is developed by a different team. I think a separate repo for the ttbar directory could be a good idea and encourage it's upkeep. |
3 participants
@Wilsker Is this the right branch to merge into? (Pointing you to the lines in question in a moment.)