MadGraph5_aMC@NLO

Which cross section to use?

Asked by Thomas Ernst

Hello,
I have a conceptual question.
So, when I want to recreate an experimental paper for recasting purposes one simple tool-chain seems to be:
MadGraph(generate samples) -> Pythia (Shower, Match/Merge) -> Delphes (Add rough detector effects) -> MadAnalysis/Root
With the integrated luminosity and the cross section i can normalize the total number of events to the correct value with N=L*sigma (I guess neglecting trigger efficiencies and so on).
Say my background is W+jets and i need to generate up to 2 partons. Which cross section sigma do I need for this? Obviously not the one before matching, but is it the matched cross section? Or is it the W+0j one, since I'm interested in this process and only generate additional jets to correct the approximations done by showering?

Another thing i don't quite get is how people normalize their number of events to higher orders. MC events seem to be most often generated at LO, but later normalized to higher order precision, sometimes citing 1405.0301 for the NLO cross section. But which one is the correct cross section? Do people simply chose the NLO value for W+0j?

Cheers!

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Olivier Mattelaer
Solved:
Last query:
Last reply:
Revision history for this message
Best Olivier Mattelaer (olivier-mattelaer) said : 		#1

Hi,

In general, the 0j cross-section and the cross-section after merging should be close to each other (at least for cases where you have a pure QCD process involve, no EW diagram opening for aditional jet/...).
In that case the difference between the two cross-section is/should be smaller than your LO theoretical uncertainty. So in a way it should not matter which one of the two you are using. Now I would just follow what your community is doing on that point.

To the question, "which one is the correct cross-section for higher order", the real answer is None, they are all estimator (with an estimated error). The advantages of higher order is that the estimated error are smaller and therefore we believe (and observe) that this closer to the experimental truth.

Revision history for this message
Thomas Ernst (th-ernst1997) said : 		#2

Hi,
 thank you for your reply! That clears it up.

Cheers!

Revision history for this message
Thomas Ernst (th-ernst1997) said : 		#3

Thanks Olivier Mattelaer, that solved my question.