Question on cut

Asked by Li

Hi,

    I am using MG5 to generate pp->ttj, if I use cut like "ptj" in run card. Is the cross section the true cross section?

    That is to say, Is it the cross section ( use cut ) equal to cross section ( do not use the cut ) * the rate ( pt larger than ptj )?

    Another question is that, is there any influence of Matching? Since if I apply a very large cut, the cut is far from the matching area so I can ignore the match?

Best,
Li

Question information

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

Hi,

> That is to say, Is it the cross section ( use cut ) equal to cross section ( do not use the cut ) * the rate ( pt larger than ptj )?

Yes this is the definition of a cut!

Now the tricky part is actually "What is a jet?".
A jet at our level and an experimental jet are obviously not the same.

For LO, we define a jet as the momenta of any quark/gluon
At NLO, we let you define a jet measure and run fastjet on all the quark/gluon

Since this occurs before the parton-shower / hardonization this is obviously not the same object/cut as the experimental cuts on jet.
(but it should be a proxy)

> Another question is that, is there any influence of Matching? Since if I apply a very large cut, the cut is far from the matching area so I can ignore the match?

If you have a very large cut on the first jet, then you should be save to not include the 0 jet multiplicity sample for the matching/merging sample.
(now to be complete, the large cut should be understood as much bigger than the merging scale obviously not large compare to other analysis)
But I do not see why you should be save to not include the 2 jet multiplicity in the matching/merging procedure.

Cheers,

Olivier

> On 16 Jan 2018, at 02:23, Li <email address hidden> wrote:
>
> New question #663113 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/663113
>
> Hi,
>
> I am using MG5 to generate pp->ttj, if I use cut like "ptj" in run card. Is the cross section the true cross section?
>
> That is to say, Is it the cross section ( use cut ) equal to cross section ( do not use the cut ) * the rate ( pt larger than ptj )?
>
> Another question is that, is there any influence of Matching? Since if I apply a very large cut, the cut is far from the matching area so I can ignore the match?
>
> Best,
> Li
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
Li (huangli-itp) said :
#2

Hi Olivier,

    Thanks very much!

    So if I generate "oo->tt" and set the "ptj=100GeV", it means at least one top pt is larger than 100 GeV but when I generate "pp->tt->wbwb", the same cut means at least b or w's pt is lager than 100 GeV. Is that right?

    If it's right, one more question is that how to require top pt in "pp->tt->wbwb"?

Best,
Li

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said :
#3

Hi,

One important detail that I forget to mention is that we consider a jet all MASSLESS quark in top of the gluon.
Therefore the "ptj" cut does not cut the top-quark since this one is massive.
If you have a model with massless b quark then "ptj" will also cut on the pt of the b. While for model with massive b quark you would need to use "ptb".

Now let me stress two points
1) the W is not a quark (and is massive) so it is not impacted by "ptj"
2) the ptj cut is applied on ALL parton that is consider as jet. Which means that for process like p p > j j j. All three jet will have a pt larger that the "ptj" cut. (if you want a cut on the leading/ second leading/... you should use ptj1min, ptj2min, ...)

> So if I generate "oo->tt" and set the "ptj=100GeV", it means at least one top pt is larger than 100 GeV.

If you want to have at least one top with pt larger than 100. You have two option:
1) You can use the cut "pt_heavy".
That cut requires that at least one particle with mass larger than 10GeV has a pt larger than pt_heavy.
(note that for a 2 to 2 process, it is actually the same as requesting that the two final state particle have a pt larger than X)

2) you can use the new cut (introduced in 2.6.1) "pt_min_pdg".
For this one, you have to specify on which particle (identified by their pdg) you want to apply a cut.
so in this case you can put {6 : 100}. Note that such cut also applied on anti-top (pdg=-6) and that applied on ALL top/anti-top
not only on the leading pt one. (as said above for 2 to 2 process they are not difference)

> but when I generate "pp->tt->wbwb", the same cut means at least b or w's pt is lager than 100 GeV.

In this case, pt_heavy will request that either the W+ or the W- has a pt larger than "X".
On the other hand, the use of pt_min_pdg will not applied any cut for this process if you only specify a cut for pdg=6
(you can have a cut on the W by specifying is pdg obviously)

  > If it's right, one more question is that how to require top pt in "pp->tt->wbwb"?

My advise here is to generate p p > t t~. with the cut explained above and ask MadSpin to do the decay.
(you have full-spin correlation and off-shell effects).

Note that like a parton-shower MadSpin performs some re-shuffling of the production event and therefore even if all the event will have top quark with pt larger than X before MadSpin, some of them will have a smaller pt after madspin. (The same is True when you go trough the parton-shower obviously)

So as usual, you have to put quite loose cut at parton level in order to not bias yourself.

Cheers,

Olivier

Revision history for this message
Li (huangli-itp) said :
#4

Hi Olivier,

    Thanks a lot for this detail explanation! And "pt_min_pdg" sounds like really interesting, I will update MG5 and try this.

Best,
Li