mmll cut behaviour in multilepton processes

Asked by Nicolas Tonon on 2020-09-19

Dear experts,

I am comparing a private tZq (actually tllbq) sample to a “default” (CMS) sample, using the same process definitions:
------------------------------
define l+ = e+ mu+ ta+
define l- = e- mu- ta-
define vl = ve vm vt
define vl~ = ve~ vm~ vt~

generate p p > t b~ j l+ l- $$ w+ w- / h
add process p p > t~ b j l+ l- $$ w+ w- / h
------------------------------

In my case I am only interested in trileptonic events, hence I force the leptonic decay of the top in the madspin card (to significantly increase my final selection efficiency) :
------------------------------
decay t > w+ b, w+ > l+ vl
decay t~ > w- b~, w- > l- vl~
------------------------------

However I noticed some disagreements in the kinematics of the 2 samples, which seem to be especially related to the lepton from the top. I would like to know whether this could be related to the mll>30 cut applied in both samples ?
------------------------------
 30.0 = mmll ! min invariant mass of l+l- (same flavour) lepton pair
------------------------------

In other words, is is possible that this introduces a bias because:
a) In the default sample, the cut only applies to the unique OSSF pair (in tl+l-bq), and *then* the top is decayed ?
b) In my private sample, the cut applies to *all OSSF pairs* among the 3 leptons (i.e. it is applied *after* the Madspin decay) ?

Thanks in advance for the clarification, cheers,
Nicolas

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Nicolas Tonon
Solved:
2020-09-29
Last query:
2020-09-29
Last reply:
2020-09-19

They are no cut applied within MadSpin. (because otherwise MadSpin does not know the cross-section since it use the Branching ratio information).

So clearly your cute are not the same between the two processes.

Cheers,

Olivier

> On 19 Sep 2020, at 12:45, Nicolas Tonon <email address hidden> wrote:
>
> New question #693003 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/693003
>
> Dear experts,
>
> I am comparing a private tZq (actually tllbq) sample to a “default” (CMS) sample, using the same process definitions:
> ------------------------------
> define l+ = e+ mu+ ta+
> define l- = e- mu- ta-
> define vl = ve vm vt
> define vl~ = ve~ vm~ vt~
>
> generate p p > t b~ j l+ l- $$ w+ w- / h
> add process p p > t~ b j l+ l- $$ w+ w- / h
> ------------------------------
>
> In my case I am only interested in trileptonic events, hence I force the leptonic decay of the top in the madspin card (to significantly increase my final selection efficiency) :
> ------------------------------
> decay t > w+ b, w+ > l+ vl
> decay t~ > w- b~, w- > l- vl~
> ------------------------------
>
> However I noticed some disagreements in the kinematics of the 2 samples, which seem to be especially related to the lepton from the top. I would like to know whether this could be related to the mll>30 cut applied in both samples ?
> ------------------------------
> 30.0 = mmll ! min invariant mass of l+l- (same flavour) lepton pair
> ------------------------------
>
> In other words, is is possible that this introduces a bias because:
> a) In the default sample, the cut only applies to the unique OSSF pair (in tl+l-bq), and *then* the top is decayed ?
> b) In my private sample, the cut applies to *all OSSF pairs* among the 3 leptons (i.e. it is applied *after* the Madspin decay) ?
>
> Thanks in advance for the clarification, cheers,
> Nicolas
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Nicolas Tonon (ntonon) said : #2

Hi Olivier, thanks a lot for the quick reply.

>> They are no cut applied within MadSpin

Ok I didn't know that, I will need to double-check all the cuts I'm applying then. Just one more thing:

>> So clearly your cute are not the same between the two processes.

Sorry for missing the point: I actually don't want to apply this particular mmll cut on the lepton from the top (just the OSSF pair in tllq), i.e. it's actually good that this cut is not applied in madspin (where I decay the top).
So then, identically for both samples, the cut should *only* be applied to the OSSF pair defined in the MG proc card (tllbq), no ?

Cheers,
Nicolas

I'm actually confused about what you call two sample here since you only specify one syntax.
Do I assume correctly that you compare one version with MadSpin versus one version without MadSpin?

Olivier

Nicolas Tonon (ntonon) said : #4

Yes exactly, sorry that this was not clear. So I was wondering if using Madspin for 1 sample (but not the other) could lead to some interplay with this cut and somehow bias my kinematics.

However based on your previous answer I understand that in that case it should not. And since then, I found out that this bias was mostly due to some additional efficiency cuts I added in the run card, which I had thought to be harmless.

Thanks !
Nicolas

Well without madspin you have additional pair on lepton on which the invariant mass can be applied so this can modify your shape as well. The easiest to check the invariant mass pair cut applied is to look at one sub-channel log
SubProcesses/P..../G1/XXXXX.log

and you have a table with the minimum invariant mass applied for all pair of final state parton.

Cheers,

Olivier

Nicolas Tonon (ntonon) said : #6

This is precisely what had me confused:

>> without madspin you have additional pair on lepton on which the invariant mass can be applied

The MG process is "p p > t b~ j l+ l-". So both with/without Madspin, there is only 1 pair of leptons at MG level (the top is not decayed here), so there can not be any difference regarding the mmll>30 cut... ?

Nicolas

> The MG process is "p p > t b~ j l+ l-". So both with/without Madspin,
> there is only 1 pair of leptons at MG level (the top is not decayed
> here), so there can not be any difference regarding the mmll>30 cut... ?
>

Oh ok I though that you were the same final state output so a syntax like
 p p > t b~ j l+ l-, (t > w+ b, w+ > l+ vl)
or
p p > b~ j l+ l- b l+ vl

Sorry for the confusion

Olivier

> On 29 Sep 2020, at 11:25, Nicolas Tonon <email address hidden> wrote:
>
> Question #693003 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/693003
>
> Nicolas Tonon posted a new comment:
> This is precisely what had me confused:
>
>>> without madspin you have additional pair on lepton on which the
> invariant mass can be applied
>
> The MG process is "p p > t b~ j l+ l-". So both with/without Madspin,
> there is only 1 pair of leptons at MG level (the top is not decayed
> here), so there can not be any difference regarding the mmll>30 cut... ?
>
> Nicolas
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.