Workings of Cutflow

Asked by Amin Aboubrahim on 2017-12-10

Hi Benjamin,

I am trying to understand how MA5 implements the cut flow on signal and background to get S/B.
My backgrounds are weighted but I don't know how MA5 is using those weights. All I can see is that MA5 scales the cross-section by the luminosity and the result would be the number of events used in the cut flow.
Could you explain what happens behind the scene in MA5? How does it use the actual number of events and weights in the sample during the implementation of cuts?

Thank you,
Amin

Question information

Language:
English Edit question
Status:
Solved
For:
MadAnalysis 5 Edit question
Assignee:
No assignee Edit question
Solved by:
Benjamin Fuks
Solved:
2017-12-11
Last query:
2017-12-11
Last reply:
2017-12-11
Benjamin Fuks (fuks) said : #1

Hi Amin,

> I am trying to understand how MA5 implements the cut flow on signal and background to get S/B.
> My backgrounds are weighted but I don't know how MA5 is using those weights.
Instead of counting a single event for 1, it uses the weight value.

> All I can see is that MA5 scales the cross-section by the luminosity and the result would be the number of events used in the cut flow.
As the sum of the weights equals the cross section of your sample, this is thus not surprizing. The contrary would have been a problem :)

> Could you explain what happens behind the scene in MA5? How does it use the actual number of events and weights in the sample during the implementation of cuts?
See above. MA5 uses the weight values to increment the cutflows and the fill the histograms.

Cheers,

Benjamin

Amin Aboubrahim (amin83) said : #2

Hi Benjamin,

Thanks for the explanation.
I am using the SNOWMASS SM background samples which come in ROOT format. The sum of the weights in each file is the NLO cross-section x BR. Also the snowmass paper tabulates the LO cross-sections of the samples.

1) I realized MA5 does not read the cross-section from a ROOT file (it doesn't exist) so I need to set it manually. I am adding the weights and giving the result to MA5 as the cross-section. Is that right? Or should I use the LO? This part confuses me!
2) Each event has a different weight. Say we pick two random events with weights x and y. Say I apply a cut and those two events were rejected, then if N is the product of the NLO cross-section (or LO?!) and the luminosity, so the remaining events given my MA5 would be N-x-y. Right? This is what I understood from your explanation.
3) Can I work with MA5's cutflow the opposite direction: asking for the integrated luminosity which gives a 5 sigma S/B?

I appreciate the help.

Best,
Amin

Best Benjamin Fuks (fuks) said : #3

Hi Amin,

> 1) I realized MA5 does not read the cross-section from a ROOT file (it doesn't exist) so I need to set it manually. I am adding the weights and giving the result to MA5 as the cross-section. Is that right? Or should I use the LO? This part confuses me!
MA5 cannot read something that is not existing… that is correct. You can normalize the sample as you want. it depends what you want to do. There is no correct and incorrect way.

> 2) Each event has a different weight. Say we pick two random events with weights x and y. Say I apply a cut and those two events were rejected, then if N is the product of the NLO cross-section (or LO?!) and the luminosity, so the remaining events given my MA5 would be N-x-y. Right? This is what I understood from your explanation.
Yes.

> 3) Can I work with MA5's cutflow the opposite direction: asking for the integrated luminosity which gives a 5 sigma S/B?
No. You have to calculate it by yourself.

Regards,

Benjamin

Amin Aboubrahim (amin83) said : #4

Hi Benjamin,

> MA5 cannot read something that is not existing… that is correct.
I meant MA5 does not add weights and read it as cross-section :)

Thank you again.
Amin

Amin Aboubrahim (amin83) said : #5

Thanks Benjamin Fuks, that solved my question.