search

cms-sw / genproductions

Generator fragments for MC production
https://twiki.cern.ch/twiki/bin/view/CMS/GitRepositoryForGenProduction
79 stars 788 forks source link

Changes on ttHH and tt4b for Run 3 #3800

Closed G-Vian closed 1 day ago

G-Vian commented 2 weeks ago

Dear GEN team,

I would like to request you to verify the following changes on the files of ttHH for run 3: 1) I have changed the value of bwcutoff from 15000 to 15 2) I have added a decay of H to b and b~ , as my team is looking for TTHH with each Higgs going into b quark pairs.

On my last pull request [1] you have asked me about the values of minimal invariant mass of a jet pair and a b pair, for the tt4b file [2].

2. In the tt4b run card you are applying the following cuts which are likely not necessary (unless you have a specific reason) and could have an impact on the analysis:

  10.0    =  mmjj     ! min invariant mass of a jet pair 
  10.0    =  mmbb     ! min invariant mass of a b pair

I have checked this value with my team and they agreed with it.

sihyunjeon commented 2 weeks ago

this is an awkward setup.

  1. you decay top in pythia8 where you lose spin correlation and is better with madgraph. but at the same time you decay higgs in madgraph where spin correlation doesn't play any role since the spin is 0. you should do other way around instead or is there any reason behind this?
  2. "I have checked this value with my team and they agreed with it." agreed with what? you do understand that this will unnecessarily cut out events with all jet pair mass at 10gev for ttbbbb sample and this could happen frequently, right? this cut applies to "all possible combinations of jets". I mean it might not matter too much in the end but who knows.
asavoy829 commented 2 weeks ago

Dear Sihyun Jeon,

1) ttHH4b: Answer to your point 1: We don't understand your point 1. What we are considering in this study is the SM non resonant production of a pair of tops with a pair of Higgs bosons. Besides, we are focusing on the Higgs pair with each Higgs decaying into a b-quark and inclusive top pair production. This is why we include the request of Higgs decay into a pair of b-quarks at the GEN level But if we understand your point 1) this is done at the PYTHIA level ? So we should not have this request at GEN and include this when going to PYTHIA stage? As for the top-pair production as said we are interested in being inclusive. If so we will not include the request on Higgs decay into pair of b-quarks in the cards.

2) tt+4b: Answer to your point 2: The request of a cut on mjj and mbb was included by the previous ttHbb MC generator convener. We don't have any issue with this and understand the effect it might have on this sample that we use for completing the study of the top pair QCD production background on our ttHH4b analysis. This is what we used for the complete study of that overall tt background for Run 2 and indeed we want to keep the same conditions on this part of the tt background for the continuation of our work at Run3.

Thanks again for your valuable assistance Aurore and Gabriel

sihyunjeon commented 2 weeks ago

Decaying a particle to its daughters is all available in both madgraph and pythia.

What you are choosing here is : "decaying top inclusively with pythia" + "decaying higgs to bb with madgraph" Computationally this is fine, no issue with "making it run" without considering physics implementations. In physics terms, it's not a valid modelling. if you decay a particle with pythia, you lose spin correlation while madgraph keeps the spin correlation. And because of their spins, top is going to be affected by spin correlation (hence the physics distributions will look different depending on whether you choose pythia or madgraph to do the work, and madgraph is correct and pythia is wrong) while higgs with its spin 0, spin correlation can be ignored (hence the physics distributions will look identical depending on your choice and both are correct). So what I mean by "awkward" choice is, you are making a choice that takes care of higgs decay in madgarph where it doesn't matter (both madgraph, pythia are correct) while you just choose top decay to happen in pythia (which is wrong). Choices are the following

  1. Decay both top and higgs in MadGraph : both top and higgs descriptions are correct
  2. Decay only top in MadGraph : both top and higgs descriptions are correct (as I said, Higgs decay will be modelled in same way, madgraph vs pythia)
  3. Decay both top and higgs in Pythia : top is incorrect while higgs is correct (it's fine if you don't care about spin correlations too much + the sample generation will be much faster) --> However you are choosing 4. Decay only higgs in MadGraph : which has no benefit over any of the above

About 2., approval of previous convener doesn't really make the modelling correct to my senses but if that's what you want no issue for me.

asavoy829 commented 2 weeks ago

Many thanks for the prompt and clear answer Sihyun, Aurore and Gabriel

On 13/11/2024 15:59, Sihyun Jeon wrote:

Decaying a particle to its daughters is all available in both madgraph and pythia.

What you are choosing here is : "decaying top inclusively with pythia"

  • "decaying higgs to bb with madgraph" Computationally this is fine, no issue with "making it run" without considering physics implementations. In physics terms, it's not a valid modelling. if you decay a particle with pythia, you lose spin correlation while madgraph keeps the spin correlation. And because of their spins, top is going to be affected by spin correlation (hence the physics distributions will look different depending on whether you choose pythia or madgraph to do the work, and madgraph is correct and pythia is wrong) while higgs with its spin 0, spin correlation can be ignored (hence the physics distributions will look identical depending on your choice and both are correct). So what I mean by "awkward" choice is, you are making a choice that takes care of higgs decay in madgarph where it doesn't matter (both madgraph, pythia are correct) while you just choose top decay to happen in pythia (which is wrong). Choices are the following

    1. Decay both top and higgs in MadGraph : both top and higgs descriptions are correct
    2. Decay only top in MadGraph : both top and higgs descriptions are correct (as I said, Higgs decay will be modelled in same way, madgraph vs pythia)
    3. Decay both top and higgs in Pythia : top is incorrect while higgs is correct (it's fine if you don't care about spin correlations too much + the sample generation will be much faster) --> However you are choosing 4. Decay only higgs in MadGraph : which has no benefit over any of the above

About 2., approval of previous convener doesn't really make the modelling correct to my senses but if that's what you want no issue for me.

— Reply to this email directly, view it on GitHub https://github.com/cms-sw/genproductions/pull/3800#issuecomment-2473862641, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARSKMLAPR23LQKLFKXVXWDT2ANSNLAVCNFSM6AAAAABRSSMAZ2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINZTHA3DENRUGE. You are receiving this because you commented.Message ID: @.***>

--------------TO7BeBlLOEgshN0LmrNxT8R1 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: 8bit

<!DOCTYPE html>

Many thanks for the prompt and clear answer Sihyun,
Aurore and Gabriel

On 13/11/2024 15:59, Sihyun Jeon wrote:

Decaying a particle to its daughters is all available in both madgraph and pythia.

What you are choosing here is : "decaying top inclusively with pythia" + "decaying higgs to bb with madgraph"
Computationally this is fine, no issue with "making it run" without considering physics implementations.
In physics terms, it's not a valid modelling. if you decay a particle with pythia, you lose spin correlation while madgraph keeps the spin correlation. And because of their spins, top is going to be affected by spin correlation (hence the physics distributions will look different depending on whether you choose pythia or madgraph to do the work, and madgraph is correct and pythia is wrong) while higgs with its spin 0, spin correlation can be ignored (hence the physics distributions will look identical depending on your choice and both are correct).
So what I mean by "awkward" choice is, you are making a choice that takes care of higgs decay in madgarph where it doesn't matter (both madgraph, pythia are correct) while you just choose top decay to happen in pythia (which is wrong). Choices are the following

  1. Decay both top and higgs in MadGraph : both top and higgs descriptions are correct
  2. Decay only top in MadGraph : both top and higgs descriptions are correct (as I said, Higgs decay will be modelled in same way, madgraph vs pythia)
  3. Decay both top and higgs in Pythia : top is incorrect while higgs is correct (it's fine if you don't care about spin correlations too much + the sample generation will be much faster)
    --> However you are choosing 4. Decay only higgs in MadGraph : which has no benefit over any of the above

About 2., approval of previous convener doesn't really make the modelling correct to my senses but if that's what you want no issue for me.

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you commented.Message ID: <cms-sw/genproductions/pull/3800/c2473862641@github.com>


--------------TO7BeBlLOEgshN0LmrNxT8R1--

G-Vian commented 2 weeks ago

Dear Sihyun Jeon,

I removed the H to bb~ decay because I will request that this decay be implemented in Pythia when asking the Monte Carlo conveners to run the simulation. I will also ask them to keep the top quark undecayed.

Thank you very much!

G-Vian commented 1 week ago

Dear GEN members,

I have created a card for the process ttZ2b, at the energy of 13.6 TeV. This is another process my group wants to simulate. I've paid attention to all the suggested parameters from the previous discussions.

G-Vian commented 1 week ago

Dear GEN Team,

We would like to kindly request if there are any updates on the running of the generation of ttHH, tt4b, and ttZ2b cards we provided.

Thank you very much!

Aurore & Gabriel

asavoy829 commented 6 days ago

Dear Sihyun Jean and GEN team We would be reolly grateful if you can give the green light for the generation of the 3 processes which are essential for our ongoing study for. Run 3, namely: ttHH, tt4b and ttZ2b. Many thanks formal Aurore and Gabriel

lviliani commented 2 days ago

@sihyunjeon do you have further comments on this?

sihyunjeon commented 2 days ago

not from my end