Fbetti dstp2hh dp2hhh

Add the momentum spectra obtained from Run-2 data:

• MomentumSpectrum_413.root obtained from prompt D*+ -> D0 (-> K- pi+) pi+ decays; selection described in Run-2 DACP ana note (https://cds.cern.ch/record/2312320?ln=en), with relaxed fiducial selection as described in https://indico.cern.ch/event/780618/#2-update-of-delta-a_cp-to-the
• MomentumSpectrum_411.root obtained from prompt D+ -> K- pi+ pi+ decays; selection described in slide 46 of https://indico.cern.ch/event/780620/#5-status-report-on-a_cp-in-d0t, with 3 GeV/c < pT(D+) < 12 GeV/c, 2.1 < eta(D+) < 4.5 and without the K-pi+ and D+h+ combined cuts

Thanks for this @fbetti. Looking ahead, do we also want to do something similar for Run 1 data?

For the moment, we think that, for ACP(KK), using the MC as a cross-check for Run 2 analysis will be enough, so we don't think the similar for Run 1 is needed.

I'm thinking for other analyses who may want to do something similar. It may be enough to also commit scripts so that anyone else could perform something similar.

I assume also that you've studied the sculpting of the distributions from TIP cuts, etc?

I don't know how much these spectra will fit for analyses that use quite different selection. The main goal of using input spectra obtained from specifically-selected data is to optimize the generation efficiency, so we did not study in detail the sculpting of the distributions from selection. Of course, some weighting on kinematics will be needed a posteriori in order to have a nice data-MC agreement, but this will happen in any case.

OK, just making sure that the discussion was had somewhere.

It is not fully obvious to me how you are dealing with the fact that what you see in data has reconstruction/selection efficiency included and when you will run MC this will be included again. Maybe I'm overlooking something, but I really want to make sure we can understand what we put in also in a few years from now.

I am not sure to understand. The goal of the dedicated momentum spectrum is to increase the reconstruction/selection efficiency during the final steps of the simulation. So, if we use an input distribution obtained from the already-selected data, aren't we going to have a very high efficiency? Of course this approach is not valid for measurements which need an evaluation of absolute efficiencies, but this is not the case since we are going to use these samples for asymmetry measurements.

@kreps, I do agree with you but since these data lay in the directory ppCollisions I think there won't be any misunderstanding in the future. It's not clear to me if the merge is not yet done because you are not convinced about the usage of real data or simply we have to wait for something else?

approved this merge request

@fbetti, you are not getting exactly what you want because you start from momentum spectrum to which efficiency is already applied and then you effectively apply part of it again (think of momentum dependent selection efficiency). Part of my problem is that while you know what you are doing, in future uses people might not realise. In my view, the better option would be a spectrum, which would correct for relative efficiency such that after you reconstruct your sample, momentum distribution shape would correspond to that you see in data.

@kreps , now I understand better your point, thanks for the clarification. In principle it's perfectly true what you say, but I believe in this particular case the effect will be negligible (but I have no strong/obvious argument to prove it). So basically this spectrum should be, as you say, corrected by the selection efficiency. The resulting spectrum will be equivalent to the spectrum obtained by a Pythia production with the generator-level cuts specific for each dec file, am I right? If no, please correct me. If yes, we could just run a generator-only production for each dec file which is involved, right? But also in this case, the momentum spectrum will depend on the particular dec file, so this information will need to be clear for other users in the future as well.

It is not guaranteed that once you correct data you will get same as Pythia, only if Pythia does the right job. I will make a few changes tomorrow and start preparing the release.

added 1 commit

• 06478c72 - spectrum obtained from Run-1 Pythia and corrected by efficiency

Compare with previous version

Hi @kreps , the spectra are now corrected by efficiency.

• MomentumSpectrum_413.root is obtained from prompt D+ -> D0 (-> K- pi+) pi+ decays. The selection is defined by the Turbo line Hlt2CharmHadDstp2D0Pip_D02{0}Turbo, with further requirements m(D0pi) < 2020 MeV/c2, D0_IP_OWNPV < 0.06 mm and Dst_L0Global_TIS == 1. The spectrum is corrected by the efficiency in the following way: a) a complete pGun production chain starting from D Run1 Pythia distribution (EvtType 27163100 with loosened generation cuts) has been performed b) efficiency as a function of (pT, eta) has been calculated comparing the reconstructed and input distributions c) the data spectrum has been divided bin-by-bin by the efficiency

• MomentumSpectrum_411.root is equal to MomentumSpectrum_413.root. We checked a-posteriori, with a PGun production that uses this new D*+ spectrum, that the agreement between simulation and data is good also in the D+ channel.

added 1 commit

• a2ccf6ca - Add options file to set source of input spectrum derived from pp collision data

Compare with previous version

mentioned in commit 25ab69ee

merged