Analysis Questions » History » Version 1
Matthew Toups, 01/03/2018 08:42 PM
h1. Analysis Questions
# At the interaction vertex, is momentum conserved? If not, why not? Is it because of a fake particle in the event? How are we handling this strictly from a kinematics point of view (not a model question)? Prove we handle missing energy properly when we move to different frames. If this is not handled right it will really screw up a lot of analyses, especially xsec analyses.
# Are particles boosted correctly? I have been on experiments that did not have the correct boosts. So check it.
# Look at the protons produced in the MEC interaction – are these protons being boosted properly?
# The solid angle of the detector is 5 mrads. Do we correct for this rotation in the simulation in order to match data? In fact, we need to think about beam width and parallax, because a neutrino with a 400 m travel distance actually has an opening angle of 6.25 mrads. Note that this will be flavor dependent.
# A consequence of not getting this right is that your muon angle distribution in the xsecs will be wrong. It also affects any analysis looking for transverse momentum or
# Check the number of neutrons coming out. Tim has said he has seen an unphysical number. What does that mean? Prove you don’t have a mistake.
# Check the dE/dx of the particles coming out against SRIM or any other data base. Prove it makes sense. I think it won’t because…
# The deltas probably have too high of a threshold, since I can pick out data compared to MC in the DL LEE images from the amount of delta sparkle. Check this. This affects the multiple scattering energy measurement, all EM energy measurements, including the michels and also may explain the PIDA problem (and why truncated dQ/dx works better).
# The Michel energy and angle spectrum is surely wrong, since the MC does not include polarization. Explain to me why this doesn’t matter to your analysis.
# Does muon polarization affect the capture cross section? (I don’t know the answer to this – I am just guessing that since capture is a weak process, it will).
# There is a claim that the case of 2 EM blobs near a 1mu1p always corresponds to a pi0, so we don’t need to worry about the mass peak. Prove to me that a relatively high energy photon splitting into 2 blobs that do not cluster is rare.
# How often do you get a pi0 from cosmic right on top of your CCQE event? According to Kazu, who has done a lot of scanning, about 1 in 3 BNBEXT has a pi0. There are so many more cosmics and so many more CCQE than CCpi0 that it seems very likely there will be some. If that isn’t right, please explain.
# Check the rate of pi+/pi0 charge exchange – is it right? Is your CCpi0 analysis taking charge exchange into account properly?
# Check the energies of the e+e- conversion pair. Does the ratio in the simulation make sense?
# Check the energies of the two pi0 photons at truth level. Does the ratio make sense?
# Check the invariant mass of the outgoing proton and pion in the resonant decays. Does it make sense? Yes, it will have smearing. Do a run with no Fermi motion as a cross check.
# Are stopped pi+’s that decay making a Bragg peak in the simulation? They should.
# Use a particle gun to produce pi+’s and check the fraction of time they hard scatter given the energy. Is it right?
# Does every single xsec analysis use the right beam windows for data and MC? Are you aware of the window-length difference when doing the BNBEXT subtraction? Can you prove this is right for every single xsec analysis by making a plot for each?
# Many xsec analyses will need exclusive samples. Is the POT counting for exclusive samples wrong? Prove it is right.
# Nearly all of the xsec analyses are making very tough cuts in order to get to a sample that is as pure as possible. When you do this, the error on the efficiency goes way up. How are you determining the error on your efficiency? Why should I think this is right? Prove to me it makes sense.