Last week: 
1. Jet calibrations especially MC calibration
Today:
1. Pile up:
   1. Assume pileup is symmetric and figure out what the offset is and subtract it from the event
   2. Why do we get asymmetric pileup? Shift in the PV. 
   3. This mean that is subtracted is not calibrated hence is not uniform in eta.
   4. In time (same bunch crossing) vs out of time (different bunch crossing) pile up
   5. Avg no. of collisions (mu formula in paper)
   6. After you subtract mean, there is residual mean and N_PV.
   7. Full procedure: rho*area where rho is the median density which is roughly 0.5 per unit area. This might not have been done at 7 TeV, check! This was definitely done at 8 TeV and later.
   8. Next week: Difference between Run 1 vs now. Pile up will be important
1. In situ subtraction:
   1. Correcting for residual between data and MC
   2. Section 21 has plots of each in situ corrections which we want to understand.
   3. Section 7.1: Calibrate forward jets w.r.t central jets. Use Z/gamma jets like events where Z/gamma is central and corrections calculated using central jets and apply it to forward ones.
      1. Why not Z->mumu+jets? Muon resolution is better for high pt and electron resolution is better at low pt. 
      2. The intra eta calibration is highest and it looks like “bat man ears”!!
      3. For Z and jets, we don’t do absolute calibration of jet pt to match that of Z but what we do is apply correction to pt imbalance in MC by comparing that to data. We apply imbalance correction to data because we want to correct only for detector effects not for QCD effects.
      4. When Jennet tried to generate multijet, there was imbalance in intra eta calibration for the third jet. Sometimes the third jet is wrong!
      5. Most jet uncertainties is dominate by modelling like Pythia vs Herwig!!
      6. In situ > 2 TeV is dominated by single particle because we don’t have any in situ there. We can use single particle calibrations and extrapolate this. Page 78, figure a shows spike at 1 TeV
      7. Dijets self calibration or top mass calibration: You can use W mass to constrain the jet energy scale. For example in Higgs, they use Higgs mass to constrain photon energy scale correction. Section 14 summarizes JES.
   1. Section 9.6: 
      1. MPF: Here, ETmiss is used as proxy for negative of the hadronic recoil calculated in the opposite direction to photon pT?
      2. Note: the hadronic activity is used here rather than the reconstructed jets, so you are not sensitive to those corrections.
      3. Why not do MPF for Z+jets? May be electrons introduce more uncertainties because of tracking etc.
   1. Jet calibration:
      1. Do it iteratively
      2. Easier to calibrate lower energy jets
   1. Single hadron correction (Section 14):
      1. Single hadron corrections are important at higher energy
      2. You get these from SPS, single hadron corresponds to pion, D meson,etc rather than jets!!
1. Figure 59 (a):
   1. Low pt: Z+jets important
   2. In situ calibration