=== Fitting background. Part 2 ===
  - Jobs failed ( ran on all.q queue that allows only 10 hours of running, but needed ~20). Rerun with long.q.
  - RooFit is implemented
  - Meanwhile, trying to fit background with RooFit.
  - Very bad fit using RooFit. {{ :users:danyyl:dijet:bkg_fit4.jpg?nolink |}} Tried to fit 800-2500 GeV dijet mass range. Much better fit (due to lower statistics){{ :users:danyyl:dijet:bkg_fit_other_range.jpg?nolink |}}. The function that I use is ''(function from EXO paper)*0.5*(1.0 + TMath::Erf( (mjj - m_eff) / sigma_eff ) )''. Took this from https://github.com/CMSDIJET/DijetRootTreeAnalyzer/blob/master/fit/fitMjj_scouting.py .
== Next steps ==
  - Wait for jobs to finish.
  - Fit background.
  - Fix ratio plot