Reproducing numu 2017 Analysis results

The working directory is CAFAna/numu/Analysis2017/ (add CAFAna to your private
release if you haven't yet).

Analysis summary

Several changes have been made to the analysis cuts and method as compared to
SA. These changes include new cuts, new weights, an updated energy estimation,
separating events by energy resolution in both detectors and an optimised neutrino
energy binning.

A detailed overview of the analysis can be found here:

Central weights

All plots should be using these weights, which adjust the flux and cross-sections to our central tunes:

kPPFXFluxCVWgt * kXSecCVWgt2017


The datasets used for this numu analysis are the same as the datasets
used for the nue analysis.

Making predictions

Main results are made in a similar fashion as SA. The statistics only result can be
produced in the same way as the SA. However, for the correct treatment of the
systematics the predictions are made using:

cafe -bq CAFAna/shared/Ana2017/make_predictions_systs.C

This creates all the systematic files you want. All the possible combinations can be found in :

cafe -bq CAFAna/shared/Ana2017/reduce_the_predfile_nue_numu.C

This creates a root file called pred_numu_reduced_v5.root. This prediction has
already been created for this analysis and is pulled in at the next step.

Making the result contours

The result surface and 1D histograms of chi-squared vs. the mixing angle and the
mass splitting are produced by running:

cafe -bq getContProf.C

This macro provides the several options for the creation of the contour via the macro
arguments. The result with or without systematics is produced depending on the
value chosen for the boolean "usesysts".

A comparison of the contours is drawn by doing:

cafe -bq plotContProf.C

Getting the cosmic background estimate

The cosmic background in the far detector within the beamspill is estimated using
both the cosmic trigger data and the timing sidebands of the numi trigger data.

Distributions of important analysis variables can be created for both trigger streams
by running:

 cafe -bq compare_cos_numi.C 

Other important details for results

Note that these results do not include FC corrections.