Further investigation of DIS tweak. Was it RES RPA that dropped it from 40% to 10% for CC events? Does the problem really not exist for anti-neutrinos? What is the right thing to do at high energy? Taper the uncertainty instead of a sharp discontinuity at W = 3 GeV?
GENIE's default FSI (hA) is more and more disfavored by external data (e.g. 2018 results from T2K, MINERvA in "transverse variables"). What do hN predictions look like?
Get events from non-GENIE generators and see what we can learn from them (NEUT, GIBUU, NuWro?)
Does our tune work for Minerva data? T2K data? --> MINERvA/NOvA workshop (Sept 29 2018) will help here
Study correct correlations for MEC systematics nu vs anti-nu
nu/anti-nu cross-section ratio? If it's currently covered, does it stay covered if we reduce other uncertainties like MEC? High energy behavior? Does flat 10% increase in DIS for neutrinos only cause a problem?
clean up code; CAFAna/Systs/XSecSysts.h may need to be split into sub-files; Vars/GenieWeights isn't just GENIE, etc
Look into using MINERvA's soon to-be published (Summer 2018?) numubar CC 1pi result to constrain resonance production. The signal includes resonant and non-resonant pion production. The latter is constrained by deuterium data. The GENIE prediction is larger than the MINERvA measurement. For the NOvA RHC ND numu sample the MC is ~20% higher than the data in the resonance region.
Fix the QE and RES RPA weights in both CAFAna and the MCReweight service to not apply the RPA effect to hydrogen
The code for the cross-section weights in future should be same for CAFAna and MCReweight service.