Tuning page for 2017 fits

CCQE

1pi

• Callum: GENIE dials
  • Phil and Callum used
    • kXSecTwkDial_NormCCRES
    • kXSecTwkDial_MaCCRES
    • kXSecTwkDial_RvpCC1pi
    • kXSecTwkDial_RvnCC1pi

• Clarence: NEUT dials and good data-sets
  • NEUT dials
    • MANFFRES
    • CA5
    • I½
  • Could include Minoo's model? Maybe a little crazy and should be left as another paper
    • Minoo and I are likely to do a data/MC + tuning paper for her model which we will label as "made by team NUISANCE".

• Everyone: Think about data-sets and tell Clarence he's wrong

• Patrick: Beer

General idea:

• Usual sigma(Enu) and Nevt/dQ2
• Use different data-sets (W < 1.4, W < 2.0, ANL only, BNL only)
• Use kitchen-sink data including the kinematic distributions in shape only

---

Patrick suggestions:

1. ANL/BNL CC1pi+p W < 1.4 GeV : Enu+Q2 : MaRES+CA5 (already done, this is the NuFact? fit)
2. ANL/BNL CC1pi+p W < 2.0 GeV : Enu+Q2: MaRES+CA5
3. ANL/BNL All Channels W < 2.0 GeV : Enu+Q2 : MaRES+CA5
4. ANL/BNL All Channels W < 2.0 GeV : Enu + Q2 : MaRES + CA5 + I1/2
5. ANL/BNL All Channels W < 2.0 GeV : All datasets (minuts phi and adler) : MaRES + CA5 + I1/2
6. ANL W < 1.4 GeV only fit all distributions : All Params
7. BNL W < 1.4 GeV only fit all distributions : All Params
8. Effect of choice of BNL flux on fits. (a bit more awkward to do as it requires extra samples to be generated for all generators. Choose just one generator? GENIE?)
9. FNAL Only Fit : All Params
10. BEBC Only Fit : All Params
11. GGM Only Fit : All Params
12. Kitchen Sink Fit : All channels and experiments : All Params
13. MiniBooNE 2D PCosMu : All Params
14. Final suggested parameter propogation. Do we want to end up with a set of 'approved' parameters for each generator. If so we should have a brief section showing their errors for all datasets and maybe also showing them at nuclear targets.

Do we want to consider floating flux normalisations? If so we should float them freely under a common parameter. e.g. float all ANL Enu distributions as one.

Clarence would add in doing a ANL W < 1.4 GeV only, including all distributions. Possibly add in the BNL W < 1.4 GeV too but only available for CC1pi+1p.

Also demonstrate the impact of choosing the BNL flux?

Maybe add in GGM?

Error Bands: Its a bit of a hassle to generate error bands whilst running the tunings. So we should just run all these different jobs and then when writing the paper make the final error bands for any plots we show. Showing error bands for every fit could easily get a bit muddled, especially if we are showing multiple generators on one plot.

Then try MiniBooNE 2D cosmu pmu