Skip to content

Latest commit

 

History

History
54 lines (47 loc) · 2.5 KB

NOTES_sterile.org

File metadata and controls

54 lines (47 loc) · 2.5 KB

“Good” contours

Take the “95cl_0” contours with >=300 points. See SterileAnalysis/dump_contour_npoints.py.

The sterile (CLs) chain

Parallel block:

Generate toy spectra trees (including dm2ee)

  • How does randomizeDm2ee work?
  • [ ] Need to generate separate (asimov) toy for each point in s22t14/dm2_14 space?
  • [ ] Add true_pars tree (necessary?)

Generate evis/enu matrix

Generate super histograms

Generate PredictedIBD (necessary)? (Note: fit_shape_3d_XXX does in fact dump the summed covmatrix)

Our rewrite of fit_shape_3d does not; we just load the nominal PredictedIBD to be safe (and wrong in the 4nu case)

Parallel block:

Generate osc prob table

Generate covariance matrices (sigsys, bgsys, dm2ee)

Run fit_shape_3d_CLs_3n.C and fit_shape_3d_CLs_4n.c (via run_fit_XXX)

Run fit_shape_sterile.C?

Run toyjobs/CLs/make_data_contours_cls.C

Plot: Run toyjobs/CLs/make_data_contours_comparison.C

Notes on FluxCalculator

CalculateFluxHistRow (idet)

  • iterm < 0 => Map core -> histogram containing fraction of (oscillated, 1/R^2) flux from the core (for each energy bin)
  • iterm >= 0 => Map core -> histogram containing 1/R^2, partially oscillated flux (without energy-independent factor?) (NON-FRACTIONAL)

ExtrapolationFactorRow (idet_far, idet_near)

  • idet_near < 0 => Map core to histgoram with osc prob to a far AD from the core (for each energy bin)
  • idet_near >= 0 => Map core to histogram with extrapolation factors (i.e. far/near including osc, 1/R^2) (per bin)

Notes on OscProbTable

  • This comment is inaccurate: 
    //      fac_extrap = Sum_enubin (M' * fac_far/fac_near)
//                 = Sum_enubin (M * fac_far/fac_near_null)

    There’s a nested inner Sum_enubin (to calculate the norm). See the actual code.

Changes to Predictor

  • Add dm2ee covariance matrix (LoadCovMatrix)
  • Add SetSin22t13Step
  • Calling convention of FixCovMatrix? (PeriodFlag)
  • Likewise for CalculateChi2Cov? Probably not; SetStage is used instead.
  • Add sin22t13-only CalculateChi2Cov?

Other action items

  • [ ] Generate data_file for randomizeDm2ee. XXX Do we want dm2ee or sigsys_w_dm2ee?

Questions

  • In Predictor::MakePrediction, why do we use mapflux_noosc * oscprob? Why is this not equivalent to mapflux? Normalization? Note that OscProbTable seems to essentially “use mapflux”.
  • How does Beda’s Predictor do the MakeSum without the PredictedIBD?

More action items

  • Add to compile.sh (in sterile branch)
  • Keep sterile branch minimal, regularly rebased onto master