cc blindanalysis ppt BCC CC group Perspectives on Blind Analyses Open order by MikeJenny


More Info
									  CC group Perspectives on Blind Analyses
                                                               D.A. Petyt        June 2004
• What should we blind?
  • The far detector energy spectrum is
    our signal. We should apply a
    blinding procedure to ensure that we
    do not sub-consciously tune
    ourselves on an “oscillation signal”
      • We are not just validating
         Super-K – there may be other
         signals out there…
• How should we blind?
    – Hold back an unknown fraction of events based on:
         • Gary’s f(Evis) – could be linear, sinusoidal…
         • Jenny’s flat Evis (or event length) model
• What do we do with the un-blinded events?
    – Check that there are no obvious peculiarities in the data
         • Function must not introduce discontinuities/spikes in the un-blinded events.
• What should we not blind?
    – The near detector data. Needs to be completely open in order to understand properties
      of the beam and of the events themselves
                When to open the box?
• We are starting with the low
  energy beam because we
  believe Dm2 lies in the S-K
  allowed region
• If Dm2 is significantly higher
  (>0.005 eV2), a higher energy
  beam is preferred for all physics
• We will have enough data to
  make this check in the first few
  months of running.
• Should plan to open the box
  after 3-6 months and look at the
• We can then close the box and
• A short me/he run early on will
  be useful to check high energy
  region, even if Dm2 is low

To top