CBI TTT

• get CBI source-finding mocks working

• rerun isos with Larry's source finding scripts for veto, instead of Jon's cruder method (JS or BSM)
  • rerun iso this with for the not-deeps
  • This all requires the GBT vs OVRO noise to be understood and believed
  • get the statistics in the tail of the iso distribution adequately

• get fully marginalized significance of excess from MCMC chains. requires some tweaking and a new chain run.

• CBI no WMAP EE run

• put in Tim comments

• iso PS
• thermal noise PS
• fix likelihoods? but this doesn't include the iso

• some other test for pointing per se -- jon's split; new per night mock with scaled RMS to allow for deck angle.

• Definitive splits (JS)
  • channel splits (something went wrong with this)

• assessment of SZ template dependence?

GBT/OVRO

• FIGURE: ISO distribution
• FIGURE: SI distribution and/or flux ratio distribution
• FIGURE: SGBT Obs and noise fit -- perhaps the noise problem is in GBT?

• FIGURE: SGBT vs SOVRO (detected flux scatterplot ; noise histogram)
• WORDS: CBI map source finding.
• WORDS: maximum likelihood analysis for SI or flux ratio.
• WORDS: translation from ML to iso distr.

• Refine GBT Noise estimate
  • write out files with individual noises -- compare to OVRO, run chain
  • vary boxcar size
  • compare medabsdev to RMS

• rerun bright/faint sources separately with +/- 20% GBT noise
  • done, suggestive but not conclusive
  • compare +/-20% noise iso distr, with best iso distr
  • rerun with +/- 40% noise?