Difference: 2018-05-02 (1 vs. 2)

Frank - data systems paper
MSIP alerts review on Friday
validating MSIP stream (choosing the right RB cut)

One possibility for the MSIP alerts is to choose top n alerts, with some SNR cut and RB cut (though evolving)

Nadia: investigating history of past candidates
RB as low as 0.17 for some genuine candidates

PDF version of the paper draft to be posted on the wiki

CCD 4 goes in tomorrow (Zooniverse campaigns)

Charlotte presented a slide showing how there are more reals than bogus events with bright stars nearby that are called a transient by the RB system.

Bright stars do not feature in reference catalog, so their distnr (as well as magnr) get -999 and that seems to be a problem.
Nadia: we can crossmatch to a brighter star catalog. Perhaps Tycho that goes to 14,
That will add a meta-data field, but will be a good solution.

Another slide showed how difference images of galaxies get faint positive residuals that are difficult to classify as real or an artifact of differencing. 

A different stretch to be tried with Zooniverse images to make it simpler to tell the artifacts apart.

GAIA cross-matching may be useful at the bright end where Pan-STARSS has many merger objects leading to an SG of 0.5.

5 minutes of Zooniverse a day can shoo the boguses away.

-- AshishMahabal - 02 May 2018