Associations calculated using the de Ruiter radius are potentially unreliable

[ddobie]

The de Ruiter radius is highly sensitive to positional uncertainty (by design). The current post-processing pipeline potentially underestimates the uncertainties, which causes a breakdown in the association calculations.

The ultimate solution is to accurately estimate the astrometric uncertainties in the post-processing pipeline, but in the meantime there are two solutions.

1. For images that have already been ingested, you can use the basic or advanced association methods with a sensible radius (a few arcsec). This would also work for fresh runs.
2. For fresh runs, you can set the ra_uncertainty and dec_uncertainty parameters to sensible numbers that accurately reflect the systematic uncertainty