Just a "counter-intuitive, seems wrong but is true" note on using star count as a quality metric, especially on under-sampled setups.
My FSQ setup runs binned Ha at 4.2"/px. At good sites, with good seeing, good transparency, good focus, and the detector square to the optical axis, the stars that appear on my frames are smaller, there are more of them, and as a result more get rejected as hot pixels by the star detector. And so star count goes down. If I use star count I end up rejecting or de-weighting the best frames. So I don't use it as a quality metric on this setup.
Mike
