On the evening of July 10th 1999 I imaged NGC 6411 which happens to contain a mag 16.25 Supernova, 1999da. The magnitude was calculated using MaximDL based on the 12.3 mag star to the lower left of the supernova.
The image was taken from my backyard in San Jose, California with my Meade 12" LX200 telescope and my SBIG ST7E CCD camera. This image is the sum of 10 one-minute exposures. Dark frames were subtracted but I did not use flat frames, which would have made the image a little cleaner.
Interestingly, I had to update 1999da's magnitude estimate twice. First, Maxim calculated a magnitude of 14.4 based on the jpeg file I created from the master FITS file that was the sum of the 10 one-minute exposures. After realizing that my estimate was two magnitudes brighter than what the Supernova Website was saying it should be, I tried Maxim on the master FITS file. There it thought 1999da was mag 15.4. Then I went back to one of my original 10 not yet merged images and tried this again. Now 1999da showed a magnitude that made sense: 16.25.
So, my guess is that flat fielding is extremely important when trying to get accurate astrometric results. In the case where I had just a single image 1999da's mag was probably just 0.10 mag too bright, while one I added the 10 images together, the resultant magnitude was about 1 magnitude too bright. When I get the time to create a flat frame I'll adjust my images with it to test this theory.