Could any astrometrist kindly tell me from their experience how much a nova's position can change between when it is caught in outburst, and when it has later faded. The assumption is that the astrometry is good even when the object is bright, but when it is fainter the image is smaller and errors [like overexposure and/or finding the photocentre] are less. Is this true, and if so does the difference usually add up to an arcsec, several arcsecs, or whatever? Cheers John