Atmospheric distortion makes the star positions waver a lot (twinkling or scintillation).
There should be times where the atmosphere calms down (very briefly) and you get a good clean undistorted image.
Do you ever then just use those few frames where the stars seem to be in their correct (or nearly correct) and undistorted positions and then just add those up?
I imagine that you would have to throw away a lot of frames to make this work but you would get a cleaner image at the end?
There are rare times the atmosphere settles down. What usually happens with planetary imaging is that perhaps 50% of image frames are used in the final image. With deep sky images generally, you must use longer exposures so the images are always slightly blurred and so you use most of them.
With the next generation or 2 of camera sensor's, it may be possible to nearly eliminate read noise so we will be able to use very short exposures. Already there are such devices that called EMCCD but they are very expensive.