[gregbradley]

How important is dithering? Data rejection in CCDstack works miracles and mounts are never perfect especially lower end ones.
median combine takes care of the rest.

Taking enough subs so the statistical averaging processes of various combine methods can do their magic is a good approach.
Its really a maths problem. To accurately identify an outlier you need to have accuarately established the bell curve of the data to see what is odd by sitting outside of it. That's pretty much how data rejection works and dither works. The odd value gets averaged out in the case of dither or some other method of determining odd values that should be taken out of the final stack.

Perhaps an example of a dithered stack of 10 versus a non dithered stack of 10 may be more conclusive but I would suggest the difference is negligible. Perhaps I am wrong and it makes a massive difference, in which case I'd love to see examples.

So worrying about something that isn't really an issue at all may prevent you from going ahead with a tandem setup and all its benefits not the least of which may be extra subs that can further increase the accuracy of your data rejection routines.

Greg.