Fredvallee the pixel shift might come partially from the natural field rotation.
Good idea.
Allowing pixel shifts is really effective with Bayer color cameras even when not using the Drizzle algorithm. It gets the star colors right, instead of having overly saturated red, green or blue stars when you are perfectly tracking with an undersampled system. I remember depending on that years ago when not autoguiding.
However, be mindful that a mount that with the ZWO case, because the pixel shifts come from a field rotation, some pixels will be shifted more than others. The field will rotate around the guide star (not the center of the FOV), if they are using multi-star guiding, the field will rotate around the averaged centroid -- I am assuming that ZWO autoguides using the actual 10-second exposures to keep images framed and centered.
Albeit, the pixel will also be shifting during a 10 second exposure itself -- using the "rule of 500", a 250mm focal length will already start to show oblonged stars even with a 2-second exposure. (I really don't know how they get away with 10 second exposures that are untracked; perhaps because the sensor is so tiny compared to the full frame case of the "rule of 500?" I suspect the scope will show lots of non-round stars, the biggest sin of astrophotography.)
(My All-sky camera with an 8mm lens and a four-thirds camera sensor is severely undersampled, and the comparatively short 3-second exposures at night show very little star trailing, with some abnormally colored stars :-).
Chen