CCDSoft should be taking the guide images using auto-darkframe subtraction, so you shouldn't have hot pixels in the guide image (within reason). Not sure what's going on there - perhaps there's a checkbox somewhere to turn it on/off that I haven't used for so long I've fogotten? It is normal that CCDSoft would do it's own auto contrast/histogram for the autoguider exposures and you can't change that (that I'm aware of) but I've never had problems with it.
Sounds like you need a brighter guide star anyway (and yes, it's often a challenge to find one - I was using 10 second exposures for guide stars for NGC 1232 the other night - colour filters were extending guide exposures considerably).
I've recently discovered myself that what's referred to Luminance is actually IR cut. I'm surprised and don't understand why. I find a large difference in detail resolved when taking an image through the "L" filter vs just clear (no filter), presumably because the ST7 picks up so much IR ... might be the same for the STL11000.. don't know. I'm yet to experiment to see what difference using L vs clear has on RGB images.
Btw.. nice start
Working at 2x2 sounds like a safe way to go, and quite tollerable considering the megapixels of the CCD. If I bin 2x2 I end up with a tiny image
Roger.