Kenny is correct guys,
Try not to hassle your selves with calculations and numbers..
What happens is as you expose the CCD to light, the bell curve starts to shift to the right. The more data you collect (Photons) the more to the right it moves. You need to position your histogram window around this bell shape.
All you need to do is start with a reasonable setting, like gain 50 and offsett 135.
Take an exposure and see how your histogram looks. Like Kenny said, you should see the whole bell curve within the hstogram. If you see an all black image an no bell curve, then you need to change the offset to bring the curve within the histogram.
Once you have done this a few times, you automatically would know the values next time you do imaging or change exposure times.
For example, i shoot with 63 gain and 135 offset for between 1 sec to 3 minute exposures thru my 14" RCX400, but if im going to increase my time to 5 minutes, or im going to use my 22" f4 Newt, then i know that i need to offsett a little more to take the extra signal into account. So i set the gain to 63 but offsett to 120 or 125 and sometimes 115 depending on sky glow.
Where other cameras would have just saturated the image with white, the offsett ability with some ccd cameras allows you to avoid this and capture the image cleanly.
Basically spend 5 minutes getting used to offsett and gain one night to grasp the basics. I use the same settings for 90 % of my exposures.
Only when i bin the image do i use a lower offsett like 115 etc.
These values are not FIXED and so documenting it as such would cause problems to users that dont have the exact same optical train as the user who set the values. Offsett changes with you lens/mirror size and f ratio.
Oh yeah, here is the link to that pdf that Kenny mentioned
http://www.lightstorm.at/pdfs/QHY8_settings.pdf .
Theo