1. 7-8 years is probably accurate for the MD/PhD portion of your training.
8 +/- 1 year is my experience. If anything, the quantitative PhDs graduated faster in my experience. See my blog entry:
http://www.neuronix.org/2011/09/meeting-about-return-this-past-week-i.html
After that, if you are not interested in practicing, you would not need to complete residency or fellowship, which are only needed if you want to practice. Most MD/PhDs end up doing a residency, but there are always some who are no longer interested in practicing and head right for a post-doc, faculty position, or industry.
I strongly recommend anyone complete a residency because, well to copy from another recent thread: Remember, the PhD job market is TERRIBLE. You do not want to give up your clinical medical abilities. If you do this, you are in the same place as the PhDs. Your MD does not mean anything without board certification/eligibility. But, being a practicing or eligible to practice physician you are much more valuable, including on the research side.
TL;DR: my post-doc advisor's advice (a 100% research MD/PhD): "Doing an MD/PhD program without completing a residency is ****ing stupid."
The number who don't enter residency from an MD/PhD program is well less than 10%. I have the percentage who went to "post-doc" as 4% from an old thread based on a PDF that I can no longer find (
http://forums.studentdoctor.net/showpost.php?p=1429866&postcount=26). That does not count graduates to industry, however I can speak for the one MD/PhD I know who went to industry instead of post-doc, and he was listed on our match list as post-doc. Graduates of the program I graduated from are heavily dissuaded from going straight from the program to industry, as it is generally assumed that one needs additional training after graduate school to become a principal investigator, which is generally the research goal of the program.
2. You would still have to take all premedical prerequisites in order to be considered for medical school or MD/PhD. Very few schools, if any, waive these requirements, even for exceptionally well-qualified candidates.
Agreed. A master's degree means nothing on its own. I generally don't recommend people do them at all if they are interested in MD/PhD. You need to focus on getting the med school pre-reqs and getting straight As in them while pursuing additional research. Master's programs include a lot of advanced coursework, teaching, and thesis writing that generally are not necessary for MD/PhD programs. This doesn't even include the cost you don't want to incur this early before a long training process. Is there some reason I'm missing about why you want the master's degree?