I don't know what counts as reliable, it works most of the time.
You can check for cosmetic filtering (CSS reapplied to your ad-divs) by checking their width or something like that. You can check for blocked domains by getting this information channeled back somehow.
There are only so much relevant filter lists (maybe 15?) out there. They are freely accessible & parsable, so generating an URL on the fly which should be blocked on your domain in one of those filterlists shouldn't be too hard.
Hmm. I can't imagine how.
Maybe if the Browser would simulate a pageview without adblock but somehow manages to hide the ad page elements. I'm sure that might break pages and (since resources can't be blocked) might reduce the use of adblockers drastically.