Ok this thread is pissing me off! Not because of all the opinions, which are great, but forgive me because I am a data analyst and am looking at this question as a very complicated one.
Pick one straintesting website, say StrainGeniuscom. If they have 5 samples of say, White Widow sent in in the last 6 months from different locations with a THC count of 17.5% to 20%, you take that average for that strain at that time.
The averages get more and more accurate with more batches being tested. They have the metric for this, I believe, a lot of these sites like StrainGenius.
Then you kind of know what the average is of the normal grower, factor in what the manufacturer, say Sensiseeds says their plant is at (it's their average). Let's say their average White Widow is 19%, then you can start gauging where the growers are going wrong or right from there.
So.. you need a "control". One from an average of a strain off of straingenius or other such testing website. Then you need the "control" that the seedbank lists. See how far off it is, and figure from there.
Remember: more numbers of tests the more accurate. Try to get 50+ straingenius reports over a year period, or something. Averaged, you will know what the average grower should expect. :dance