Ok, I'll bite...
In the article it mentioned that the tutor in question had a greater than 90% success rate.
What's the control group success rate? 95%? 85%?
If I start an SAT tutoring group and manage to limit myself (let's say, through pricing) to a group that is already quite likely to succeed, than I have effectively priced myself into a track record.
As long as the stats already look impressive, it should be relatively easy to ignore that part of the question, as long as I am not actively harming the chances of my students.
After that, it becomes a question of the intangibles mentioned in the grandparent post: look, dress, body language, etc.
I agree, however, that it's not like faking wine, but for the opposite reason. I would bet that it's much easier to uncover the wine fraud, since there will be some sort of paper trail, whereas there will never be a control group study between your SAT tutoring and no intervention at all.