I always assumed this site [...] tries to avoid driving readers to any particular conclusion, but simply away from misguided ones[.]

As a larger point, separate from the context of lukeprog's particular post:

What you assumed above will not always be possible. If models M0...Mn are all misguided, and M(n+1) isn't, driving readers away from misguided models necessarily drives them to one particular conclusion, M(n+1).