Compare these two posts:
http://dearauthor.com/wordpress/2009/04/12/amazon-possibly-using-category-metadata-to-filter-rankings/ http://community.livejournal.com/brutal_honesty/3168992.html Even if this is true, it doesn't absolve Amazon. This is not an exploit of a bug, it's an exploit of a policy decision. This is not a software glitch -- the software is doing exactly what it was designed to do. It was designed to allow exactly this sort of thing, intentionally, because the people making the policy decisions asking for this feature (deranking) didn't think through how it could be exploited.
ETA: I find the policy decision of hiding things from me that Amazon thinks are inappropriate for me to see incredibly objectionable, if that wasn't clear. The fact that said policy decision is also easily exploitable is just icing on the cake.
ETA2: More on the specifics of this particular claim, and the likelyhood of it being a metatroll:
http://bryant.livejournal.com/672165.html That said, I still think it has signs of being some sort of exploit.
ETA3: What they said:
http://tech.blog.extendance.com/2009/04/13/amazonfail-a-architechtural/