scooterdude - 9:59 pm on Dec 14, 2012 (gmt 0)

Ok i'll throw my outlandish theory into the mix

Starting with Panda Google has been manually creating the "seed" set of websites others have referred to , but, they've also assembled a fairly massive list of webmasters, linked those webmasters to their attributable output, via gwmt, crawling tools, manual sleuthing, assigned ranking scores +ve and -ve, then apply this to the output of the other Algos

How about webmasters not "identified" you ask, well, in writing an Algo, the +ve and -ve bias would probable have reducted impact for queries where none of the identified have a presence,

I think they've been telling us this for a while , who was talking about brands being the answer to a certain open sewer,,,