aristotle - 7:02 pm on Jul 12, 2011 (gmt 0)

My impression is that Panda tries to measure the intrinsic value of a page. This should be fairly easy to do for pages of very high value, and also for pages of very low value. But that still leaves the middle ground, or what could be called medium value. But there probably isn't a strict cutoff point, such that everything below a certain level is thrown out. Instead there is probably some kind of "quality score", as others have mentioned, which can be factored in with relevance scores, authority and other factors.