This is really interesting. I think I had always hoped that the metric might include something other than (or in addition to) pure plays to reflect quality and engagement — like number of tracks copied from a playlist or maybe quantifying the “originality factor” of a list by measuring how high popular tracks were on the “uploaded-first-by” list.
When the MySpacers arrived with their 200-comment emo playlists that always seemed to sit at the top of the popular list, I wondered whether comments had something to do with it. But honestly, it often seemed totally arbitrary, with odd four-dead-track playlists inexplicably bubbling up to the top. I always surprised myself by having even the slightest interest in the list…I suppose that even though it was sometimes a mess, I did find lots of good music through it, exposing things I might not have otherwise found through the other mechanisms in the site.