indyank - 6:33 am on Jun 5, 2011 (gmt 0) [edited by: indyank at 6:58 am (utc) on Jun 5, 2011]
As regards winners, it is clear that panda is not run for all niches.The research paper of Biswanath Panda also makes references to the intensity of the process meaning the resources that these panda runs consume are huge.
They have probably excluded certain niches (examples are niches on which there are not many as in the case of netmeg), certain platforms (examples are youtube, blogspot etc.) and several other sites.
However, I believe that panda is a run on sites picked using some criteria (or picked randomly) than making explicit exclusions (though there could be exceptions to it). Those that have not been hit are probably sites that have never been tested by panda.
Two sites i.e. DI and cultofmac were initially picked by panda but were dropped later on.It does suggest that they had signals for panda to trap them but they probably removed those signals later on. It could have been done with some help, though they may not obviously share them, without requiring google to make exceptions for them.
[edited by: indyank at 6:58 am (utc) on Jun 5, 2011]