On Sun, May 20, 2012 at 7:31 PM, Gwern Branwen <gwern0(a)gmail.com> wrote:
On Sun, May 20, 2012 at 6:09 PM, David Levy
<lifeisunfair(a)gmail.com> wrote:
Yes, there is. Your methodology has been
challenged
I don't recall any challenges
You haven't gone over your methodology. I highly doubt you've
selected the links randomly. And you don't seem to have done any
analysis of whether or not the links should be there or not.
That was my point "what percentage of the links were actually good in
the first place". Not to try to rationalize results which you hadn't
already presented, despite what you think.
On Sun, May 20, 2012 at 6:22 PM, Anthony
<wikimail(a)inbox.org> wrote:
Removing 100 random external links? For a few
weeks? Then adding
back the ones that deserve to be added back?
I think it's less questionable to just re-add all the links, no
questions asked about 'deserving'.
I have no idea which way would be less "questionable", nor even what
that is supposed to mean. But the right way to do it is to only
re-add links which should be added back.