On Mon, May 21, 2012 at 11:39 AM, Gwern Branwen gwern0@gmail.com wrote:
On Mon, May 21, 2012 at 8:15 AM, Anthony wikimail@inbox.org wrote:
So, you are not removing random links at all.
.< I should just link XKCD here, but I'll forebear. I am reminded of an anecdote describing a court case involving the draft back in Vietnam, where the plaintiff's lawyer argued that the little cage and balls method was not random and was unfair because the balls on top were much more likely to be selected. The judge asked, "Unfair to *whom*?" Indeed.
---
From the beginning you seem to be under the mistaken impression that I
am trying to defend Wikipedia or defend the current Wikipedia processes or something. I am not. I find your experiment interesting. I think it would be more interesting if your selection of links were truly random, though.
I don't think you should describe your experiment as "removal of 100 random external links by an IP", because your selection was not at all random. I don't say this because I am trying to prove something about the results. I say it because it is a flaw in your methodology.
And I'd note that my methodology, while being quite as random as most methods, carries the usual advantages of determinism: anyone will be able to check whether I did in fact remove only last links which are not official or template-generated in External Link sections, and that I did not simply cherrypick the links that I thought were worst and so least likely to be restored.
How could we do that? You could have just cherrypicked the worst links that were last links which are not official or template-generated in External Link sections. I'm not saying I think you did that. But you certainly could have.
Anyway, the main thing I'd like to say about all of this is simply that your selection is not random. Your sample is biased. Biased in which direction, I don't know. Biased intentionally, I doubt. But your sample is biased.