30 May 2011

Vandalism, second chances, and bots

Three interesting observations about the English Wikipedia:
  1. Almost half of all the text added gets reverted.
  2. Over three quarters of contributions from registered users are from someone who's had a contribution reverted.
  3. Half of the top 50 contributors (by amount contributed) are bots.
(This measures the size of contribution as the number of bits of entropy of a revision when compressed with a model primed with the prior five unreverted revisions of the page. It's imperfect, especially for large deletions that don't cause a page to get reverted, but it's a first approximation.)