Tim Starling, 21/10/2011 02:29:
There's no bias towards or away from porn, however. The distributions of page_random gaps are independent of any variable you might want to study, like quality or age.
If you try to get a lot of random pages from Special:Random, eventually you will notice that some pages are missing and some pages come up more often. But if you are only fetching a small fraction of the total number of articles, then the statistics of the returned sample will look more or less the same as a true random sample.
So this is a problem only for users who (inefficiently) visit Special:Random thousands of times (perhaps millions combined) looking for (untagged) things to be fixed. Poor abandoned less popular pages, what a sad and unjust condition. :-(
Nemo