3 ms·
One dimension of the problem is that misleading, attention-whoring headlines work. (Maybe the sites even like these, because they generate traffic in discussion
by gojomo 19y ago
One dimension of the problem is that misleading, attention-whoring headlines work. (Maybe the sites even like these, because they generate traffic in discussion threads.)
Headlines could be rated separately from the link; perhaps a smaller group of trusted editors, or a random panel of unassociated readers, could approve/edit headlines. Further, votes without having visited the target article could be deweighted.
Another dimension is vote-gaming, including outright fraud. Increasing the cost of votes, with captchas or even registration/poll fees, might help.
Actually making vote-mobs explicit, as in the 'Uberfact' idea [+], I also think has potential. That is, rather than making recommendations/correlations for every user account, let users create their own named cliques. (Let the founders of such cliques have ultimate say on who else can join as voters.) Then, every article gets a different score per clique, and you can view the 'top' items from any clique's (or combination of cliques') perspective, whether you share their views or not.
Some cliques could be ideological, others by membership in some external community, others purely automated (perhaps mirroring some external scoring, like PageRank or Digg-score). Then you channel a lot of the gaming into a more productive and filterable expression of passion; the default view of the site might degrade, but even slightly savvy users can extract a elite, useful view.
[+] http://unqualified-reservations.blogspot.com/2007/08/uberfact-ultimate-social-verifier.html http://unqualified-reservations.blogspot.com/2007/08/uberfac...