3 ms·
Would it be possible to apply this research to the new page in order to reduce or eliminate information cascades? http://www.overcomingbias.com/2007/04/popular
by Alex3917 16y ago
Would it be possible to apply this research to the new page in order to reduce or eliminate information cascades?
http://www.overcomingbias.com/2007/04/popularity_is_r.html http://www.overcomingbias.com/2007/04/popularity_is_r.html
If we split the new page into three or four parallel universes then it seems like that could go a long way to improving the quality of what hits the front page.
- jacquesm 16y agoThat would likely backfire, already a lot of stuff that is good or even great does not get enough votes to make the homepage. Splitting the new page in to multiple segments would make that worse. I can see the case for randomizing the entries on the newpage, and maybe making it a bit longer to offset the variation in traffic on the site at different times. Happy to see PG work on more voting ring busting, there is definitely still quite a bit of that happening, the new page sometimes shows pretty weird things.
- Alex3917 16y ago"Already a lot of stuff that is good or even great does not get enough votes to make the homepage." The reason this happens is because stuff scrolls by very quickly, and people are most likely to click on the stories that already have votes. And the first stories to get votes are always the ones from outlets that people are familiar with, like Mashable and TechCrunch. If you split the new page into three universes then all of the stories that would normally get free votes have their vote count divided by three, whereas the other stories do about the same, which is comparatively better than they'd do otherwise. At least this is my thinking. And then you just add up all the votes from all the new pages to determine the front page.