zlacker

>>Robin8+(OP)
> The assumption is that a Story cannot go from the top 30 to a position higher than 90 in a single minute, without having been explicitly removed.

That's wrong. Both the flamewar detector (a.k.a. the overheated discussion detector) and user flags do that, and there are other software mechanisms that do it too. For example, if a story has been on the front page for more than (IIRC) 18 hours, it gets an automatic downweight unless we manually override it.

Also, keep in mind that user flags affect a submission's rank long before the [flagged] marker appears.

>>dang+wg
How do you keep user flags from being used as a way to squash articles on a particular topic before they have had the chance to be exposed to the wider HN community?

Meaning if someone were to theoretically get a real time feed of HN submissions, and flagged articles that they didn't want seen as well as messaging a group of friends to do the same thing. Do you have protections for this type of behavior that would prevent this person from having undue influence on what can and cannot have a chance at being seen by others?

>>nathan+Av
What you’re asking about is referred to as “voting ring detection” and it’s something social networks keep very secret.

>>kortil+IA
I don't have any proprietary knowledge of how HN does voting ring detection, but to offer an intuition about how it might work in the scenario proposed by GP, this voting ring would be detectable because their flags are highly correlated and clustered together in time. The more stories they attempt to flag down (successfully or otherwise), the more obvious the pattern will be.

I'm sure you're already thinking of ways to bypass that, and yes what you're thinking will probably work, it's a game of cat and mouse and no one technique will be sufficient or work forever. (See also https://en.wikipedia.org/wiki/Defense_in_depth_(computing) )