3 ms·
There's a fuzzy line. I think it's ok (or at least undetectable) if the voters are genuine HN users. There are limits to what they'll upvote, unless they genu
by pg 17y ago
There's a fuzzy line. I think it's ok (or at least undetectable) if the voters are genuine HN users. There are limits to what they'll upvote, unless they genuinely like it. Whereas it's not ok to email a bunch of people and tell them to create HN accounts for the purpose of upvoting your post; that's basically sockpuppets by proxy. The threshold is somewhere between these two cases.
There is a lot of code for detecting voting rings, and it itself embodies varying levels of punishment (ranging from votes not affecting the position on the page, to votes not affecting the score, to the post getting flagged to be killed) depending on how abusive a vote seems. So far things seem to be under control, probably because the way I decide when to add more countermeasures is when I notice lame posts doing suspiciously well on the frontpage.
- jacquesm 17y agoOk, thanks. So how about a group of people that collude on a back channel in order to categorically boost their posts/comments and flag others? I know that it is annoying to have to define this but apparently it needs to be spelled out. I've found that it is very easy to detect this kind of stuff by simply monitoring a post in the first few minutes after it goes live, anything that gets 3 or more upvotes in the first ten minutes but sinks off the homepage is a good candidate, 3 or more upvotes in the first 3 minutes is more or less a guarantee of having people that game the system.
- pg 17y agoIf they did it frequently enough to create a detectable pattern, the votes would start to be ignored to varying degrees. As you point out (though this wasn't your theme), the ranking algorithm on the front page also helps mitigate the problem. Stuff upvoted by a ring will sink back down if it doesn't catch on with other users.
- tokenadult 17y ago3 or more upvotes in the first 3 minutes is more or less a guarantee of having people that game the system. It could be three or more people all multi-submitting the same article that just got posted to a frequently visited site. Each may not know that anyone else just submitted the same link to HN, if they are simultaneous enough. (It has happened to me.)
- 3pt14159 17y agoI only up vote a friends article if I like it enough to submit it myself (not just normally vote for it). I think that should be the test.
- swombat 17y agoWell put. For the record I regularly upvote things passed to me by other HNer friends (some of them YC alums), but I always read it first and make sure it's something I'm comfortable upvoting. And I occasionally bring articles (both my own and other people's) to those same friends' attention. The ultimate decider has to be whether the article is good and worthy of an upvote. I think I have mistakenly upvoted something that was actually bad maybe 3 times in the two years I've been here, and each time I felt really bad about it. Basically, if you wouldn't happily submit it yourself, don't upvote it.