Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Well unfortunately, in the large debates, really many duplicates are being submitted. Take the UBI one with close to 3k unique claims. It's really hard to suggest claims that aren't already in there.

If you have specific examples, please point us to them, but we are not aware of any really large (>500 claims) biased debates.

The idea is for claims to be canonical for each debate, otherwise you have to rewrite them a gazillion times, including their children. That's what we have linking for.

Generally speaking, if you think your claim was unfairly rejected, just hit report and we'll look into it.

Having claims rejected as duplicates, happens to us too. :)

If you have any suggestions, other than reporting, and roaming powerusers (which we will introduce soon), please let us know!



Have you considered democratizing moderation? If many people agree that a claim is a duplicate of another claim then it's probably a duplicate or at least the community wants to treat it as such.

Also instead of removal just make it less visible. With a big arrow pointing towards the main discussion. If people still choose to discuss there then they must see value in doing so.

PS: thank you for making this


Regarding democratizing, the problem are the troll hordes out there. One person with many followers tweets out a particular claim and then you'll have the strangest things suddenly appearing at the top. It's the same reason Internet polls don't really work. And the reason we consider the default perspective ALL votes, problematic (we are building something here soon). This is especially a problem due to our topics, primarily very controversal issues. These are the type of topics that very often get edit-locked on Wikipedia because of edit wars and we deal exlusively with them. That's why we built the suggestion system (https://support.kialo.com/hc/en-us/articles/115003791445-Sug...) and 5 permission levels (https://support.kialo.com/hc/en-us/articles/115003877705-Par...).

Suggested claims that were not accepted, aren't deleted. users often argue back and forth with admins on those claims, often edit them and then get them accepted. Very often it's really just a clarification or "could you please add sources" or "do you mind moving this claim over there".

Duplicates are thornier. What we have considered, is linking them to the ones that they supposedly are duplicates of, so you could see them there.

That's possible and we might do it, but we really don't get that many complaints about them. It's really just a problem in the debates with >1k claims. but as more debates grow, this might become more of an issue and then we'll build that.

Thanks for your input.


Thanks, it's interesting to learn what is a practical problem when growing a discussion website.

Is it possible to deter troll behavior by using a reputation system and investigating bad behavior? Reputation that takes a long time of good behavior to build but a single act of malice to destroy.

It also has the statistical beauty of adding a prior probability estimate to the likelihood estimate that sorts Reddit comments.


Agreed.

We have a "reputation system" it's the thank count next to usernames. We have not handed out any privileges based on those yet, but will probably soon take all with a certain thank count, look at their contributions and grant them some poweruser rights, e.g. commenting into all public debates and see how that goes.


delegated democracy (page rank) is pretty ideal but complex


Are there simpler heuristics than page rank? Perhaps a reputation score determines how much your vote is worth?


page rank is a reputation score


Yes, but if computational cost is a concern there are other options.


I fear that's too easy to game at our scale, given the horde out there.


page rank in this context is morally relativistic


How would you apply it in our context?


Not OP. In page rank you have a graph of pages as vertices and links as edges. In this case you'd want users as vertices and up votes as edges.

Page rank would give users a high score if they are up voted by other users with high scores.

I'm not sure if page rank can deal with negative weight edges for down votes but probably.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: