I would like the LW codebase to scan for any cases where multiple upvotes/downvotes were made to a single post or comment from the same IP. If any accounts are found which do this frequently, I’d like the results posted publicly.
That’s a weird thing to post publicly—while there’s no explicit guarantee of privacy here, it’s very asymmetric and unexpected, so runs afoul of privacy best practices.
I would not be surprised, for example, if my wife and I end up upvoting a lot of the same things frequently, and don’t see why we should be singled out for it.
ETA: wouldn’t this cause particular problems for, say, the SIAI house?
People should have the opportunity to explain the circumstances. It could be done by contacting people privately, but that would involve work for someone.
Not that I do this, but isn’t there a limit to how much karma an account can add to or remove from other posts? One based on how much karma that account has gotten? So if someone is using two accounts to do multiple up/down votes to the same comment, isn’t this already limited by the fact that they have to earn karma in the second account first?
I would be much more interested in a feature that discounts (but not entirely) the effect of votes when the voter is participating in the thread whose comment they voted on. Such voting is suggestive that the user is doing an “I disagree for the wrong reasons/ I don’t like you” vote, which is disinformative.
This is similar to Slashdot’s system of not allowing you to moderate and post in the same discussion. (I think that goes too far in this direction, but has the right general idea.) It’s also a good policy to adhere to on an individual level, as it prevents you from voting when you’re biased.
Then people would have to choose between voting and commenting. But we often see people complain about people who downvote without commenting. It’s a little rude to downvote without explaining why.
Not that I do this, but isn’t there a limit to how much karma an account can add to or remove from other posts?
There is a limit on downvoting (I think 4 times karma), but not on upvoting. And sockpuppets could upvote each other, giving themselves more power to downvote. Noticing that an account gets all (or most of) its karma from the same IP address could help to guard against this exploit.
Though IP tracking only makes the exploits less convenient. There are proxy services that let you hit the website from multiple different IP addresses.
Okay, that idea makes sense then and I support it.
I would only suggest that it not let users know about mere comments from sockpuppets at the same address, as people often have valid reasons to post as a sockpuppet (i.e. not attach the personal information to their name if their primary account uses their real name). It should only be a tool for restricting sockpuppet karma abuse, not mere posting.
I have no special knowledge about it. I think that there are a lot of people on LW whose stated beliefs should lead them, logically, to cheat to advance their views; therefore it’s not unlikely that at least one of them does so. The harm that could be caused by even one person doing that, multiplied by the probability it is happening, is great enough to justify looking into it.
I would like the LW codebase to scan for any cases where multiple upvotes/downvotes were made to a single post or comment from the same IP. If any accounts are found which do this frequently, I’d like the results posted publicly.
That’s a weird thing to post publicly—while there’s no explicit guarantee of privacy here, it’s very asymmetric and unexpected, so runs afoul of privacy best practices.
I would not be surprised, for example, if my wife and I end up upvoting a lot of the same things frequently, and don’t see why we should be singled out for it.
ETA: wouldn’t this cause particular problems for, say, the SIAI house?
Good point about the SIAI house.
People should have the opportunity to explain the circumstances. It could be done by contacting people privately, but that would involve work for someone.
Not that I do this, but isn’t there a limit to how much karma an account can add to or remove from other posts? One based on how much karma that account has gotten? So if someone is using two accounts to do multiple up/down votes to the same comment, isn’t this already limited by the fact that they have to earn karma in the second account first?
I would be much more interested in a feature that discounts (but not entirely) the effect of votes when the voter is participating in the thread whose comment they voted on. Such voting is suggestive that the user is doing an “I disagree for the wrong reasons/ I don’t like you” vote, which is disinformative.
This is similar to Slashdot’s system of not allowing you to moderate and post in the same discussion. (I think that goes too far in this direction, but has the right general idea.) It’s also a good policy to adhere to on an individual level, as it prevents you from voting when you’re biased.
Then people would have to choose between voting and commenting. But we often see people complain about people who downvote without commenting. It’s a little rude to downvote without explaining why.
There is a limit on downvoting (I think 4 times karma), but not on upvoting. And sockpuppets could upvote each other, giving themselves more power to downvote. Noticing that an account gets all (or most of) its karma from the same IP address could help to guard against this exploit.
Though IP tracking only makes the exploits less convenient. There are proxy services that let you hit the website from multiple different IP addresses.
Okay, that idea makes sense then and I support it.
I would only suggest that it not let users know about mere comments from sockpuppets at the same address, as people often have valid reasons to post as a sockpuppet (i.e. not attach the personal information to their name if their primary account uses their real name). It should only be a tool for restricting sockpuppet karma abuse, not mere posting.
Do you have reason to suspect this is happening often enough to warrant creating a feature about it?
I have no special knowledge about it. I think that there are a lot of people on LW whose stated beliefs should lead them, logically, to cheat to advance their views; therefore it’s not unlikely that at least one of them does so. The harm that could be caused by even one person doing that, multiplied by the probability it is happening, is great enough to justify looking into it.