New page with info on the DT1 live "voting": https://bitcointalk.org/index.php?action=trust;dtview


All of the "DT1s" would also have to be of Member rank, which isn't trivial.

Anyway, there are many conceivable ways to abuse the system, but if it happens, you can just shoot me a PM and I'll fix it, probably in less than 24 hours. To do it in a way that's non-obvious, where I wouldn't blacklist the accounts, would require tons of time and sMerit, and is unlikely. I'm inclined to let the current criteria go for a while and see if "unknowns" actually start habitually getting into it, in which case changes would be needed.


It doesn't. The algorithm would never work that way.

Once there's more than 100 users who would be selected to DT1, a random subset of those eligible will be selected each month. Or that's my current thinking, at least.

This will help prevent people from being unfairly attacked by even a majority of DT1, since some month you might lose the dice roll and find yourself in the minority for a while. As I mentioned in the OP, it creates more of a credible threat of retaliation, which I believe will have a moderating effect overall.


Yes, that's my intention. Once someone thinks, "DT is basically OK, but it's a bit chaotic, and every now and then I see something that I disagree with," then they've graduated from trust-newbie and are ready to build their own list. The old system was too safe/static/universal/untouchable.

(But I don't want DT to be absolute garbage, either, since trust-newbies need some guidance.)

If you ignore the "DT1 voting":
 - ~ing everyone who trusts user X without ~ing X himself often won't result in X being excluded from trust networks.
 - ~ing X himself without ~ing anyone who trusts X might result in X being excluded from trust networks.
 - ~ing X himself and everyone who trusts him might result in X being excluded from trust networks, but to the exact same degree as the above option. It doesn't function as some sort of "super-exclusion".

If you take DT1 voting into account, then there's one exception: if you're on DT1 and your exclusion causes someone to drop off of DT1, then this could help you in your goal.

I do not view it as appropriate for trust ratings to relate primarily to non-trust matters. By giving someone negative trust, you're basically attaching a note to all of their posts telling people "warning: do not trade with this person!". If we can get DT working well enough, in the future I'd like to prevent guests from even viewing topics by negative-trust users in trust-enabled sections, so you have to ask yourself whether your negative trust would warrant this sort of significant effect.

In particular, in my view:
 - Giving negative trust for being an annoying poster is inappropriate, since this has nothing to do with their trustworthiness. If they're disrupting discussion or never adding anything, then that's something for moderators to deal with, and you should report their posts and/or complain in Meta about it.
 - Giving negative trust for merit trading and deceptive alt-account use may be appropriate, but you should use a light touch so that people don't feel paranoid.
 - You should be willing to forgive past mistakes if the person seems unlikely to do it again.
 - It is absolutely not appropriate to give someone negative trust because you disagree with them. I'm disappointed in the reaction to this post. Although H8bussesNbicycles is perhaps not particularly trustworthy for other reasons, the reasons many people gave for neg-trusting him are inappropriate. You can argue that what he's advocating is bad on a utilitarian level, but he would disagree, and his advocacy of a certain Trust philosophy doesn't by itself mean that he's an untrustworthy person. DT selection is meant to be affected by user lists, and it is totally legitimate to try to honestly convince other (real) people to use a list more in-line with your views.
 
I'm not going to blacklist people from DT selection due to not following my views, since a big point of this new system is to get me less involved, but if a culture somewhat compatible with my views does not eventually develop, then I will consider this more freeform DT selection to be a failure, and I'll probably get rid of it in favor of enforcing custom trust lists.

~snip~

If the "gaming" takes the form of strategically sending a lot of merit, creating sockuppets, and stuff like that, then no. That sort of gaming might get me to blacklist people, in fact. But if it looks more like politics, then that's OK, and that's what H8bussesNbicycles's thread looks like to me.

Not anymore. That's clear abuse, awarding merit for political reasons rather than any idea of quality. Only because he was a source, I effectively undid those merit sends. If he had not been a merit source, I still would've blacklisted anyone who got into DT1 through that type of shenanigans.

I hadn't read into the thread deeply enough to see that stuff. Those are better arguments against the trustworthiness of H8bussesNbicycles & co., but note that the current negative-trust-ratings were sent long before that. Before February, the thread looks like politics to me.


Since some people view account sales as fundamentally untrustworthy, I think it's an appropriate use.


I don't find it unreasonable to have a restrictive selfmod thread. You can guess from the banner & deletion stats that it's going to be a restrictive, single-viewpoint thread. You could always start another topic.

That said, using selfmod topics in a deceptive way can be an appropriate reason for negative trust.

The majority of ratings seem to be warning people about red flags, not punishing provable scams. IMO this isn't a bad thing, since once someone has scammed, it's kind of too late.

What do you think about splitting the scam rating, with a "warning" rating for scammed previously OR you strongly believe that they will scam in the future, and a "scammer" rating for scammed previously AND you strongly believe that they will scam in the future? And then if you only have warning ratings, the indication displayed next to posts will be softer.

---

I think that you and I have a fundamental disagreement on this stuff, though:

Predictability and guidelines are often good. I wrote some Trust guidelines recently, and I may write more. But I don't believe in having a set of hard rules which is to be applied to all cases. Whenever an argument starts looking like it was written by a lawyer, or relying overmuch on precedent, you've stopped thinking about the real case and have started using rules to retreat into moral and intellectual laziness, divorcing yourself from the decision you're about to make. If you're making a decision about a case, then you're responsible for that case, and you can't say, "I don't agree with it, but I was just enforcing the rules." Every case needs to be handled individually.

129 users who were wearing a yobit signature and had at least 1 good report against them in the last 14 days are banned for 14 days. All yobit signatures are wiped. Signatures containing "yobit.net" are banned for 60 days.

Some people were talking about neg-trusting spammers for spamming. This is not appropriate; report the posts, and if that doesn't seem to be working well, come to Meta with specific examples and suggestions.

Logged-out users will now see a warning in trust-enabled sections if more DT members neg-trust the topic starter than positive-trust him.

This increases the responsibility of DT members not to give negative trust for stupid reasons, but only for things that cause you to believe that the person is a scammer.

I agree with this approach. I considered giving merit proportional to activity, but I decided not to because doing so would probably give far more undeserved merit than deserved merit in total. But undoubtedly some people got screwed by this, and if they have decent posts, by all means, give them the 250 or 500 merit that they need to rank-up.

I also agree with the idea of (free) "reviewer" topics in general, for finding high-quality posts that went unnoticed. Barcode_ created one in the Chinese section, as well.

The current status quo seems OK. If you have to pay $20 per merit plus a risk of getting red trust, then that's a situation that doesn't concern me at all. I'm not going to lose sleep over people going to ridiculous lengths to buy merit, since only a small number of people will be willing and able to do that.

As long as merit sales are a black market, I'm happy, since that makes it far more difficult and expensive to buy merit. If that's the case, then the small volume of black-market merit trades don't themselves bother me much, and I think that it does more harm than good to get too witch-hunty about it.

If they complain about amounts, tell them to complain to me. It's best if sources try to exhaust their source allocations, even if it means giving posts higher amounts than is typical. If you have 150 source merit and you only see 3 merit-worthy posts in a month, then I'd rather you over-give each of them 50 merit than let the merit expire. That way there are more people capable of sending merit, and the "merit economy" is less top-down.

If a DT member tags you for something stupid involving merit (ie. probably anything less than selling merit), then they're not going to be a DT member for much longer.

Aside from that, if people complain about whether things deserve merit at all, then that's something to perhaps think about, but if you conclude that they're wrong, then that's that. You don't need to stress about it or defend yourself constantly. It's conceivable that someday you and I will end up disagreeing too much about this stuff and I'll remove your source status, but it's really not a big deal.

The topsendban list is just a first indication of abuse, and many excellent people are on it. Your place on there acts as a sort of benchmark: eg. chandra12 has a similar score there, but whereas you are an extremely active merit-giver with a diverse selection of posts merited (most of which anyone would agree with), chandra12 only has two large merit sends. His behavior in comparison to yours while having a similar topsendban score is what creates a strong abuse impression.

I appreciate the work of you and other sources who take it seriously!