Topic: [POLL] DT2 Status; how many net inclusions should it take? (Read 951 times)

Accordingly, Dabs has dropped from 20^th to 37^th on BPIP's "most recognised" list.

Hopefully that won't be necessary.  If theymos is willing to modify the requirements for DT2 according to the recommendations with the most votes, then it's not going to be an issue in the near future.  I hated to have to exclude him myself, I don't know him well enough to specifically distrust his judgement, but his use of the trust system is obviously self-serving. 

Posting my name usually does the job
I've removed him, for now, from my list. If I get to properly read that topic about him and properly check the info, I will consider ~, but not right now.

I should have done this few days ago when LoyceV was taking about inclusion from theymos but now it's done. Considering the overall situation I think for me right now a "~" for Dabs is appropriate.

In 400 years there will be thousands or at-least millions? 🤣

I get the feeling you're missing one small factor; no DT1 inclusions are required for a member to remain on DT1, a net score of 0 is all that's needed.  A member could have the required number of votes and be selected through the lottery process without having a single DT1 inclusion (or exclusions, of course.)

Here's where we are at this month:

https://loyce.club/trust/2022-07-09_Sat_05.08h/54791.html (Week 182) - NOTE non DT1 UID's have been removed to enable context to be seen clearly.


Such are the nuances of the DT system (such as it is) were @BitcoinGirl.Club to remove their DT trust in Dabs, Dabs would then be at DT1 (0) given theymos has seemingly become aware of this issue.

Dabs' only two users they trust that are on DT1 ( @SFR10 and @julerz12 ) are both Known Alts Investigators and it would be interesting to get their thoughts on the whole Dabs DT trusts the very users who then gave positive trust feedback for escrow related services up to the point Dabs seems to have no longer been an escrow in 2018.

The between a rock and a hard place for julerz12 would be that their only DT1 vote is from Dabs.  Should they "distrust" Dabs, then Dabs could (probably) then remove julerz12 from DT1 by removing their trust.

OP: Has anyone tried to reach out to @NeuroticFish, @GreatArkansas or @sheenshane for their thoughts given their trust of Dabs' judgement is why we are here in this thread?

---

---

Interesting also to see be.open ( DT1) and their known alt sne.su both distrust Dabs.   

I voted ‘3 or more DT1 inclusions’ simply because the harder it is to get on default trust, theoretically at least, it lessens the chances of abuse in the trust system.

Getting on default trust by having 1 DT1 inclusion is wide open to abuse & not good enough imo.

Is it the case that a person has to "trust" ten UID's to included in the monthly raffle, or ten UID's that trust them (with 10/250 merit as the case may be) - (or a combination of both)? ^{(Not including the other things like online activity prior to the raffle etc)}

If they're inactive, they won't reach DT1 anymore.
Exclusions don't matter for the DT1-election, only inclusions.

I think you mean DT2 members, those don't have an activity requirement.

Their inclusions can still vote in the DT1-election, but their exclusions only matter to people who have them on their Trust list.

Being active in the past 30 days is a requirement to reach DT1.

---

---

400 years from now, thousands of dead people will include thousands of dead people on their Trust lists. And people on other planets will read back the posts from Bitcoin's early days. LoyceV says hi!

How about inclusion or exclusion of no longer actively DT1 members?

Time can change and scam accusations can be handled or be proven as real scam. So if an exclusion is because of temporary scam accusation, then a DT1 member passed away, what will happen if that scam accusation later will be handled and it will be no longer a scam.

Like Zepher, TMAN, Lauda. Their trust list (inclusion or exclusion) will affect others.

Does theymos exclude inactive members from his trust list (DT1 list)?

Click the link and scroll down: 1640 users earned 100 or more Merits.

No, I check for that. But I don't check if someone was online in the past 3 days, nor if they posted in the past 30 days. And votes are limited, so if for example one user with 500 earned Merits votes for 4 people, only 2 of them actually get a "supervote".
And a few users are blacklisted from DT1.

The threshold to qualify for the lottery is also rather low, I agree.  One battle at a time, however. 


I wouldn't mind seeing the "supervote" go away and have the merit threshold for voting raised to 100.  It's been four years since the introduction of the merit system and I think enough merits have been passed around by now to justify it.


That's actually not as many earners of 250 merits as I would have thought.  Can I trouble you to scrape the number of users who've earned 100 or more?



Since there were 133 members who were included in the lottery, is it safe to assume the remaining 86 haven't built their trust lists?

Just 10 people (with some earned Merit) is enough to put someone on DT1, after which they can create as many DT2-members as they want, or downvote all other DT1-members.

The requirements are 250 Merit for a "supervote", 100 isn't enough.

I never checked, but (using data from a week ago), there are 811 users who've earned at least 250 Merit, and 11,650 users who've earned 10 or more.

I can't answer all your questions, but 3 years ago I made my DT1-election: Rank up pipeline: loyce.club/trust/ranking/ shows that 477 users are included by at least one user with 10 or more earned Merit. Those at the bottom of that list are very far from reaching DT1, but 219 users have enough inclusions from users with enough Merit to qualify for DT1.

I no longer have a database of forum posts, so I cannot answer these questions.

Of the 133 people who are eligible to be on DT1, how many people included them on their trust lists that have 10/100 merit (how many people actually played a role in selecting those eligible to be on DT1)? Of those who potentially can play a role in selecting who is on DT1, how many people have no one on their trust list who is eligible for DT1? How many people who are potentially able to play a role in selecting DT1 have a trust list, but no one on their trust list is eligible for DT1? How many accounts are actually active by various metrics (such as making a post in the last 60 days, logging in during the last 60 days, etc).

As I said, I don't know the answer to the above questions. But I do think the answers will give context to assertations that a "very small minority" of the community selects DT1. The headline number of users is obviously inflated due to spam bots, and inactive users.

It's been almost two weeks since I updated the poll, and so far no one has voted for a net score of 0, which is the current setting.

Suchmoon's suggestion of net 1 with a minimum of 2 inclusions seems to be the crowd favorite.  It is interesting that all who've voted would like DT2 inclusion more restrictive that it currently is.

---

Which of the "Ok" items have not been implemented?  Most of those seem like old features to me, lol.  Several of the "Maybe" items seem to have been implemented as well, and the Bpip extension takes care of at least one.

I'm tempted to resurrect that thread to suggest an onion address for the site.


Lol, I don't use it as such, but it is a handy way to "bookmark" posts to which you'd like to refer at a later date.


I was still in diapers, probably running around in the alts board during those days. 

Right, so while theymos didn't necessarily categorise them based on priority, you can kind of read between the lines, and establish what might be implemented before something else. Ok/maybe obviously being better than the no category.

Obviously, this was back in 2018, and some of these changes have actually been implemented.

Cheers, I tried searching through my merit history with theymos as I was sure I would've merited it. However, didn't.

Also, 2018! Time flies.

Here you go: https://bitcointalksearch.org/topic/m.44432901

I'll have to find it, or rely on someone who's a little bit quicker, but there was a time where theymos gave his opinion on certain features that were suggested over the years, and they categorised them in terms of priority or at least what they somewhat agree with, and don't.

It was a pretty extensive list. So, we got a little insight into their priorities, as well as how their mind works when it comes to implementing things into the forum. I don't specifically recall the trust system being mentioned though. Although, it might have been. Think this post is a few years old at this point.

I missed that topic. Good job for letting him doing it.

What I understand, he has a list of priorities and none of us know what's in the list. Something is very important to us could be less important for him, not saying that changing the net score is something very important right now. It's working but it would be better if the score is higher. Hopefully he also see the importance like us.

I remember when I was asking to change the display name, his response was very quick in PM.

Yeah, it's somewhat bad that so many users are potentially blindly trusting the DefaultTrust. However, 5000 users voting isn't actually to bad. I mean there's a common belief with statistics that at the very minimum 1000 should be in the dataset, which we're above at the least. Obviously, the more, the better that goes without saying.

Although, 5000 users including various different users, isn't a terrible representation. It would be interesting to see how those 5000 trust lists differ. You know, if they aren't changed all that much, and only adding a few users to the list, or excluding some, that would could be more of an issue. However, 5000 users on face value alone, doesn't strike me as terrible. I'm actually surprised it's that many to be honest. Not everyone trades on the forum, and not everyone necessarily needs the trust system.

No, that's not how it works.  No amount of inclusions or exclusions affect my trust network.  Only my inclusions and exclusions have any effect.  If I have my trust system depth set to 1 or (the default) 2, then those who are included by the members I've included will have an effect.  Again, you're missing the point; this proposal has nothing to do with inclusions and exclusions on their own.  It's about those 100 or so members who are on DT1, and the affect their inclusions have on the system as a whole.



Indeed, they have been chosen by a very small minority of the community:


A little over 5,000 users have voted, and those votes affect the forum's trust system for millions of users.  Not my idea of good representation.  Doubling the threshold for DT2 inclusion isn't going to fix all the abuses from which the trust system suffers, but it'll make it a bit harder to abuse.

---

I disagree.  Let's be real, this is theymos' forum and he may disagree with the premise of this thread.  What he chooses to do is purely up to him, but he's been accommodating in the past.  Not long ago I made a stink about all the AWS scammers in the digital goods section, and he responded by banning those sales:

https://bitcointalksearch.org/topic/m.52093309

I voted for third option Net of 1 with a minimum of 2 DT1 Inclusions.
If we are already changing how DT2 members are elected maybe we should think about reducing number of DT1 members as well.
One part of DT1 members could be moved to DT2, and I don't know exact number but I think we do need some reduction, like some members suggested before.
I am not great in math, so I will leave calculations to other people.

No shit; it's definitely not the way it used to be.  I haven't read this thread in full (yet), but IIRC before Theymos made the rotating system change, a member needed two DT1 inclusions in order to be on DT2--and even then, if some other DT1 member(s) decided to exclude you, off DT2 you went.  

I say all of that because that's exactly what happened to me at some point in 2018, right around the time the merit system came out.  Blazed and Hilariousandco included me on their trust lists, but OgNasty and Tomatocage subsequently excluded me.  T'was a very strange period indeed.

I voted for two inclusions, but seeing as how there are so many DT2 members the standards really ought to be higher.  People might forget that even being on DT2 brings with it a lot of power with respect to the weight of your feedback and to a lesser extent whether people will automatically trust you when doing a deal.  It's not a trivial matter--not then, not now.

G'damn, you are the reigning master of statistics and of forum history.  I bow humbly before you, sir LoyceV.

Whatever the net everyone is thinking, it is obvious that a net of 0 does not work properly, very easy to misuse.

This is a very good question. History tells us that there are very little chance for him to do something about it. He must have other priorities 😉

However, those changes would only affect a small number of people who use the auction sub-board; we want to see something more general, and we have had a few suggestions over the years, but none have been implemented. This OP pool is just one of many.

Edit: thanks for the reminder! 

I can: May 9, 2022

I noticed theymos only contribute to a topic he likes/support the idea.

Well, i can't recall the last time theymos implemented anything on this forum; I believe he's been preoccupied with the new forum software, I could be wrong, but I believe that most of our suggestions will be implemented in the new forum. theymos appears to be only concerned with the weekly DT elections and the yearly merits source appointment. All other suggestions have been like throwing water on the back of a fowl.

Do you think there is any chance that theymos will implement a change in this regard?

At the moment the most voted option is: "Net of 1, but with a minimum of 2 DT1 Inclusions"

From what I have seen in all the threads that make proposals for changes to the forum, in the end nothing is implemented.

IEEE computer floating-point arithmetic also has a negative zero but nobody with their fancy compiler optimizations respects that anymore.

---

Anyway, maybe it'd be better for at least a net 1 DT2 score with two inclusions.

How? They have blocked PM's from me since their tirade in 2017...


Given on the 2020-09-12, 22:09:04 you wrote this: (context - In relation to their two negative trust feedbacks)


You don't suppose they should have gone back over the dozen or so trust feedbacks on their alt accounts and made corrections?

No.  They consciously used two alts for negative trust feedback AND they have also in more recent times used the DT trust to engage in DT trust abuse.  Plain and simple.

---

---

Changing the number of DT1/DT2 required to enable the changes proposed by the OP will only mean someone with alts can and will use them to slam their distorted votes on others.

Oh, you are referring to the default, default settings. I made my argument in the post you cited that the status quo should remain.

The option to customize the number of net inclusions for someone to be included in your trust network is not currently available. Currently, if someone has a net 0 number of inclusions (with at least one inclusion), they will be in your trust network, and the only way around this is to manually exclude them. I was proposing to allow for you to require that at least x number of inclusions to be required for someone to be in your trust network if they are not explicitly in your trust list.

There is one thing you are leaving out -- the underlying reason why someone has this power is that many people have included the person in their trust list. Those who are on DT1 is (somewhat) random, and does rotate, however the people who are on DT1 are not arbitrary, they have effectively been chosen by the community to have this power.

I did at least two and a net of one.

although your net +2 is a tad stricter. I would think either one is better than current set of rules.

so far everyone voted for either of the two above.

I like the proposal too. At least it's better than having just one inclusion. But I will always like to have higher number of net inclusion. It reduce the chances of spam DT2 members. Right now it is too easy to be in DT2. No disrespect to anyone but many users are using it in their favour very easily.

Voted again but unfortunately I pressed the wrong option LOL
I wanted to go with Net of 3 or more DT1 Inclusions 😂

I really like this proposal, so I reset the poll and included this option.  If you've voted, please vote again.

Here is another nuance with the new proposals: DT1 lottery.

(double check the math below before making long-reaching conclusions, as it could be completely wrong because Saturday and... you know...)

If there are 125 candidates to select DT1 from (AFAIK typical these days), the chance of any candidate to be in DT1 in any given month is 80%. Under current rules, that's also the chance of a DT2 member to be in DT2 if that DT2 member has bare minimum of inclusions (one).

If we bump this up to two inclusions, then even if a DT2 member has the required minimum of inclusions (two), their chance to be in DT2 is ~64% since they need both of their "includers" to be in DT1 at the same time (0.8 * 0.8 = 0.64).

Of course it gets more convoluted once you add exclusions and/or more inclusions. For example someone with one inclusion and two exclusions (barely out of DT2 under current rules) still has a ~16% chance to "sneak" into DT2 if one of their "excluders" is not in DT1 (20% chance) and their "includer" is in. Adding two required "includers" drops that chance to ~13% (0.64 * 0.2 = 0.128).

If the DT1 candidate pool is 150, then the numbers above change from ~ 67% to 44%, and from ~ 33% to 15%, so a larger pool appears to make the proposed rules even tougher. But this doesn't consider the increased chances of someone being included or excluded by a potential DT1 member because there are more potential DT1 members.

Have you asked hilarious about this? Considering the 2 years between feedbacks, he might not even have noticed he left feedback from 2 different accounts.
Since hilariousetc has DT2 strength (8), DireWolfM14's proposal isn't going to change this.

I've voted "None of the above", because I agree with this.

I've said this before:

I used to browse the forum on occasion with the ;dt tag appended to every page, because sometimes it is useful to see what the vast majority of other users are seeing. I don't anymore, though, because DT is a complete mess. The number of users on DT2 with a single inclusion who have been included for all the wrong reasons is far too high, as Loyce's numbers above show.

I used to think like this too, but given what a mess DT is now, any inclusions/exclusions I make are based entirely on what I want to see when I browse the forum. It is my trust list, after all. Any effects this has on DT (short of blatantly inflating my own trust score) do not factor in to my decision at all. If I go way off base with my inclusions, then I'm certainly people will let me know and exclude me in turn. But I still agree that any one person should not have the power to add someone to DT2 without any input whatsoever from anybody else.

Here is an example of how a DT1 member is abusing the system:



by promoting their alt to DT2 status, their alt can then go on to give positive trust feedback to others - in this case unnoticed:

https://bitcointalk.org/index.php?action=trust;u=507936;page=trusted;offset=0;dt


Why does this matter?  The person who received these two trust feedbacks then has a higher overall score on BPIP and also in the forum:

https://loyce.club/trust/2022-06-25_Sat_05.09h/507936.html


on BPIP: https://bpip.org/Report?r=mosttrusted their score would be 614 not 624 - their overall recognition score would be 897.333 not 900.333 and put them into 17th place behind Lauda on a score of 898.000

(Let's not forget this user has in the past knowingly used the trust feedback by both their DT1 and DT2 to dispense *negative* trust feedback).

---

---

This is the slippery slope you go down - when you set the bench mark at one level, there will be those who are able to clear that hurdle and then are able to elevate those around them - sometime unnoticed, other times noticed after the fact.

I have no doubt DarkStar_ did not notice this - and I have no doubt DarkStar_ was NOT a participant in it occurring.

Counting zero as a positive score is an oddity I'll give you that.

I like it, this makes even more sense than requiring 2 inclusions.

Most of the members with 5 inclusions on DT2 have been on DT1 at some time. Asking for 5 DT1 inclusions makes DT2 harder to reach than DT1, so DT2 will be mostly limited to members who lost the monthly DT1-election.

I only added the negatives to show how many users are excluded. But it doesn't really mean anything for DT2. Most (2443) of the 2723 users with DT1 strength (-1) are excluded by a single DT1-member.

That's not right. It turns out the data I posted on "DT2" in this topic includes DT1-members. I overlooked that last night. The 614 DT2-members you mentioned from my other topic don't double count DT1-members.
I've added a warning, but won't do a recount.

But not exclusions that include a member in DT2.

So as far as I can see the list gives us information about the members in DT2 from 0 onwards but the negatives are only the exclusions, not being able to deduce if there is among these any DT2 member  (someone for example with one DT1 inclusion and excluded by another two DT1).

According to my count from 0 to 12 inclusions there would be 614 members, and according to the last Complete overview of users on DT1 and DT2 and their ratings there were 611 DT2 members.

It will probably be because there are at least 3 new DT2s since the last update.

In any case, it seems that there is no one of what DireWolfM14 said:


It is far from being a generalized problem.

In the end, the conclusion I draw from the discussion is this:


The 5 inclusions proposed by yahoo62278 are too radical a change in my opinion and I doubt that theymos would be willing to make such a change, besides it would mean moving to a somewhat aristocratic system.

I'd like to see 5 inclusions or more personally. It has already been shown multiple times that quite a few members have no idea how the trust system works, but if multiple users include the same person it makes it more likely that person is a little more trustworthy.

Of course we would have to monitor the recent inclusions at 1st and make sure lowlifes aren't just adding their scammer buddies just to try n scam the system.

I think it should be Net 1 but with 2 inclusions minimum (i.e. like the current formula but bumped up by 1).

One inclusion, no exclusions - not in DT2.
Two inclusions, no exclusions - in DT2.
Two inclusions, one exclusion - still in DT2.
Two inclusions, two exclusions - no longer in DT2.

And so on.

This would reduce the chances of (lone) rogue "includers" and rogue "excluders" from messing with the system (self-scratching etc).


I don't like this asymmetry (you need to gain two inclusions to get in, but one asshole can kick you out). The current system is also asymmetric in the other direction (you need just one inclusion to get in, but only a gang of two assholes can kick you out). I prefer a balanced amount of assholery.

Correct.

That's how exclusions work.

No, on second thought she shouldn't be in DT, the same way as if Claire was excluded by 33 DT1 members and included only by one.

I said earlier that I wasn't sure but I think it would be better to count net inclusions.


What does that mean?


As you say that I understand that there are 94 users with 2 net positive inclusions, 61 with 3 net inclusions and so on. But continuing the progression with the negative numbers is not clear to me, as it could appear that there are 2723 users with one negative net inclusion (or one net exclusion), 284 with two net exclusions, and so on.

And there are not 2723 users in DT2.

Yes, but you're going to be disappointed:
That doesn't include users with zero posts.

I'm not sure what you mean by this.  So, if Claire was excluded by 3 DT1 members, and included by only two she would remain on DT2?


I agree, there should be at least two DT1 inclusions at the very least.

---

It is customizable, as you very well know.


I'm not talking about my network, I'm talking about Default Trust.  You obviously know that the Trust System is customizable, or you wouldn't have customized your own.  This wouldn't change anything for those of us who've customized our trust lists, it would only have effect on those who have not customized there's, i.e. newbies and those who've yet to bother.

---

Here's an example of why I think this is worth discussing; I recently included a member into my trust list who's not very active but has been here for a significant amount of time, and seems to have a good head on his shoulders.  He came to my attention when he questioned me about a tag I left for a suspected scammer, suggesting that I may have jumped the gun.  His concern was admirable, and showed restraint.  I've had the same concerns in other situations.  After looking through the reviews he left for others, and reading some of his posts I decided I wanted to see his reviews in my custom trust system.

Since I'm on DT1, now that I've added him he's on DT2.  My actions have an affect of the trust lists of the majority of forum users.  I don't believe that I alone should have that power.  If some other DT1 member feels the same way about said user, then great, I'm not alone.

Also, if Bob believes he should have that power, that would make Bob specifically the type of person that I say shouldn't have it.

---

WTF?

---

@LoyceV, any way to tell how many members have a custom trust list?

I counted DT2-inclusions on Trust settings: there are 33 users with 0 (net) inclusions on DT2, and 308 with 1 inclusion.
Longer list (Update: this list includes DT1-members (when included by another DT1-member, they're on DT2 too)):
I'd say the minimum should be DT2 strength (2). That removes about half the users, and makes it less of a "burden" to include someone.

Perhaps this should be a customizable option for users to set.

I generally agree that in order for someone to be excluded from your trust network, there should be a strong consensus that the person should not be in your trust network, so a maximum of net 0 is appropriate for someone to be included (assuming one inclusion). I also don't think that having a limited number of "bad" people in your trust network is not the end of the world, especially if this person is not regularly sending trust ratings.

I realize that some may disagree with my above opinion, and they should be able to customize how their trust network propagates. I don't think this setting should propagate to others who include you on your trust list, so this particular setting would only apply to you.

I explained exactly the same. Right now only one DT1 member adds Claire to his list and it does not matter what equation it is, as long as the score is not negative, Claire is DT2 . It's x >=0.

No wonder we have so many DT2. I do not know what should be the standard score to be in the DT2 but the higher number is always reduce the number of spam in the system. I voted for 1 anyway.

It sounds to me that we were talking about this topic in some other meta thread about this. I'm sure LoyceV participated in the conversation, so I'm guessing he'll come around here to comment.

I'm not sure about subtracting exclusions. What I would be fine with is if there were two instead of one inclusion by DT1 to be DT2, I think it would reinforce confidence, and at the same time reduce the number of DT2 that I see it a bit long.

According to Bpip you have a net score of +4 inclusions.  Just to eliminate any confusion; for Claire to be on DT2 at least one DT1 member must have her included.  But, if one includes Claire and one excludes he, she will remain on DT2 even though her net inclusions is 0.

I am of the opinion that a net score of +2 should be required for inclusion to DT2.  So, for Claire to be included in DT2, at least two DT1 members would need to include her.  If one DT1 member excludes Claire, she would need a total of three inclusions to be on DT2.  I think this would reduce the Trust System spam that is so common today.  It would also reduce the potential for trust system cliques from developing.

What does DT2 mean?

We know how it works. If you list someone in your trust list, they are DT1 for you regardless of who have distrusted them. If they have someone on their trust list, they become DT2 by default (keep the default trust depth apart). So, your 1 inclusion is enough to make someone DT2 for you unless they are on your distrust list.

The same also applies here. But with net score as zero it doesn’t make sense but you need something in between +1 (DT2), -1(off DT2). I don't think this makes a lot of sense. You can't make someone DT status as neutral, something like neither shown nor hidden by default. I don’t know if I was able to make sense but that's the case here in my opinion.

I think having a positive score all it requires to gain DT2. Mathematically 0 is neither positive nor negative but I think in computer language it considers as positive. So when the score after at least an inclusion and exclusion from two DT1 executes, the algorithm considers it positive. The user gains DT2. Hopefully I get your explanation correct. But I will consider x >=1 not x>=0. Right now it must be x>=0.

Wait them to see it, you will get the data shortly 😉

I'm just curious about what you folks think.

Currently, only one DT1 inclusion is required for a member to be included in DT2 status.  In fact, a net score of 0 inclusions is all that is required, for example:  If Alice and Bob are both on DT1, and Alice includes Claire while Bob excludes Claire, Claire will be left with a net score of 0 DT1 inclusions, but that's all that is required for Claire to gain DT2 status.  In this example Claire's reviews will be displayed as "Trusted Feedback" if the user viewing the forum has the default Trust System settings.

Of course anyone can change their own Trust System settings, including the trusted depth of your included members' inclusions.  This is a good way to customize how you see "Trusted Feedback."  However, this is not something most newbies do when they first join the forum.  I don't know what percentage of members use custom trust lists, but I suspect it's on the low side of total active members.

I have mentioned my opinion about this in the past, but I'd rather not sway the voting by expressing it in this post.  I do encourage anyone with an opinion to share it here.