The new DT system, updated 04 MAY 2019

[o_e_l_e_o]

@LoyceV, how difficult would it be to give us some numbers for different scenarios? For example, how many DT1 members would there currently be if the requirements were raised to the following (with the "upper requirement" of 2x 250 merit votes staying the same in all cases):

A) 10x 25 merit votes
B) 10x 50 merit votes
C) 20x 10 merit votes
D) 20x 25 merit votes

The problem with the DT list growing too large is only going to get worse over time as more users set default trust lists and more users earn 10+ merit.

[1714028727]

1714028727

[1714028727]

1714028727

[1714028727]

1714028727

[1714028727]

1714028727

[Blazed]

Wow... just noticed how much DT1 has changed. I think having the higher merit requirement is a good idea as 10 merits are not very many. Seems I am now just a regular user without that massive DT1 abuse power  

[LoyceV]

@LoyceV, how difficult would it be to give us some numbers for different scenarios? For example, how many DT1 members would there currently be if the requirements were raised to the following (with the "upper requirement" of 2x 250 merit votes staying the same in all cases):

I've done that before. It's probably faster to play around with http://loyce.club/trust/ranking/ and manually see who still qualifies. You for instance would be DT1 if it requires 15 votes from users with 500 Merit, but you don't make it to 30 votes with 10 Merit.

Seems I am now just a regular user without that massive DT1 abuse power  

You're still DT2, but you need to be more active to be DT1. Say make a post on the 3rd of each month.

[TMAN]

Wow... just noticed how much DT1 has changed. I think having the higher merit requirement is a good idea as 10 merits are not very many. Seems I am now just a regular user without that massive DT1 abuse power  

Sup blazed. Get on slack

[Blazed]

@LoyceV, how difficult would it be to give us some numbers for different scenarios? For example, how many DT1 members would there currently be if the requirements were raised to the following (with the "upper requirement" of 2x 250 merit votes staying the same in all cases):

I've done that before. It's probably faster to play around with http://loyce.club/trust/ranking/ and manually see who still qualifies. You for instance would be DT1 if it requires 15 votes from users with 500 Merit, but you don't make it to 30 votes with 10 Merit.

Seems I am now just a regular user without that massive DT1 abuse power  

You're still DT2, but you need to be more active to be DT1. Say make a post on the 3rd of each month.

Not too worried about being DT1/DT2 with there being so many people already added. Do you feel the system works better now rather than the old way?

[nutildah]

Pretty nuts. Honestly I think the number should be capped at 100 total. I don't care if I'm on it or not, but we cant have everybody on the forum being DT.

Agreed. It is becoming unwieldy, and with every update and bunch of brand new inappropriate accounts make their way on to DT2 as well. I'm sure theymos previously said that if DT1 reached >100 users, then each month a random selection of 100 users would be picked from the pool. I'm not sure that's the best option, as you run the risk of scammers losing their warning if all their tags suddenly moved to "untrusted". As I said at the initiation of this new system, I think the requirements are too lax - 10 votes from users with 10+ merit is too easy to achieve. Either the number of votes or the required merit (preferably both) should be increased. And as I mentioned in another thread, there should be a merit and not-inactive requirement for DT2, same as there is for DT1.

I agree with all of this, but instead of having the 100 "randomly" selected they should be selected based on the criteria used in Loyce's DT "Rank Up Pipeline." Maybe have the selection based on "who fits into DT1 the most" rather than relying on random elements.

[suchmoon]

I agree with all of this, but instead of having the 100 "randomly" selected they should be selected based on the criteria used in Loyce's DT "Rank Up Pipeline." Maybe have the selection based on "who fits into DT1 the most" rather than relying on random elements.

I think it's time to set the criteria to include only true believers, i.e. ones who have "moon" in their names.

Seriously though, I think default depth should be set to 1 and then DT1 should be capped at ~200 or so based on most votes. Or leave it at depth 2 but cap DT1 at 50-ish. This DT2 business isn't working out, it looks like we can't find 100 users who would be 100% attentive to their trust lists, let alone feedback ratings of their "subordinates".

[TMAN]

I agree with all of this, but instead of having the 100 "randomly" selected they should be selected based on the criteria used in Loyce's DT "Rank Up Pipeline." Maybe have the selection based on "who fits into DT1 the most" rather than relying on random elements.

I think it's time to set the criteria to include only true believers, i.e. ones who have "moon" in their names.

Seriously though, I think default depth should be set to 1 and then DT1 should be capped at ~200 or so based on most votes. Or leave it at depth 2 but cap DT1 at 50-ish. This DT2 business isn't working out, it looks like we can't find 100 users who would be 100% attentive to their trust lists, let alone feedback ratings of their "subordinates".

So just lose DT2? Not a bad call although the work for DT would increase

[o_e_l_e_o]

Seriously though, I think default depth should be set to 1 and then DT1 should be capped at ~200 or so based on most votes. Or leave it at depth 2 but cap DT1 at 50-ish. This DT2 business isn't working out, it looks like we can't find 100 users who would be 100% attentive to their trust lists, let alone feedback ratings of their "subordinates".

DT2 is becoming less and less useful as time goes on. Every time theymos recalculates DT1, then a bunch of inappropriate accounts are showing up on DT2; proven scammers, long dead accounts, accounts whose only left feedback is a positive to the DT1 member who is now including them, that kind of thing. I still think stricter voting criteria for DT1 (higher merit required and more votes required) and some new criteria for DT2 (not inactive and minimum of 10 earned merit) would be a good first step.

There are a bunch of good DT2 users who I don't think we would want to lose by setting default depth to 1, but I would agree to setting a much lower cap on DT1 users coupled with stricter requirements as above.

[LoyceV]

Do you feel the system works better now rather than the old way?

I feel like it caused a lot more drama with the power-fights. It also "inflated" trust: users got either a lot more green, or a lot more red.

I agree with all of this, but instead of having the 100 "randomly" selected they should be selected based on the criteria used in Loyce's DT "Rank Up Pipeline." Maybe have the selection based on "who fits into DT1 the most" rather than relying on random elements.

I expect theymos to use a combination of randomness and votes. See https://bitcointalk.org/index.php?action=trust;suggest for example: my name shows up most of the time (refresh the page a few times), but not always.

There are a bunch of good DT2 users who I don't think we would want to lose by setting default depth to 1, but I would agree to setting a much lower cap on DT1 users coupled with stricter requirements as above.

An easy and effective restriction would be to require 2 (my suggestion) or 3 (LFC_Bitcoin's suggestion) inclusions from DT1 to be on DT2. That takes away the responsibility from just one person, and prevent users from directly influencing their own trust score.

[o_e_l_e_o]

An easy and effective restriction would be to require 2 (my suggestion) or 3 (LFC_Bitcoin's suggestion) inclusions from DT1 to be on DT2. That takes away the responsibility from just one person, and prevent users from directly influencing their own trust score.

Another good suggestion. As per your numbers, requiring 2 inclusions would cut the current DT2 list down to 128 from 378 - a much more reasonable number. 3 inclusions would cut it down to 65 which I think is probably on the small side. I'm not sure whether requiring net inclusions of 2 or more is better, or requiring a minimum of 2 inclusions but still just a net above 0.

Having said that, I still think DT1 is becoming too large.

[tranthidung]

Another good suggestion. As per your numbers, requiring 2 inclusions would cut the current DT2 list down to 128 from 378 - a much more reasonable number. 3 inclusions would cut it down to 65 which I think is probably on the small side.

I think that 3 inclusions from DT1 members  for DT2 members is a better one. I don't think more DT2 members will help to solve things better. There are abusements on DT2 already, so if there will be a implementation to cut it down, it should be 3. It is nearly the same as when merit system started in 2018, people complained and proposed to require 5 merits for Member rank, 50 merits for Full member rank and so on.
More DT2 members will result in more fake DT2 members, and annoying wars from users and DT2 members if they use their role wrongly.
Maybe they are good users, but who knows when they have power in hands, they might change somehow, more arrogantly, and arbitrarily use their power, then wars will occur.
I think we should have a vote on which cutoff should be chosen, 2 or 3 inclusions and consequent estimated DT2 members?
Quality should be prioritized than quantity.

[xtraelv]

Another good suggestion. As per your numbers, requiring 2 inclusions would cut the current DT2 list down to 128 from 378 - a much more reasonable number. 3 inclusions would cut it down to 65 which I think is probably on the small side.

I think that 3 inclusions from DT1 members  for DT2 members is a better one. I don't think more DT2 members will help to solve things better. There are abusements on DT2 already, so if there will be a implementation to cut it down, it should be 3. It is nearly the same as when merit system started in 2018, people complained and proposed to require 5 merits for Member rank, 50 merits for Full member rank and so on.
More DT2 members will result in more fake DT2 members, and annoying wars from users and DT2 members if they use their role wrongly.
Maybe they are good users, but who knows when they have power in hands, they might change somehow, more arrogantly, and arbitrarily use their power, then wars will occur.
I think we should have a vote on which cutoff should be chosen, 2 or 3 inclusions and consequent estimated DT2 members?
Quality should be prioritized than quantity.

I think part of the Theymos experiment is to decentralize the power. By having DT1 effectively vote on all the DT2 there is a lot of power by the existing  DT1. (I have a lot of respect for some of the DT1's)

In my opinion this is the Theymos gold dredge experiment. By putting all the sluice into the dredge (including the nuggets and grains that were already found) and tossing it around the theory is that all the gold gets stuck behind the riffles. The risk is that the gold that was already found gets washed away in the process.


But in terms of trust lists and trust networks being build the process appears to be working.

[OgNasty]

The problem with the DT list growing too large is only going to get worse over time as more users set default trust lists and more users earn 10+ merit.

That isn’t the problem, that’s the point. Eventually there will be so many eligible users that it can’t be manipulated with sock puppets as easily and rampantly as it currently is.

[o_e_l_e_o]

Sure, I take your point.

My concern would be the requirements are currently low enough that it would be fairly easy for one high ranking user, particularly if that user is a merit source, to earn enough merits on a few alts to vote on a couple of sockpuppets. With thousands more merit being distributed every month, it will only get easier as time goes on. I would rather be in a situation where the majority of sockpuppets weren't eligible at all, rather than relying on random chance to prevent them from being included.

[LoyceV]

I would rather be in a situation where the majority of sockpuppets weren't eligible at all, rather than relying on random chance to prevent them from being included.

That could happen indeed, but when I posted evidence of sockpuppets voting, it took theymos less than 2 hours to blacklist them.
It's likely we can't catch all of them, but I'm not really worried about large numbers of sockpuppets on DT1.

[OgNasty]

My concern would be the requirements are currently low enough that it would be fairly easy for one high ranking user, particularly if that user is a merit source, to earn enough merits on a few alts to vote on a couple of sockpuppets.

Sure, that's a valid concern.  Take a look at how much merit Foxpup has sent The Pharmacist.  I think this is why the random feature was introduced.  There would be too much crying and backlash if theymos took a hardhanded approach to stopping this sort of abuse, especially among the higher ranking members, which is why I assume he prefers a community based approach.  Unfortunately, theymos used some really poor judgement when selecting Merit Sources, which will have a negative effect on this forum for quite some time.

[nutildah]

My concern would be the requirements are currently low enough that it would be fairly easy for one high ranking user, particularly if that user is a merit source, to earn enough merits on a few alts to vote on a couple of sockpuppets.

Sure, that's a valid concern.  Take a look at how much merit Foxpup has sent The Pharmacist.  I think this is why the random feature was introduced.  There would be too much crying and backlash if theymos took a hardhanded approach to stopping this sort of abuse, especially among the higher ranking members, which is why I assume he prefers a community based approach.  

How does sending a bunch of merits to an already Legendary member aid sockpuppeting? Or are you suggesting The Pharmacist is funneling merits to sockpuppets?

Unfortunately, theymos used some really poor judgement when selecting Merit Sources, which will have a negative effect on this forum for quite some time.

I agree with you here. I can't see any rational reason why Quickseller should be a merit source.

[suchmoon]

Or are you suggesting The Pharmacist is funneling merits to sockpuppets?

Perhaps The Pharmacist is a foxpuppet.

[marlboroza]

Or are you suggesting The Pharmacist is funneling merits to sockpuppets?

Perhaps The Pharmacist is a foxpuppet.

I think he is trying to say that randomness was introduced because The Pharmacist has received too much merits from Foxpup:

Take a look at how much merit Foxpup has sent The Pharmacist.  I think this is why the random feature was introduced.