Decentralised merit sources

[Zedpastin]

I have an idea which will operate in similar terms to the current trust system possibly implementing flags for making things easier. This being a Bitcoin forum I think we should encouraging decentralized approaches as much as we possibly can. The current trust system has its flaws but I think at heart it has a good foundation and isn't bias. How would we feel about introducing a similar system to determining who is merit sources? At the moment we are relying on theymos who has a million other responsibilities to review and appoint merit sources. The primary concern here is members who have showed that they are a valuable asset to the forum and are both good posters and good at rewarding others for good posts are not being appointed merit sources or are being appointed long after they have made an application.

Just a few examples:
https://bitcointalk.org/index.php?topic=5158075.0
https://bitcointalk.org/index.php?topic=5156799.0

Both of these members I would say fit the criteria for a merit source but at the moment we are waiting for the approval of theymos which could take a long time. To counter this I would suggest implementing a system similar to the trust system. Basically members can put people who they deem good quality posters into their list. The more people who have you on their list the more weighting you have. Anyone in the top 5%/10% or whatever we deem suitable for the size of the forum would then become a merit source. We could have several tiers like depth 1/2/3 which members in the higher tiers give their custom lists more weighting and thus anyone on their list more weighting to becoming a merit source. To avoid abuse we could implement this right away by allowing current merit sources to have more weighting and prevent account farmers from having an impact by reducing their weighting to almost nothing. I know that this is a semi decentralized way of implementing the system but I think this would be the most effective. This way we can have a continuous flow of merit sources and remove inactive merit sources quite easily without waiting for review from theymos and a more community driven effort. I understand with the current system there isn't much room for abuse when appointing merit sources and this new solution might introduce some sort of abuse but by allowing merit sources to have more weighting hopefully this abuse would be minimal because after all the first generation of merit sources would have already been appointed by theymos.

[Quickseller]

The amount of merit a person has determines some limitations, such as posting limits and restrictions a person has, so I don't think having people setup a network of merit sources would really be feasible.

theymos has used a fairly diverse group of people acting as merit sources, with the goal of having merit sources give out their sMerit supply to a wide range of people who are deserving merit.

[Jet Cash]

The proposed system sounds a bit too complex to me. Besides, I'm already a decentralised merit source, nobody tells me what to do, even when I'm behaving like a loose cannon.

[LoyceV]

To counter this I would suggest implementing a system similar to the trust system. Basically members can put people who they deem good quality posters into their list. The more people who have you on their list the more weighting you have.

This reminded me of this:

This thought occurred to me recently:

If you wanted to implement Merit in a decentralized forum (ie. one in the vein of Freenet's Frost or FMS), you could do it in this way:
 - Everyone can, from their own perspective, give unlimited merit to posts, and these merit transactions are put into files which each user publishes via the decentralized system. (Like a merit.txt.xz which every user publishes.) Unlike on bitcointalk.org, you can also give people merit without an associated post.
 - For everyone who has merit, you download their merit-transactions-list, but scale down/up all of the numbers so that the total merit that they send is equal to the actual sMerit that they own. It might or might not be useful to do this via some sliding time frame scheme so that merit transaction amounts aren't just continually diminished over time as they increase in quantity.
 - Apply the above step recursively, creating a web-of-trust-style merit network

Then every user has a subjective merit score for each post (sort of like the bitcointalk.org trust system, which was inspired by FMS). And if you wish, you can assign people to be merit sources from your perspective by sending them large amounts of merit directly; these might or might not appear in the merit-transactions-list which you publish.

[DdmrDdmr]

<…>

I doubt there is much of an issue in terms of the amount of time it takes to decide to appoint a Merit Source on @theymos’ behalf. 36 were added over a weekend back in September 2018 (see Enhanced newbie restrictions & requirements), and recently a handful were added when one was let of the hook.

Merits Sources are added or reallocated when there seems to be a specific need to pump-up the overall merits in circulation, and monitoring that number (perhaps in relation to the number of active profiles) is likely the driver to make adjustments. Specific board that bring up an issue may drive for a specific course of action to take place if the case is well defended. Selecting merit sources on a quasi-DT based scheme looks like an overly complicated thing to do, prone to wars, vetting, abuse and so on.

Note: I get your intent though, as within my first 50 posts I gave it a go at proposing an idea to democratize merit assignment to all levels (complementary).

[Heisenberg_Hunter]

This seems like a good but complex idea to put in practice. The allocation of merit sources should more or less be based on theymos himself as he is the one who allocates the limit to everyone. For instance, o_e_l_e_o applied for being a source but he was given less smerit allocation per month and he uses them fairly. Similarly certain local boards are far more inactive when it comes to really good quality posting. Let's consider the Indian board as an example where there would be only 10-15 posts per day made by various users and among them only 2-5 of them would be merit worthy posts. So in such cases, smerit allocation to the Indian board could be limited if I am made a source based on my application. If at the end of the recycling period I am having more of them, I would be spending them in the technical board and Beginners & Help Board (where I am a common visitor).

But if there is a dedicated merit source, other people in the board would be much more interested in posting good quality posts and contributing to the board subsequently. On the contrary, Merits and Trust are inversely related to each other and can never be compared. Trust is based on other person's perspective whereas Merit is based on the sender's perspective, some may deem a certain post to be of merit-worthy while others may not.

[tranthidung]

theymos has used a fairly diverse group of people acting as merit sources, with the goal of having merit sources give out their sMerit supply to a wide range of people who are deserving merit.

By now, theymos shows that he has not had plans to expand merit sources too much, but total merit sources will be increased over time, when the forum will become much bigger than now. With low inflation of merit sources, maybe when new merit sources added, distributional merits on old sources might be adjusted (of course, weeks after operations of new merit sources, when they demonstrate that they actually do their works as sources well).
The latest expansion of merit sources increased total merit sources from 123 to 130 (after the withdrawal of @Vod weeks ago).
One year anniversary of merit system
On 24th Jan. 2019, total merit sources are:

Basic statistics of merit system
How many merit sources in total by now?

Code:

There are 123 merit sources with a total merit generation of up to 20735 sMerit per 30 days

Merit stats page

After checking (today), I saw that after the biggest latest change, there are 130 merit sources, but today I checked and found that we already have 131 merit sources in total (one new merit source added, but don't know when it was added)

There are 131 merit sources with a total merit generation of up to 21045 sMerit per 30 days

[Welsh]

My personal opinion is It's a good suggestion, but when put into practice it just doesn't seem feasible this is just to complicated. We already have issues with users understanding the trust system, and that is often misused. At the moment theymos chooses who he believes is right to distribute merit fairly, and I think this system is already working. There aren't too many issues with abuse, and the system seems to be somewhat stable in that there aren't a massive amount of users ranking up, but good quality ones are.

The forum currently isn't that active that merit isn't being distributed enough. The current user base, and the amount of merit sources are somewhat balanced at the moment. The only things we could likely improve on his merit distribution in local boards. This could be looked into in the future if we ever hit the point where the forum activity justified a system which "takes care of itself", but at the moment I don't think we need to be implementing complicated systems when the current system is good enough.

[o_e_l_e_o]

The primary concern here is members who have showed that they are a valuable asset to the forum and are both good posters and good at rewarding others for good posts are not being appointed merit sources or are being appointed long after they have made an application.

I don't think you can chalk this up simply to theymos being too busy. Some sources were appointed within a day of application. I think he is simply happy with the amount of sMerit currently being generated.

My worry with your system is that you would end appointing people who didn't want to be sources. Many users would simply nominate their friends, some would nominate trading partners, and many more wouldn't understand the difference between customized merit lists and customized trust lists, and just put the same people on both. You would end up giving sMerits to people who have little interest in spending them, and total merit circulation would drop.

[CryptopreneurBrainboss]

The only reason those two users and many more qualified members haven't been appointed is because theymos (forum in general) haven't seen any need for more sources at the moment. The suggestion isn't that bad but like others have said, it's too complex and considering the trust system haven't had much impact in regards to forum users customizing their trust list don't think the merit system will be any different.

Have you notice, it's easier to be a DT member if you're actively contributing positively to your local baord same fate might befall the merit system if your suggestion is implemented which doesn't seem right as more sources will be centralized to the local boards will might result to abuse. Not saying the local boards don't deserve source but giving them too much might be too much responsible.

[tranthidung]

The forum currently isn't that active that merit isn't being distributed enough. The current user base, and the amount of merit sources are somewhat balanced at the moment. The only things we could likely improve on his merit distribution in local boards.

If we don't need too much bitcoin in total supply, low inflation with scheduled halving each four years, I believe that we don't need to see more merits distributed each month to merit sources. I agree that local boards should have more merit sources, but it raises another difficulty issue on how to prevent abusements in local boards, that usually more chaotic and hardly to control, in my opinion. There are suggestions on increases of moderators on local boards, and I think such suggestions make sense. I don't think we should increase merit sources in local boards if moderators in local boards won't be risen concurrently.
Or new moderators for local boards simultaneously play as merit sources.

[Welsh]

If we don't need too much bitcoin in total supply, low inflation with scheduled halving each four years, I believe that we don't need to see more merits distributed each month to merit sources. I agree that local boards should have more merit sources, but it raises another difficulty issue on how to prevent abusements in local boards, that usually more chaotic and hardly to control, in my opinion. There are suggestions on increases of moderators on local boards, and I think such suggestions make sense. I don't think we should increase merit sources in local boards if moderators in local boards won't be risen concurrently.
Or new moderators for local boards simultaneously play as merit sources.

I don't think merit abuse, and the amount of moderators are directly related. Sure, if a merit source is abusing their merit in the local boards, then this may be a admin issue, but not a moderator. Anyone, who thinks merit abuse is going on can flag that up via reputation, and in cases involving merit sources to theymos. I also have a different viewpoint on Bitcoins limited total supply, and merit. Bitcoin is scalable, and merit requirements seem to be staying as they're to prevent injustice to those that have already earned it. Therefore, as activity increases, and hopefully the amount of good quality posts too we will need more merit to distribute if we are to reward good quality posters, unless your goal is to have an elite group of users at the higher ranks. Although, I have a different perspective to that, and believe good posters should be rewarded, and as the demand rises so should the supply. I'm not implying to drastic measures either. Having it challenging enough to combat account farmers, and spammers from reaching the higher ranks without hindering genuine quality members is the way to go about it in my opinion.   At the moment I think the forum is in a decent place when it comes to distribution of merit like I said previously. Generally, as activity rises in the forum more content is being buried, and in different time zones. Therefore, to cover all of this we would need to carefully select merit sources in sections, and timezones which merit aren't being distributed, and determine whether its because there's no good quality posts being made or there's not enough merit source eyes seeing the content.

[Upgrade00]

. To avoid abuse we could implement this right away by allowing current merit sources to have more weighting and prevent account farmers from having an impact by reducing their weighting to almost nothing.

I am curious to know how you'd determine who an account farmer is. Earned merits, rank, trust?
It think it would be susceptible to abuse on a small scale.
Overall, I don't think k you have to fix something that isn't broken. Delays in merit source isn't so much an issue that warrants changing the whole system.

The more people who have you on their list the more weighting you have. Anyone in the top 5%/10% or whatever we deem suitable for the size of the forum would then become a merit source.

If you make it an automated process, then members don't need to apply, or show interest. Take into account that not everyone would be interested in being charged with smerits to distribute within a certain time frame.

[The-One-Above-All]

Merit is poorly named. Should just be renamed "likes"

It seems that merit was initially brought in to prevent bots and account farmers powering up their accounts. This it was reasonably well equipped for.

Sadly people seemed to have placed some other FAKE value to these scores and tied them to TRUST or VALUE or MERIT.

Merit is currently quite a meaningless metric. It seems pointless to worry about who gives it out before first setting up some strict enforceable criteria for giving merit. Until all posts are reviewed and all run through the same criteria then the score is meaningless. It is actually the most misleading and damaging aspect of the entire forum.

1. decouple it from rank after full member
2. come down a lot harder on people giving merit to garbage that gets debunked or has been debunked already on that thread. Remove them from merit source.
3. Set up some strict criteria that sets a threshold that must be met to give merit.
4.Max merit per post range from 1-3, good, Vgood, excellent...giving merit of 50 per post - sometimes for complete garbage destroys the entire system.

The "sources" don't make that much difference so long as each post gets the merit is deserves as matched against a strict criteria that ensures it is a valuable post in terms of reaching the optimal solution/outcome.

Best thing would be some AI that could be run over posts that are submitted to a "merit worthy" pool. Where anyone including the poster can submit the posts but with each post deemed "non merit worthy" then their ability to submit to that pool is reduced. That point is not likely anytime soon, so for now merit sources could review that pool. Also even though you apply merit to the post it should not be visible to other members for a delayed period, I see a lot of lemmings merit behavior as the thread develops.

To be fair a real meritocracy is only possible if the VERY smartest people are merit sources. Else how they will decide which posts of a thread are the most valuable in terms of reaching the optimal solution?  We have merit sources like Tman? issuing merit to posts that are observably the lowest value posts in a thread so you may as well just rename "merit" to "garbage" for now. Who cares who doles out the garbage? it is still meaningless garbage for now.

[bitart]

Merit is poorly named. Should just be renamed "likes"
...

Sometimes people use merit like "likes" but the original idea is different.
You can "like" a post if it's funny, or if you agree with the content of the post.
But merit is for a post that contains something excellent stuff, it can be technical stuff, explanation about a difficult thing, etc...
However, like is used for facebook and the forum is really different from facebook (and I don't know if like is copyrighted by facebook or not, but it's possible, I just don't have the time to google for it)

[mikeywith]

I am not a big fan of this idea, due to the reasons that other members have mentioned , decentralization is not always the best option, many things have to be a one man's decision for them to work smoothly.

but if this idea comes to existence then a a "demerit" system which functions the same way (election based) must come with it, just like we have negative/positive , support/oppose in the trust system.

given the fact that some untrustworthy members made it to DT , it's guaranteed that some merit abusers will be merit sources , we don't want Meta to be full of "Merit source abuses" , the community should handle these issues , if not then Theymos should keep his hands dirty and the current system remains the same. 

[nakamura12]

What theymos did is already working for theymos by making a certain forum user as a merit source as a substitution to theymos when handing out merits to deserving forum users. What if theymos didn't made someone as a merit sources then theymos would not be able to give rewards to those who are deserving that is why theymos assigned a group of people to do the job without someone ordering a merit source to send that person a merit so it's also decentralized.

[tranthidung]

but if this idea comes to existence then a a "demerit" system which functions the same way (election based) must come with it, just like we have negative/positive , support/oppose in the trust system.

Decentralization and / or demerit system might make the forum more crazy. By writing crazy, I imply more drama might pop up, especially with demerit system. If there are things should be implemented, I think demotion is one of them.
It is not a essential thing to do, because basically over time, more and more self-made users will appear, and within next few years, we will have more self-made Hero and Legendary members (from current self-made Full Members and Senior Members). They will certainly be more competitive than grandfather Hero & Legendary members (whom can not earn a single merit). Moreover, for such low quality higher ranked users, they might sooner or later will violate something like plagiarism and others thing. Consequently, such low quality generation will be wipen out gradually.

[Heisenberg_Hunter]

Merit is poorly named. Should just be renamed "likes"

That's very much true. I don't think merit sources are going by the way like this but other's who have been airdropped with merits and who aren't really aware of forum policies and workings are abusing the system for sure. The 50 merit posts should be of high quality standards but at times I do see they are being rewarded for a supportive post/ for a high bid in the collectibles section too. Recently I was browsing the collectibles section, where one of the guy gave 50 merits for just a high winning bid of another member. Doesn't this seem to be stupid enough? Merits are introduced for identifying high quality posts and not for just throwing them away as a reddit upvote or a facebook like.

but if this idea comes to existence then a a "demerit" system which functions the same way (election based) must come with it, just like we have negative/positive , support/oppose in the trust system.

Just like the negative feedback, a demerit button will create too much drama and widespread havoc all over the forum. Meanwhile the trust actors play a very much different role in terms of merits and not just me most of them would be against the usage of a demerit button.

Let us consider an instance where a spam one-liner post was merited with 50 merits which doesn't even deserve a single merit and a demerit source as elected through a more decentralized way be demeriting them. If this continues on a larger extent, bullying and targeting of users will certainly increase and finally we would see the Meta and Reputation boards being filled with utter spam topics on demerit system. Not just this, it might severely deter the quality post being made by individuals if they are being hated by a demerit source.

For instance if a good poster who is contributing positively is on a clash with some of the demerit sources, will they be constantly demerited until they reach the negative merits? Demerits or negative merits are difficult to put in practice and wouldn't be good for the forum in longer run.

[Macadonian]

but if this idea comes to existence then a a "demerit" system which functions the same way (election based) must come with it, just like we have negative/positive , support/oppose in the trust system.

Demerit cannot and should not become a feature. If we think trust and merit gets abused imagine the after effects of de merit at the moment we definitely have bias merit sources and that's something we can't avoid and isn't necessarily a bad thing in the grand scope of things. Just think of the hate fueled arguments that happen within the Defaulttrust and with particular well known members who have a bad reputation for whatever the case. I would predict that these members would be nuked down to oblivion yet these members have and still do in some cases contribute meaningful to the forum. At the moment merit sources can only have positive effects on members by meriting them which means merit abuse is easy to spot on a mass scale and bias isn't much of a problem. Where as if you had the power to derank someone and having a negative impact on people you dislike on this forum then you are going to have issues.

The merit system is already used as a liking system by many even some merit sources are meriting the post because they agree with it and not because its a quality post. This is fine like mentioned above because it has no negative effects but if we had the ability to de merit even if this was restricted to merit sources it would unleash the ability to have a negative effect on people you dislike, people who have scammed and people who disagree with your opinions. Bias will always be present no matter what and believe it or not the moderators and admins will have bias. I bet higher ranked members and members they like get a little bit more leeway than "normal" users like I have said though this isn't necessarily a bad thing and is just part of human nature. Even in the some of the highest court cases in the world judges will have bias. The de merit system sounds like a horrible idea to implement there are enough personal attacks and grudges on the forum as is.

A decentralized version of merit does not sound like it would work although even if someone like this was implemented demerit should still not be implemented.

[Nnuego]

How many decentralized merits source do we have in a forum that we have millions of account. Remember not all of them will be able to read all the post by day. I think the system its very difficult to manage cos it will somehow confusing for them to merit good post per minute