Topic: Newbies restrictions, pattern from the aftermath - data and few cosiderations (Read 401 times)

Just a quick update on the graph on the main post, i just got the merit data coming from today and updated it, it seems to be converging back to the values pre-restrictions indeed:

I have suggested this before but I think that those memebers who have recived red trust for merit abuse should be stripped of merit because they are likely to use their merit to farm accounts.We already have seen this tendency , users with red trust upranked 10s of Newbies to Jr.

It is much easier to do this than chase after them after they have spilled their merit on their farm.

I have some interesting data in works regarding activty of members ( posting activty and last activity).

I am sure that there are high ranking alts out there.

it may be a bit drastic 50 merits, however nothing it's preventing to raise the requirement of a small amount of merits, followed by a checking period to verify what are the effects. Then rinse and repeat till you reach the "sweet spot".

personally I think it should be 20, Members get a warning now and have 3 months to gain the needed 50 merits, if they cant then they lose sig rights - even existing members+

that should make this an interesting place

It's just an idea, but perhaps would be useful to have some projections about how many Hero & Legendary are out there, how many of these account are active/not locked, how many are still actively posting, then break everything down by their merit activity and potential merit.

I figure in the pre-merit time it was relatively easy to farm several account and somebody could have easily done it for years. I would not even be surprised at this point if we happen to find users which controls several legendaries, since for certain people this became a way to make quite a lot of money.

I would support that as well, and other members such as hilariousandco do as well.  I'm actually very glad he's still fighting for more restrictions on signatures, since Theymos obviously took hilariousandco's proposal of needing 1 merit to move from Newbie to Jr. Member seriously.

The best thing would be to just disable signatures up to a certain rank, like Member, which hilariousandco has also suggested.  That would instantly cause an uproar in Meta but likely a concomitant reduction in shitposts.  Theymos probably thinks this is a radical solution, but I think it's the only one that really has a chance of working.

The only part of that I would worry about is older, high-ranked accounts with airdropped sMerits being sold or hacked and used to rank up shitposters, which has apparently already happened.  I'm not too worried about it, since there are only an obviously finite number of those accounts.  I'm not sure if anything should be done at all, but I'm more sure that nothing actually will be done.

Another matter this may rise up is what to do with the unspent airdropped merit. Erase them could avoid this situations as well, but is not an easy decision to make.
Given the total unspent merits still kicking around, somebody may have enough to handle higher limit as well.

I believe this will only reduce spam a little. We could put 10 merits restriction and they would still able to cheat the system.
I think bounty rule should be updated from bounty managers side. For example, as we know that a full member requires 100 merits then its stake should be 10x more than a member.
This case, honest full members won't be much affected by spammy members, let alone spammy jr. members (100x). The reduction of alts incentives for bounty is a much better solution.

alright i was using this to do the counting:

Select count(*) FROM
(
SELECT  Count(UserData.UserName)
FROM meritdata
INNER JOIN UserData ON UserData.UserId = meritdata.fromid
WHERE date between  '2018-01-24' AND '2018-01-26'
Group by UserData.UserName
);

Well research, I strongly be believe required 10 merit to ware signature will reduce lot of spam. On the other hand people's will not able to abuse merit system easily. Because it's not easy to get 10 merits. Only require 1 merit for Jr. Members reduced lots of spam. So if require 10 merit than more spam will reduce.

In my opinion let them(Jr. Members) spam on bounty section only. If they are not able to ware signature I think others board will spam free almost. We can realized current situation, so many people's are motivating for make constructive post in order to gain single merit. So no doubts people will more constructive if need 10 merit to join signature campaign.

I match this data exactly, so the userbase we use is the same. The issue must be on the merit TXs and/or the dates perhaps as @LoyceV suggests. I didn’t get the same results on the change of rules week, so that is why I contrasted the first date period in your chart.

Without linking to the Senders (just so we can contrast the DBs), the number of TXs and sum of merit for the three day period I get is as follows:
nTx   nMerit   nReceivers   nSenders
6972   20053   2888           2988

I used in this case:
select count(*) nTx,sum(amount) nMerit, count(distinct m.user_to) nReceivers, count(distinct m.user_from) nSenders
from merit m
where m.converted_date >= convert(datetime,'24/01/2018 00:00:00',103)
and m.converted_date <= convert(datetime,'26/01/2018 23:59:59',103)

The Senders (2.988) is basically everyone that would be counted during the first three days (since all accounts that merited another (but one) were created logically before the 24th of January at the time per-se).

Edit: Found it!: The end-date on the chart is an open interval, not closed, so I should exclude it and therefore really use as an interval: 24/01/2018 00:00:00 .. 25/01/2018 23:59:59 That gives me the same results as the chart.

As far as I know none of the profiles from Merited users has been deleted. The post history on the other hand disappears when Newbies get nuked.
If you tell me which time zone you used, I'll do my count for unique Merit senders up to January 26.

Im not really sure why is that, i have about 16418 user registered before 24/1/2018 which sent merits, this seems to be consistent with what i can count in the graph posted above so i think i don't have them.
The total number of users i have is 29039, which comprehend anybody who ever received or sent merits.

While scraping i remember to notice some user that had completely vanished (personal page does not exist anymore) but i didn't keep track of who or why.

It could also be a time difference indeed.

Since it's so close to the introduction of Merit, with large amounts per hour, could this be caused by using a different time zone?

Sure, a 10 merit requirement for a signature would go a long way to reduce many of the scammy attempts we saw during the first few days after the rule changed for Jr. Members. Nevertheless, the change has likely had a rather large impact as it is, and I figure the idea is to give Newbies a bit of a chance, even if there will be cases (so long as it is not hordes of them) that bypass the new requirement through their alts.

The spike was large during the first week on the new Sender’s side, but seems to be reducing its pace as can be seen on your chart (and my prior update sMerit participants – A quick follow-up after the changes – 20180928). So even if bypassing through Alts is a thing, it seems to be decreasing rather rapidly, as the number of new Receivers is falling promptly to numbers that seem more logical than the first week’s spike.

Note: I tried to contrast the numbers in the chart against mine (in the DB), and we don’t seem to match for some reason I cannot see. For example, for the period 24/01/2018 (00:00:00) ..26/01/2018 (23:59:59), and account created before the 24/01/2018, I get 2.987 first time senders, while your first column shows 2.076. That is quite a large gap … Any idea why?

Yes, I support raising the requirement to have any kind of signature to 10 merits. It would be one step further, wouldn’t affect traffic much and would make it more difficult to abuse the system.

With the 1 merit requirement we have seen some people making efforts to improve, although not many of them. I think with the 10 merit requirement more people would try to make a decent contribution to the forum, because it is not a too hard requirement, it is achievable, even if you are not a great poster.

Few considerations about what happened between 17/9 and 21/9, users which got and sent their first merit ever and how they are linked together.

After the new restriction applied on jr. members (17/9) in the merit system, there has been a spike on the merits distribution, part of it is because more merit sources were added and more merits were granted to them, part of it, is what i’m trying to explain in here.

I looked at the number of users which actually started sending merits for the first time ever during the period from 17/9/2018 and 21/9/2018, while being registered in the forum before the introduction of the merit system 24/1/2018 and found out there are 390 users which meets this requirement, which account for 1402 merits sent, about 20.24% of the 6928 sent in that period of time.

If we compare a period of 1 week  antecedent 17/9/2018 and we go back till the starting of the merit system we can actually see 390 is a considerable amount, to see that kind of users we must go back till March when the merit system was relatively new. Moreover just one week before the restrictions, we were in a descending trend with about 95 new merit senders. Have a look here:


On a side note: We can also notice this graph seems to “follow” the merit distribution graph by week, this could be a sign that airdropped merit could still play a big role.


To close the circle we need to find how many of these 390 users sent merit to people which never had merit before (note that in this category can also fall a legendary which never got a merit or any other rank for what can matter, but we can see later who actually was the receiver of these merits). The number of people fitting the profile is 353.

The total number of user that received their first merit was 1691 and of these, 606 were merited by these 353 first timers, for a total of 1207 merits

This is surprising imo, since i was actually not counting on finding these many connections and by putting together all the information provided here it's pretty clear that is not a coincidence in most of these cases.
This make me think about the suggestion to raise the minimum amount required to have a signature to 10 merits or more than the single actual merit. If that ever happen it may be certainly interesting to see how many new users will appear and the number of merits poured in the aftermath, as it clearly an indication that many can bypass it quite easily.

---

I provide in here a list of the Sender - Receiver and since the merit assigned are still fresh you can actually verify everything is as i said. You can also double check with the full merit history using any of the tool the community made available so far.

This is the couple Sender receiver, with rank(there may be legit people as well in here, everyone is innocent until proven otherwise):
https://drive.google.com/open?id=1l4p3_85bf_OLvnadQSRv9hmjubrJHJ5Q0ejcTHSX_Nc