Topic: Which Ranks send sMerit to which Ranks - and who ranked up (Read 960 times)

You dont need to be an excellent one to contribute for this forum. You just need a fresh idea and deliver it to this forum, and i guess there are still a lot of idea to discover out there. Even though i made some statistic back there, but DdmrDdmr took it to another level with a great new idea, and that is what you must aim for your posts/thread. You already have an idea in your post upside, and you know that it is an interesting idea, why not do it yourself? Although the result will have a lot of variable because it is highly subjective.

Ahh!

I know this is going to come across as whining, but I knew people were hoarding their sMerit!!! It would be an interesting idea to have a poll of people with sMerit saved and ask them what they think is a meritworthy post. If it was broken down by member rank that would be even more interesting. I think the legendary and hero crowd have very different reasons for meriting a post than regular members.

This is some really great data. Hopefully it will help me learn how to earn more merit. I don't have much to add to most technical discussions since I am not as technically savvy as many on this forum.

Agree with your serious work, I have no Merit, otherwise I will send a Merit to you!

Right, cheers for the insights.

I'm taking a look at how things are going for new users that registered on the day the merit system was set in motion (just the new registered users), and the first surprise is that, out of the first 40k accounts created, 30k roughly are still ranked Brand New with cero posts. What you say covers a large part I guess, but it is still a tremendous amount (roughly 75% of first week post merit kickoff created users).

I'm still gathering data and seeing what can be made of it, but my idea is to work on the first 2 weeks of created users and see how they've got on in the forum.

Lots of accounts are signed up by spambots and then get 'nuked' by the moderators. This bans them and also deletes all their posts. It is still possible to login to a banned account as it is just banned from posting and sending PMs. So it could be the bots still trying to post from a banned account.

Other reasons to register an account beyond the search function are to use the watchlist and send PMs.

I don't quite get your point. Users are being rewarded if active, provided that their posts are seen by the community as worthy (of sMerit). Inactive users are not punished directly, nor should they in my opinion.
By the way, there seems to be an enormous amout of users that create an account and leave it there with cero posts, although they may be login in every now and then. I don't get the point of creating the account in these cases, since the forum may be browsed (unless it is to gain access to the search feature).

Selling accounts is not prohibited but, as I see it, buying one is not the smartest way to start promoting projects, since the starting point is already flawed by having a user pretending to have a certain status and implied knowledge in the forum which he has not.

I'm sure there will be a massive decrease. I don't know if the old data is available but I would be very interested to see the number of Member > Full Member after the first 120 days are over compared to the 120 days before.



Don't think it's working as I keep saying 'Please Sir' but my politeness won't budge 

True that in order for the comparison to be fair we would have to compare individuals that have had enough time to rank up in the two month window, ideally comparing users starting from the same baseline. A quick comparison would give us an idea of throughput decrease though.

Regardless, theymos’s new rank system will shatter it all, at least for the remaining of the day...

I understand the first bit now.

On the second point, I still don't think it can be a valid comparison yet. As a Sr. Member when the system was introduced I was 7 weeks away from getting the activity but still had to start at 250 Merit like every other Sr. Member. Until they have all had 8.5 months to earn the Merit we can't compare rank up rate to previous data. In order to rank up at the same rate as before each member needs to earn Merit at a rate of ~1 per day, not 250 in about 60 days. It is still an interesting stat that you are doing, but the comparison will take time to be meaningful.

Khalis_3, I'm the thread OP. Don't hit your head over language on this forum.  TMan's comment is not offensive at all; rather the opposite. I'ts usual common english language and, when used properly, as is the case has the opposite meaning to what you may be interpreting.

What does you told men, you say so dirty in this forum. Why?

Hi,
would you mind sharing me your dataset, please.
I would like to do basic statistical analysis about merit circulation in the forum. Unfortunately, I have no skills to get data from the forum. The raw dataset Theymos gave us contains too little information.
Thanks in advance for your help.

Ty for data mate, I tried to do something with those infos. I've ignored sMerit sent from copper/VIP/staff/founder/mods cause in this way graphs are cleaner and I think they are under 0.01% of population of the forum, only pure ranks from members!

% sMerit sent from -> to

I did some work on this a couple of weeks ago, and posted my findings here: Additional sMerit Analysis (smerit.txt):    https://bitcointalksearch.org/topic/additional-smerit-analysis-smerittxt-3167331

It's a rather tough post I must admit, but the interisting part to work on what you mention is down on "4) Given/Received sMerit I (Tab 4):" There I give a full relation of who sent how much sMerit to whom and got how much in return.

These cases are where we can find candidates to account farming that pass sMerit around their accounts. My Google Sheet has all the data needed to spot these cases:
https://docs.google.com/spreadsheets/d/1Su1vulMzsbkYPOsU8llr8PrQXRPvy4oM_E218klwHh8/edit#gid=968451010.
Look at Tab 4 specially. It is not Rank related since I did not have this information at the time.

Unfortunately, to catch cases of merit abusers or farms you need to go one by one on each suspect, since proof is down to a qualitative inspection after using a quantitative instrument such as my worksheet to narrow down suspects. I don't know who's task that would be but this is how I would approach the problem.

Also I would make the caught in action users numbers more visible in the forum as a warning for farmers and merit abusers. Currently, therer are a lot of reports of farmers and merit abusers, but no clear report on overall results.

Yes, that too is part of what can be looked into, although from the Rank point of view, which is what is being treated in this analysis, it would be maybe a bit too far fetched (trying to dysect which part of one's rank came from which forum section).

If your interested, there's a good breakdown by Zentdex ->   Where the Merit Pours:     https://bitcointalksearch.org/topic/where-the-merit-pours-3093768

That would be the ideal data for the comparison, but if the data called whos ranked up since merit introduced, then i guess it is a fine data, although we can not make a comparison between old member who already has the activiry point or half way with new member who joined after merit system introduced, but this data could be a motivation for all people who thinks that ranking up is impossible in merit age.

I bet that probably at least 25% of the higher end users that upgraded have worked together as some point. 1/4 of the new Hero members are a proven farmer, so you can only imagine how much easier it would be for the lower end members to achieve the next rank. I would suggest looking deeply into many of these accounts, as I am pretty sure that most are suspicious in some way.

True, everybody's before number should be equal to their rank. I´ll explain what I meant in my referenced comment about the Before Data not being 100% exact:

- Merit Txs were based on the extraction that runs up to the 23/03/2018.
- Rank data is extracted with a snapshot for each user somewhere between 28/03/2018 and 29/03/2018 (quite a few hours are required, so snapshot is not simultaneous for every user whose data was extracted).

I worked back from the current Merit Balance (28/29th), substracting the Merit awarded, to get to the initial Merit status of each user.

There are nevertheless some gaps due to the underlying raw data being analyzed having a few days difference. Basically, I don't have in this analysis all the Txs until the exact moment that each individual user's snapshot was taken (lacking 5 days transactions to be precise).

Let's say for example user X had 60 Merits on the 28/03/2018 and the sum of Txs for the user shows he was awarded 50 Merits since the kickoff until the 23/03/2018. That would mean that 10 Merits are non justified from the Tx, so they could either be:

- Given in the initial airdrop (10 lets say)
- Or have been generated between the 23rd and the 28th.

I cannot tell the exact origin of these 10 Merits with non-100% synchronized data sets, but I do know that the user X does indeed have 60 Merits (at least on the 28/29th), and at leat 50 received through the TXs.

The above should, notwithstanding, have little effect in the overall picture of the analysis presented. Future updates (If demanded..) could be made to shorten that window to a max. of 1 day lag between the two datasets implied.
 

In relation to the second point you mentioned, I think it can be done: If we take any consecutive two month window from last year and knew how many users had ranked up in that window frame (regardless of how it took them to get there), we could compare to current two month timeframe.

Say for example Feb/Mar last year has 30 users rank up to Sr. Members (however long it took them). We could compare those to current rank up data (18 Sr. Members). Ofcourse context is different and should be stated to interpret the results.

Fuck me OP this is decent work. You keep this up every month and you will rank up based on repeating these stats. REally is great to see someone embracing the system and using talent and time to put something to the community, well done dude

Just take a look on ranked up hero members.
One of them Turkish-local mostly, hard to evaluate.
Second with red trust for merit farming.
Third looks like truthworthy member.
And for fourth (gawlea):
no red trust, posts in English, but:
1) some posts with 50 merits from one person:
Those oneliners really deserve 50+ merits?
2)
a) January 28, 2018, 06:42:59 PM:  gawlea receives 2 merits from S3cco
January 28, 2018, 06:56:31 PM: S3cco receives 5 merits from gawlea
b) January 30, 2018, 08:16:53 PM: S3cco receives  2 merits from gawlea
January 30, 2018, 02:21:19 PM:  gawlea receives 1 merit from S3cco
January 31, 2018, 12:38:25 PM:  gawlea receives 2 merits from S3cco
c) February 14, 2018, 12:52:46 PM: gawlea receives 1 merit from S3cco
February 15, 2018, 11:09:13 AM: S3cco receives  3 merits from gawlea

Not proofs anything for sure, but seems strange to me.

Heroes that we deserve.