Quote from: Jawhead999 on March 07, 2020, 01:03:50 PM

I think this thread is the answer [CLUBS] Top Merited-Users Classified into 4 Clubs
This data has been started from February 28,2019 till now, usually update every week.

That one is for whole time since the merit system was kicked off, but I can also do a list of top merited users in a given period, like this one.

Quote from: tranthidung on January 14, 2020, 07:10:55 AM

Due to respect to merit sources and the core ideas of keeping sources as in secret as possible, I don't give data on them here.

I have some summary statistics and box plots for the top 100 meriters, during whole time (2018w4 to 2020w2), and in specific years (2018, 2019).
From the raw tables, you can see the median of weekly gave-away merits from the top 100 meriters, over all 4 groups have increased over years. You can also see it from the box plot, medians represent by these lines inside each boxes.

There are some factors:

They earned more by themselves.
Their merit giving-away behaviour changed.
Their monthly sMerit allocation increased

Expectation:
I do expect the median of weekly gave-away merits will keep increasing in the year 2020. New merit sources and new rising stars rock!

Box plot:

Summary statistics:

Whole time (2018w4 - 2020w2):
* First and last week are incomplete.

Code:

1-25 | 2575.0 64.2 185.5 35.0 11.0 77.0 0.0 8462.0
  26-50 | 2575.0 17.8 26.6 10.0 1.0 24.0 0.0 401.0
  51-75 | 2575.0 9.5 15.9 5.0 1.0 12.0 0.0 331.0
76-100 | 2575.0 7.0 11.9 4.0 0.0 9.0 0.0 197.0
--------+--------------------------------------------------------------------------------
  Total | 10300.0 24.6 97.0 8.0 1.0 24.0 0.0 8462.0
-----------------------------------------------------------------------------------------

In 2018:

Code:

Summary for variables: amount18
   by categories of: group

  group | N mean sd p50 p25 p75 min max
--------+--------------------------------------------------------------------------------
   1-25 | 1225.0 44.2 62.0 22.0 9.0 53.0 0.0 652.0
  26-50 | 1225.0 10.6 17.4 4.0 0.0 14.0 0.0 181.0
  51-75 | 1225.0 6.5 10.8 3.0 0.0 8.0 0.0 121.0
76-100 | 1225.0 5.9 10.8 3.0 0.0 7.0 0.0 167.0
--------+--------------------------------------------------------------------------------
  Total | 4900.0 16.8 36.7 5.0 0.0 16.0 0.0 652.0
-----------------------------------------------------------------------------------------

In 2019:

Code:

Summary for variables: amount19
   by categories of: group

  group | N mean sd p50 p25 p75 min max
--------+--------------------------------------------------------------------------------
   1-25 | 1300.0 84.1 252.3 48.0 16.0 101.0 0.0 8462.0
  26-50 | 1300.0 24.1 31.3 16.0 5.0 31.0 0.0 401.0
  51-75 | 1300.0 12.3 18.8 8.0 2.0 15.0 0.0 331.0
76-100 | 1300.0 7.9 12.7 5.0 1.0 10.0 0.0 197.0
--------+--------------------------------------------------------------------------------
  Total | 5200.0 32.1 131.2 11.0 3.0 31.0 0.0 8462.0
-----------------------------------------------------------------------------------------

However, I don't see convincible reasons to do that. If I want to create a list of most regularly merit-earners, I would go to the median of their daily earned merits (the p50) and sort them out descendingly to get the top most regularly merit-earners.

Such list reflect both three factors:
- Their total earned merits.
- Their regularity in merit-earning.
- Their level of stable activitites.

I also have this analysis, from which I calculated the difference between medians of top merit-earners in the two period (two years, 2018 and 2019). See details
You can look at it and see the changes (represented by medianchange) shows how one user changed activities and helpfulness on the forum between 2 years.

As being written, I would not go with it (even I can do this).