Author

Topic: Stats on most actively users in merit activites. (Read 245 times)

tranthidung

legendary

Activity: 2310

Merit: 4085

Farewell o_e_l_e_o

The table updated a little bit. Initially, I got the top 1% by their ranks, so 11537*1% ~ 116 (that figure you used) but then I realized there are two users have same total earned merits, so I turned to use the cutoff for the top 1% by their total earned merits >= 645. After all, there are 177 users in the top 1%, not 116.

Fortunately, results don't change too much.

I have what you need too (I did not totally understand what you ask for in my previous post).

Code:

+---------------------------------------------------------+
   | temerit_group totalmerit pemerit_group group50 |
   |---------------------------------------------------------|
  1. | 20405 433079 4.7 50% bottom |
  2. | 274318 433079 63.3 50% top |
  3. | 138356 433079 31.9 1% top |
   +---------------------------------------------------------+

Your roughly estimations are not too different from correct ones. Congrats! Cheesy

Another shocking fact, the top 1% richest in term of total earned merits own 32% of total earned merits in the forum, on the last week data dump by theymos. And, the richest will get richer more fastly than the others.

Fortunately, if one can manage to earn around 8 merits per week, that one will surely move upwards to the top, after months. Because of the top 100 merited users (take at 2019w36) earn about 8 merits per week, in median, over 27 weeks (2019w23 to 2019w49).

Quote from: tranthidung on December 15, 2019, 10:06:51 PM

For four groups of the top 100 merited users:

Code:

. tabstat meritchange , s(n mean sd p50 p25 p75 min max) by(group)

Summary for variables: meritchange
by categories of: group

group | N mean sd p50 p25 p75 min max
-------+--------------------------------------------------------------------------------
1-25 | 650 20.12462 28.1251 13 4 28 0 501
26-50 | 650 12.98308 15.66713 8 2 18 0 129
51-75 | 650 10.14769 12.49282 7 1 14 0 107
76-100 | 650 10.27692 12.1469 7 2 14 0 112
-------+--------------------------------------------------------------------------------
Total | 2600 13.38308 18.737 8 2 18 0 501
----------------------------------------------------------------------------------------

Visually, we have the one

nutildah

legendary

Activity: 3010

Merit: 8114

Quote from: tranthidung on December 20, 2019, 01:45:17 AM

Code:

Summary for variables: temerit_user
   by categories of: group50

   group50 | N mean sd p50 p25 p75 min max
-----------+--------------------------------------------------------------------------------
50% bottom | 5949.0 3.4 2.3 3.0 2.0 5.0 1.0 9.0
   50% top | 5472.0 50.2 85.7 18.0 11.0 45.0 10.0 645.0
   1% top | 116.0 1187.2 752.1 1012.0 795.5 1258.5 645.0 6354.0
-----------+--------------------------------------------------------------------------------
   Total | 11537.0 37.5 151.9 9.0 3.0 18.0 1.0 6354.0
--------------------------------------------------------------------------------------------

That was fast, thanks.

I think the numbers I'm looking for here are

Number of merits owned by the top 1%: 116 * 1187.2 = 137,715
Number of merits owned by the bottom 50%: 5949 * 3.4 = 20,227

That means the top 1% of posters combined possess 6.8x as many merits as the bottom 50%.

The bottom 50% have less than 15% of the top 1%'s ownership.

Furthermore, the top 25% of the top 1% still possess 180% more merits than the bottom 50%.

That means that 29 members together have almost twice as many merits as 5,949 combined.

Its not fair that in this great society, half the accounts on this forum should be expected to live on 9 merits or less, especially when 9 merits isn't enough to earn a Member level status.

If I am elected president, I will pledge to fight merit inequality in a system which disproportionately favors the merit rich and punishes the merit poor. Thank you for your vote.

tranthidung

legendary

Activity: 2310

Merit: 4085

Farewell o_e_l_e_o

Quote from: nutildah on December 20, 2019, 01:24:01 AM

What is the number of merits earned by the top 1% of the 11,537 legit users as compared to the bottom 50%?

You got it.

* Cutoffs:

p50 and p99 in the table.

50% bottom: temerit_user <= 9
50% top: temerit_user > 9 & < 645 (excludes 1% top)
1% top: temerit_user >= 645

Code:

variable | N mean sd p50 p25 p75 p99 min max
-------------+------------------------------------------------------------------------------------------
temerit_user | 11537.0 37.5 151.9 9.0 3.0 18.0 645.0 1.0 6354.0
--------------------------------------------------------------------------------------------------------

Code:

group50 | Freq. Percent Cum.
------------+-----------------------------------
50% bottom | 5,949 51.56 51.56
   50% top | 5,471 47.42 98.99
   1% top | 117 1.01 100.00
------------+-----------------------------------
   Total | 11,537 100.00

Code:

Summary for variables: temerit_user
   by categories of: group50

   group50 | N mean sd p50 p25 p75 min max
-----------+--------------------------------------------------------------------------------
50% bottom | 5949.0 3.4 2.3 3.0 2.0 5.0 1.0 9.0
   50% top | 5471.0 50.1 85.3 18.0 11.0 45.0 10.0 641.0
   1% top | 117.0 1182.5 750.6 1011.0 783.0 1255.0 645.0 6354.0
-----------+--------------------------------------------------------------------------------
   Total | 11537.0 37.5 151.9 9.0 3.0 18.0 1.0 6354.0
--------------------------------------------------------------------------------------------

nutildah

legendary

Activity: 3010

Merit: 8114

What is the number of merits earned by the top 1% of the 11,537 legit users as compared to the bottom 50%?

I'm trying to do some Bernie Sanders math here.

tranthidung

legendary

Activity: 2310

Merit: 4085

Farewell o_e_l_e_o

From this (Merit - simple poll on operational fundamentals), I create this one.

Users who have recorded activities with merits: 39341
users who sent merits only: 7398
Users who received merits only: 17922
Users who both received & sent merits: 14021

From that 14021 users:

Autobanned: 2470
Nuked: 14

Then we have the rest of 11537 legit users who both sent & earned merits and have not yet been autobanned or nuked.

See details here

Now, let's move further by looking at only the 11537 legit users' merit history.

Merit sent activities:

Means +/- standard deviation: 21.5 +/- 153.7
Median (interquartile range): 3 (1 - 6).
Minimum - maximum: 1 - 5088.
Total sent merits: 472 545.

Code:

variable | N mean sd p50 p25 p75 min max
-------------+--------------------------------------------------------------------------------
nsent | 11537.0 21.5 153.7 3.0 1.0 6.0 1.0 5088.0
----------------------------------------------------------------------------------------------

Merit earned activites:

Means +/- standard deviation: 19.5 +/- 81.4
Median (interquartile range): 3 (2 - 9).
Minimum - maximum: 1 - 2394.
Total earned merits: 433 079.

Code:

variable | N mean sd p50 p25 p75 min max
-------------+--------------------------------------------------------------------------------
nearned | 11537.0 19.5 81.4 3.0 2.0 9.0 1.0 2394.0
----------------------------------------------------------------------------------------------

In a nutshell, we have the medians at 3 times (from the sample size of 11537 legit users).

The upper population

Let's assume users who have stats >= p50 (for each stats) are more active ones. I skipped data for that p50-above population, to see how stats change.

The results show that:

The upper population have nearly double statistics compare to the whole population.
The upper population have 5776 - 6922 users only. You might be shocked.

Details
Merit-earned activities:
Sample size: 11537.

Total number of times of merit-earned activities.

Code:

Total earned merits over users.

Code:

variable | N mean sd p50 p25 p75 min max
-------------+--------------------------------------------------------------------------------
temerit_user | 11537.0 37.5 151.9 9.0 3.0 18.0 1.0 6354.0
----------------------------------------------------------------------------------------------

Let's see data for the upper 50-quartile users (I skip users have stats >= p50).

Total number of times of merit-earned activities:
Median changed from 3 to 7.
Sample size decreases to 6922.

Code:

variable | N mean sd p50 p25 p75 min max
-------------+--------------------------------------------------------------------------------
nearned | 6922.0 31.6 103.4 7.0 4.0 17.0 3.0 2394.0
----------------------------------------------------------------------------------------------

Total earned merits over users:
Median changed from 9 to 17.
Sample size decreases to 5776.

Code:

variable | N mean sd p50 p25 p75 min max
-------------+--------------------------------------------------------------------------------
temerit_user | 5776.0 71.7 209.2 17.0 10.0 47.0 9.0 6354.0
----------------------------------------------------------------------------------------------

Merit-sent activities:
Sample size: 11537.

Total number of times of merit-sent activities.

Code:

Total sent merits over users.

Code:

variable | N mean sd p50 p25 p75 min max
-------------+--------------------------------------------------------------------------------
tgmerit_user | 11537.0 41.0 377.4 5.0 2.0 13.0 1.0 22228.0
----------------------------------------------------------------------------------------------

Let's see data for the upper 50-quartile users (I skip users have stats >= p50).

Total number of times of merit-sent activities:
Median changed from 3 to 6.
Sample size decreases to 5812.

Code:

variable | N mean sd p50 p25 p75 min max
-------------+--------------------------------------------------------------------------------
nsent | 5812.0 41.3 214.8 6.0 4.0 16.0 3.0 5088.0
----------------------------------------------------------------------------------------------

Total sent merits over users:
Median changed from 5 to 11.
Sample size decreases to 6412.

Code:

variable | N mean sd p50 p25 p75 min max
-------------+--------------------------------------------------------------------------------
tgmerit_user | 6412.0 72.2 504.1 11.0 6.0 34.0 5.0 22228.0
----------------------------------------------------------------------------------------------