Statistics of user registrations on Bitcointalk 2017-2020

[bitart]

Are there some statistics about the number of users that have been banned/nuked?
And in the case of banned members, about their ranks?

I think it would be interesting to see how many of the freshly registered users are shown the exit door daily.

I have a feeling that each time the mods are following the lists of LoyceV and others, after a few hours the number of fresh registrations spikes again  

Are the nuked users marked also as Newbies, without any post? If mods nuke a newbie, who already posted something, all of the posts are deleted too. This means for me that the newbies with 0 posts could be partially nuked users also. (And I'm not surprised if the number of these kind of users are growing...)

[1713305785]

1713305785

[1713305785]

1713305785

[1713305785]

1713305785

[1713305785]

1713305785

[Alex_Sr]

UPD: Added statistics for August

Month	Registered	Total at the beginning of the month	% increase	First user of the month
August	54 384	2 318 752	2,35%	First user of Aug, 2018

[LoyceV]

Are the nuked users marked also as Newbies, without any post? If mods nuke a newbie, who already posted something, all of the posts are deleted too. This means for me that the newbies with 0 posts could be partially nuked users also.

Nuked users are shown as "Brand new" (example).

I'm curious how many users are nuked. I'll make an estimate based on a sample of the data: I start with the first post of the month, then add 10 to the userID. This means I scrape 10% of all users from http://bpip.org/, to get an estimate of the percentage of users that have been nuked in a month.

Update: I'm not sure if BPIP has ban-data from January 2018, and I didn't want to continue the slow scraping process.
I expected a large percentage of accounts to be banned, but it's really not that much. Most new accounts (about 75%) never made a single post.
That leaves just data for August 2018: I've scraped 10% of all accounts (starting at 2318753, ending at 2373137): 3 users have been Autobanned and 65 users have been Nuked (5439 accounts checked). That's only 1.25% of all accounts, or approximately 5% of all accounts that have ever posted.
I'm currently checking July 2018, then I'll stop. I'm curious to see if a higher percentage gets banned a month later.
Update: 18 users have been Autobanned and 167 users have been Nuked (6984 accounts checked).
In short:
2.65% of the accounts created in July are banned.
1.25% of the accounts created in August are banned.

[bitart]

I don't know the background, but I think farmers are not registering the new accounts only on demand. I think they are creating accounts here continously, and they sell/rent them when a customer drops in.
So to cut the long story short, I think they are using the FIFO method when they sell/rent their farmed accounts. This means when a customer gets some accounts from farmers, those accounts are quite old already (but without any posts). So if this is true, your statistics can be true that there are only a few accounts banned in a few months time after the registration.
If you are after the spammer accounts, I would do it just the opposite way (but I don't have the knowledge to the technical and programming part to e.g. write a query, etc...) If you could check how old the accounts who were banned at a specific time, it could show something different, I guess.

[Alex_Sr]

Added statistics for September

Month	Registered	Total at the beginning of the month	% increase	First user of the month
September	45 362	2 373 136	1,88%	First user of Sep, 2018

[Alex_Sr]

Added statistics for October 2018

Month	Registered	Total at the beginning of the month	% increase	First user of the month
October	39 354	2 418 498	1,60%	First user of Oct, 2018

[Alex_Sr]

Added statistics for November 2018

2018

Month	Registered	Total at the beginning of the month	% increase	First user of the month
November	35127	2 457 852	1,41%	First user of Nov, 2018

[stompix]

August	54384
November	35127

Wow, that's quite a drop and pure coincidence, but still -35.40%  is so close to the fluctuation in bitcoin price since mid-August till now (6500>4000), around 38%

Probably it's not just price but also the introduction of the 1 merit requirement, the diminishing interest in ICOs and thus the demand for bumping bots but if this bear market continues we will probably see less than 500 new accounts a day.
I wonder how many of those are genuine...

[Alex_Sr]

August	54384
November	35127

Wow, that's quite a drop and pure coincidence, but still -35.40%  is so close to the fluctuation in bitcoin price since mid-August till now (6500>4000), around 38%

Probably it's not just price but also the introduction of the 1 merit requirement, the diminishing interest in ICOs and thus the demand for bumping bots but if this bear market continues we will probably see less than 500 new accounts a day.
I wonder how many of those are genuine...

It is impossible to find out the number of genuine accounts - too many bots are registered. Perhaps genuine is 1 in 50 and perhaps 1 in 1000...

[DdmrDdmr]

<...>It is impossible to find out the number of genuine accounts - too many bots are registered. Perhaps genuine is 1 in 50 and perhaps 1 in 1000...

This could give us a clue: I’ve used @Piggy’s full dataset (covers full months up to October 2018) to count the number of registered accounts with 0 posts. Now there could be a number of reasons for registering and having no posts (don’t want to, nuked (*), etc.), but I’d say, more likely than not, many cases are botty created accounts.

On overall average, 73,18% of all created accounts have 0 associated posts.
 

Code:

Year      Month     nAccounts 0Posts    %0Posts
2018      10        39353     31859     80,96%
2018      9         45362     35917     79,18%
2018      8         54384     42815     78,73%
2018      7         69840     53517     76,63%
2018      6         77681     57388     73,88%
2018      5         91993     66755     72,57%
2018      4         86898     63823     73,45%
2018      3         97644     69923     71,61%
2018      2         112211    82573     73,59%
2018      1         214546    171502    79,94%
2017      12        191455    155404    81,17%
2017      11        97286     68276     70,18%
2017      10        89443     59339     66,34%
2017      9         46186     26730     57,87%
2017      8         42254     25514     60,38%
2017      7         41011     27543     67,16%
2017      6         32733     20032     61,20%
2017      5         23978     15127     63,09%
2017      4         15807     9820      62,12%
2017      3         14760     9495      64,33%
2017      2         12677     8227      64,90%
2017      1         13354     9003      67,42%
2016      12        11199     7484      66,83%
2016      11        11187     7254      64,84%
2016      10        9935      6780      68,24%
2016      9         9923      6595      66,46%
2016      8         11193     6832      61,04%
2016      7         11257     7794      69,24%
2016      6         20630     16177     78,41%
2016      5         16918     13434     79,41%
2016      4         18493     14798     80,02%
2016      3         24945     21438     85,94%
2016      2         40436     36777     90,95%
2016      1         39178     35480     90,56%
2015      12        39950     36841     92,22%
2015      11        44059     40887     92,80%
2015      10        68831     65930     95,79%
2015      9         7727      5046      65,30%
2015      8         10942     8105      74,07%
2015      7         12616     9489      75,21%
2015      6         9310      6630      71,21%
2015      5         8037      5571      69,32%
2015      4         8914      6094      68,36%
2015      3         38297     34399     89,82%
2015      2         34357     30799     89,64%
2015      1         20794     17206     82,75%
2014      12        8391      5061      60,31%
2014      11        7996      4490      56,15%
2014      10        8332      4606      55,28%
2014      9         10158     5825      57,34%
2014      8         11443     6699      58,54%
2014      7         11119     6434      57,86%
2014      6         11952     7129      59,65%
2014      5         14264     7685      53,88%
2014      4         17948     11318     63,06%
2014      3         33521     24153     72,05%
2014      2         25450     13680     53,75%
2014      1         26619     12079     45,38%
2013      12        31775     15383     48,41%
2013      11        19680     10776     54,76%
2013      10        2927      1562      53,37%
2013      9         5155      2738      53,11%
2013      8         6052      3410      56,35%
2013      7         7413      3764      50,78%
2013      6         9372      5197      55,45%
2013      5         15512     8771      56,54%
2013      4         18712     8986      48,02%
2013      3         7357      3453      46,93%
2013      2         3532      1794      50,79%
2013      1         3184      1859      58,39%
2012      12        2765      1691      61,16%
2012      11        2708      1763      65,10%
2012      10        3199      2067      64,61%
2012      9         2345      1268      54,07%
2012      8         2394      1230      51,38%
2012      7         2109      1197      56,76%
2012      6         2079      1318      63,40%
2012      5         2017      1265      62,72%
2012      4         2071      1417      68,42%
2012      3         2117      1518      71,71%
2012      2         1991      1275      64,04%
2012      1         2051      1287      62,75%
2011      12        1719      1304      75,86%
2011      11        1990      1545      77,64%
2011      10        2060      1469      71,31%
2011      9         2443      1706      69,83%
2011      8         3856      2511      65,12%
2011      7         5451      3278      60,14%
2011      6         14434     7318      50,70%
2011      5         5866      2947      50,24%
2011      4         2822      1864      66,05%
2011      3         2423      1760      72,64%
2011      2         1090      564       51,74%
2011      1         376       179       47,61%
2010      12        422       161       38,15%
2010      11        218       115       52,75%
2010      10        174       76        43,68%
2010      9         167       92        55,09%
2010      8         163       82        50,31%
2010      7         333       58        17,42%
2010      6         34        8         23,53%
2010      5         15        2         13,33%
2010      4         22        1         04,55%
2010      3         5         0         00,00%
2010      2         22        0         00,00%
2010      1         8         0         00,00%
2009      12        8         0         00,00%
2009      11        3         0         00,00%
 

(*) According to @suchmoon’s post on banned accouts (see [MERIT][NSFN] We're destroying more merits than we're sending. Like, way more), between 1,2k and 1,7k accounts are nuked every month, thus causing these accounts to have 0 posts.

[stompix]

Weird numbers and there is a clear trend upwards from 71% to 80%, merit impact?

But if only 1-2k are nuked there are still something like 20k just lurking, again strange, but probably they are leaving them to age a bit before starting to use them as bumpers  Although I see no reason for it...

80% leaves only 7000 real users if we remove another 90% for alts and another 90% for bounty only and again 99% for future shitposters    it's probably what Alex_Sr said, one in a hundred.

[DdmrDdmr]

Weird numbers and there is a clear trend upwards from 71% to 80%, merit impact? <…>

I don’t think so. Merit should not lead people to create accounts but not create a single post (which is what I’ve measured). Likely it’s something like the newer the non-posting account, the less chance it has had to post over time. The metrics, going backwards in time, show less number of accounts with cero posts as we go back in time (likely because of what I mentioned above).

There are a few exceptions though, that seem to surge when there are abnormal account creating periods (compared to previous months), and that have a higher number of 0 post accounts than expected:

  - December 2017.. January 2018 (big boom)
  - November 2015.. June 2016 (a few months after the account hack)
  - January 2015 .. March 2015

I went over @Piggy’s dataset to see bot creating patterns and bursts. I think the largest I detected with recent dates was between the 28/09/2018 and the 29/09/2018, with over 7K accounts created in that period, where most of the usernames included upper/lower/number combinations.
I’m not sure why they do it. I think it may be a phased feature: First create a bunch of bot accounts (possibly not too tough), and then move on to trying to use bots to provide them dummy content for their posts (and join campaigns). The latter is likely way more difficult than the former, and many will not succeed, leaving bags of created accounts with cero posts. Of course I could be wrong.

[bitart]

Maybe they're testing the forum software or their own code for bots. I don't know how many accounts the forum software can handle but his could be a kind of attack to slow down the forum software day by day... Or, they write a code to create accounts, they have to test if if it's working or not. Also they have to run a stress test, to be sure what is the maximum amount of bot accounts they are able to create when a customer asks, they should be able to give exact timing, like: we're creating you 2,000 fresh accounts in 3 hours time... just wondering...
Maybe it's easier for the customers to just buy some already registered accounts instead of trying to register hundreds of them from the same IP (but if I remember well, recaptcha will stop them and their IP will be banned too if they want to register too many accounts).
So for an average bounty hunter, who's skills are coding rather than creating accounts, it's just easier to buy some of these accounts, and all he have to do is start running his own code, nothing else.

So to cut the long story short, why don't we just lock all the newly created accounts if they don't post in a few weeks time (I know there are newbies who are here to read first and they will post only if they feel confident about the forum, bitcoin, etc... and it takes time, but they are not the majority now unfortunately...)

[thongexiphstaff]

For the stats in December 2017 between the topic figure and the forum official stats.
I known the reason, that is due to the lack of official stat on Dec 2017 (only counted till the 13th December 2017, or nearly the first half of the Dec, not a full Dec).
Therefore, it's normal to see the big difference between topic stats and forum official stats.

Comparing your stats to the forum’s oficial stats (https://bitcointalk.org/index.php?action=stats) there are a few discrepancies:
< ... >
-   December 2017 is off quite a bit (191 537 vs 63 756 official)
December’s number of registered users seems a bit small on the official stats, considering that BTC peaked on the 16th of December, so I always wondered if they were correct for December 2017. I've just seen that, oddly enough, they only go up to the 13 of December (clicking the detail on the official Stats), so yours seem coherent with events.

Have you known why?
If you known it, can you share it here, please.

Forum official stats are not available for some reason for 2018.

[Alex_Sr]

Added statistics for December 2018

2018

Month	Registered	Total at the beginning of the month	% increase	First user of the month
December	23 504	2 492 979	0,93%	First user of Dec, 2018

2019

Month	Registered	Total at the beginning of the month	% increase	First user of the month
January	??	2 516 483	??	First user of Jan, 2019

[Alex_Sr]

Added statistics for January 2019

2019

Month	Registered	Total at the beginning of the month	% increase	First user of the month
January	22 234	2 516 483	0,88%	First user of Jan, 2019
February	??	2 538 716	??	First user of Feb, 2019

[tranthidung]

This one is not exactly what you want, but maybe helpful for your interests.
I made it by using reliable data dump by theymos, but it is not exactly monthly statistics.
Please use it with potential biases.

As I promised yesterday, I made some quick descriptive analysis today.

Time-series plot

Basic statistics:
- Mean +/- standard deviation: 21916 +/- 34795
- Median (interquartile range): 10145 (2311 - 25509)
- Min - max: 3 - 238433.

Interquartile range (from p25 to p75 in raw result table)
Data interpretation:
- 50% of observed total monthly registered accounts (just assumed the dataset is for exactly monthly data) ranges from 2311 to 25509 (the interquartile range);
- 50% of observed months have total registered accounts above 10145 (the median), whilst the rest 50% of months have figures below 10145.

Code:

. tabstat regacc, s(n mean sd p50 p25 p75 min max)

    variable |         N      mean        sd       p50       p25       p75       min       max
-------------+--------------------------------------------------------------------------------
      regacc |       112  21915.65  34794.89   10144.5    2310.5   25508.5         3    238433
----------------------------------------------------------------------------------------------

Data source was dumped by theymos there:

We're at a local minimum, but registrations jump around a lot:

Code:

+---------------------+---------------+
| 30day starting      | registrations |
+---------------------+---------------+
| 2009-11-02 00:00:00 |             3 |
| 2009-12-02 00:00:00 |             8 |
| 2010-01-01 00:00:00 |             8 |
| 2010-01-31 00:00:00 |            22 |
| 2010-03-02 00:00:00 |             5 |
| 2010-04-01 00:00:00 |            22 |
| 2010-05-01 00:00:00 |            15 |
| 2010-05-31 00:00:00 |            33 |
| 2010-06-30 00:00:00 |           325 |
| 2010-07-30 00:00:00 |           159 |
| 2010-08-29 00:00:00 |           148 |
| 2010-09-28 00:00:00 |           190 |
| 2010-10-28 00:00:00 |           204 |
| 2010-11-27 00:00:00 |           383 |
| 2010-12-27 00:00:00 |           384 |
| 2011-01-26 00:00:00 |           938 |
| 2011-02-25 00:00:00 |          2202 |
| 2011-03-27 00:00:00 |          2737 |
| 2011-04-26 00:00:00 |          4936 |
| 2011-05-26 00:00:00 |         14048 |
| 2011-06-25 00:00:00 |          6250 |
| 2011-07-25 00:00:00 |          3903 |
| 2011-08-24 00:00:00 |          2938 |
| 2011-09-23 00:00:00 |          1983 |
| 2011-10-23 00:00:00 |          1965 |
| 2011-11-22 00:00:00 |          1835 |
| 2011-12-22 00:00:00 |          1812 |
| 2012-01-21 00:00:00 |          2082 |
| 2012-02-20 00:00:00 |          1901 |
| 2012-03-21 00:00:00 |          2158 |
| 2012-04-20 00:00:00 |          1969 |
| 2012-05-20 00:00:00 |          2057 |
| 2012-06-19 00:00:00 |          2040 |
| 2012-07-19 00:00:00 |          2145 |
| 2012-08-18 00:00:00 |          2419 |
| 2012-09-17 00:00:00 |          2661 |
| 2012-10-17 00:00:00 |          2958 |
| 2012-11-16 00:00:00 |          2763 |
| 2012-12-16 00:00:00 |          2611 |
| 2013-01-15 00:00:00 |          3320 |
| 2013-02-14 00:00:00 |          4933 |
| 2013-03-16 00:00:00 |         13816 |
| 2013-04-15 00:00:00 |         18162 |
| 2013-05-15 00:00:00 |         11318 |
| 2013-06-14 00:00:00 |          8095 |
| 2013-07-14 00:00:00 |          6703 |
| 2013-08-13 00:00:00 |          5505 |
| 2013-09-12 00:00:00 |          4137 |
| 2013-10-12 00:00:00 |          5632 |
| 2013-11-11 00:00:00 |         26804 |
| 2013-12-11 00:00:00 |         28776 |
| 2014-01-10 00:00:00 |         29576 |
| 2014-02-09 00:00:00 |         29839 |
| 2014-03-11 00:00:00 |         22534 |
| 2014-04-10 00:00:00 |         18065 |
| 2014-05-10 00:00:00 |         13071 |
| 2014-06-09 00:00:00 |         12180 |
| 2014-07-09 00:00:00 |         10561 |
| 2014-08-08 00:00:00 |         10925 |
| 2014-09-07 00:00:00 |          9418 |
| 2014-10-07 00:00:00 |          8218 |
| 2014-11-06 00:00:00 |          7974 |
| 2014-12-06 00:00:00 |          8136 |
| 2015-01-05 00:00:00 |         23256 |
| 2015-02-04 00:00:00 |         36913 |
| 2015-03-06 00:00:00 |         33500 |
| 2015-04-05 00:00:00 |          8450 |
| 2015-05-05 00:00:00 |          8577 |
| 2015-06-04 00:00:00 |         10303 |
| 2015-07-04 00:00:00 |         11773 |
| 2015-08-03 00:00:00 |          9725 |
| 2015-09-02 00:00:00 |          7644 |
| 2015-10-02 00:00:00 |         68601 |
| 2015-11-01 00:00:00 |         44059 |
| 2015-12-01 00:00:00 |         39244 |
| 2015-12-31 00:00:00 |         37106 |
| 2016-01-30 00:00:00 |         42233 |
| 2016-02-29 00:00:00 |         24213 |
| 2016-03-30 00:00:00 |         19336 |
| 2016-04-29 00:00:00 |         16031 |
| 2016-05-29 00:00:00 |         19935 |
| 2016-06-28 00:00:00 |         12582 |
| 2016-07-28 00:00:00 |         10646 |
| 2016-08-27 00:00:00 |          9986 |
| 2016-09-26 00:00:00 |          9844 |
| 2016-10-26 00:00:00 |         10718 |
| 2016-11-25 00:00:00 |         10691 |
| 2016-12-25 00:00:00 |         12429 |
| 2017-01-24 00:00:00 |         13664 |
| 2017-02-23 00:00:00 |         13893 |
| 2017-03-25 00:00:00 |         15276 |
| 2017-04-24 00:00:00 |         20585 |
| 2017-05-24 00:00:00 |         31381 |
| 2017-06-23 00:00:00 |         35866 |
| 2017-07-23 00:00:00 |         40726 |
| 2017-08-22 00:00:00 |         45026 |
| 2017-09-21 00:00:00 |         70477 |
| 2017-10-21 00:00:00 |         95568 |
| 2017-11-20 00:00:00 |        144762 |
| 2017-12-20 00:00:00 |        238433 |
| 2018-01-19 00:00:00 |        136872 |
| 2018-02-18 00:00:00 |        107237 |
| 2018-03-20 00:00:00 |         87087 |
| 2018-04-19 00:00:00 |         89659 |
| 2018-05-19 00:00:00 |         83999 |
| 2018-06-18 00:00:00 |         72430 |
| 2018-07-18 00:00:00 |         61360 |
| 2018-08-17 00:00:00 |         43451 |
| 2018-09-16 00:00:00 |         43516 |
| 2018-10-16 00:00:00 |         35335 |
| 2018-11-15 00:00:00 |         31502 |
| 2018-12-15 00:00:00 |         20733 |
| 2019-01-14 00:00:00 |         20046 |
| 2019-02-13 00:00:00 |          2763 | [PARTIAL PERIOD]
+---------------------+---------------+

In both medians and means, the year of 2018 has been the highest year with its median and mean of monthly registered accounts are 66895 and 64782, respectively.

- The median means that 50% of months in 2018 have total registered accounts above 66895 or below 66895.
- The interquartile range for 2018 is 36477 to 88373. It means 50% of months in 2018 have figures in range from 36477 to 88373.
I always use median and interquartile range for descriptive statistics because they automatically exclude effects of outliers data points.

Code:

. tabstat regacc year10 year11 year12 year13 year14 year15 year16 year17 year18, ///
>         s(n mean sd p50 p25 p75 min max) c(s)

    variable |         N      mean        sd       p50       p25       p75       min       max
-------------+--------------------------------------------------------------------------------
      regacc |       112  21915.65  34794.89   10144.5    2310.5   25508.5         3    238433
      year10 |        13       146  144.3185       148        22       204         5       384
      year11 |        12  3795.583  3560.192    2469.5      1900    4419.5       938     14048
      year12 |        12  2313.667    353.35    2151.5    2048.5      2636      1901      2958
      year13 |        12  11433.42   8798.91      7399      5219     15989      3320     28776
      year14 |        12  15041.42  8092.689   11552.5      8818   20299.5      7974     29839
      year15 |        12  24992.25  19299.52   17514.5      9151   37009.5      7644     68601
      year16 |        12  16553.67  9350.784   12505.5   10668.5   19635.5      9844     42233
      year17 |        12  63804.75  67452.29     38296   17930.5   83022.5     13664    238433
      year18 |        12  64781.75  38659.33     66895   36476.5     88373      3535    136872
----------------------------------------------------------------------------------------------

[Alex_Sr]

Added statistics for February 2019

2019

Month	Registered	Total at the beginning of the month	% increase	First user of the month
February	16 202	2 538 716	0,63%	First user of Feb, 2019
March	??	2 554 918	??	First user of Mar, 2019

[DdmrDdmr]

Registered accounts is on a steep descending trend, although we need to take into consideration that many are registered and never active at all. If we look at some data I interpreted back in November 2018 (see Analysis - Bitcointalk Accounts – Date of Register vs Last Active per Rank), there is an interesting table that speaks for itself:



As we can see above, as of November 2018:

- Brand New accounts that stay as Brand New accounts are massive, and have been steadily increasing over the years

- Most of those Brand New accounts stay baron on creation, not being active (login-in) in the years after their creation. Take for example 2017: out of the 434.487 Brand New accounts, 348.196 did not log-in during 2018 (only 86.291 did).


The above just goes to show how many accounts are not active in terms of posting, and likely many of them are bot created accounts that the creator managed to create but did not get on to the next step: exploiting them with automated posts in campaigns. Current reduction monthly figures are therefore subject to what trend these type of accounts have. Nevertheless, reduction is noticeable.

Clicking on the profiles of OPs listed first user of the month, for each month, kind of hints what is going on behind the scenes: out of the 27 listed profiles, only 11 made at least 1 post, and out of the pre-2019 profiles, only 2 out of the 24 have been active during 2019.


Another complementary reading can be taken from the number of posts created per month:



The number of created post per month is dropping to similar numbers as those around the first half of 2017. There is an obvious correlation to crypto hype and value variance, and the numbers are interpretable according to one’s perspective. For a person signed-up during early 2018, things are really quiet now. For a person who signed-up in 2016 or before mid 2017, numbers seem to be training back to those days.

Note: Number of posts reading for the 25/03/2019 is a linear projection of todays reading minus the reading on the 23/02/2019. Actual value may differ and I really expect it to be lower.

[tranthidung]

Registered accounts is on a steep descending trend, although we need to take into consideration that many are registered and never active at all.

For new registered account figures, I suggest to exclude all newbies, that registered, and almost have no activity count, or even got banned or nuked.
Those sort of bot-created or farm-accounts don't represent true trend of new users in the forum.
They are extreme outliers, and should be excluded all.