Bitcoin Forum

From first 2'000'000 users there are 1'443'323 "Brand New" and 417'536 "Newbie" users. Together these 2 positions amounts to 1'860'859 , this is 93% of all users !! This is huge amount and this information was seen already in my last posts*, where users started to speculate of possible reasons behind this huge new account number. Therefore I decided to take a closer look to these users. There are limited statistics which can be extracted from users profiles thus limited information what we can analyse.

http://i67.tinypic.com/20z5tus.jpg


Lets take a look at the biggest group of "users"or id's - "Brand New', term "users" does not seem right since almost all of them do not contribute to the forum in anyway. There are 1'443'323 "Brand New" users. The big question, why so many? My first instinct was to think that bitcointalk.org being an old ,influential domain which is highly ranked in search engines , is heavily abused for creating backlinks. I filtered those accounts which has backlinks in them and it turns out that they are not so many, but the number is notable 1,32% (19'051 Id's) (1.Graph) of " Brand New's" have them in their profiles. While going trough data, I noticed that there are profiles which has last activity " Never" , I filtered them and it turns out that they are a lot -10,3% (1.Graph) from all " Brand New's". I am not sure why these users have this kind of " Last Activity" in their profiles , my guess is that they are "nuked" by moderators .

Other stats that can be extracted from user profiles are: post count (2.Graph) and last activity (3.Graph). Most of these accounts have not made any posts , there is a small fraction that has made 1+ posts.

If we take a look at (3.Graph) ,we can see " Brand New" account activity representation, I calculated the difference between last activity and registration date , grouped  data and plotted it in respect to days.I excluded the 10,3% with last activity " Never" . We can see that about 76 percent of " Brand News"  analysed last active where less then 5 days after registration. 



http://i66.tinypic.com/ih8ho5.jpg


I did similar analysis on " Newbie" accounts , Newbie accounts also has has some notable percentage of links in them, there are about 10 000 accounts with links in their profiles and most likely they are meant as back-links. (Graph.4) If we take a look at the post count (6.Graph), we can see that that about 32% of Newbies have made only one post and the rest have made more than 1 post. In the " Activity plot"  (5.Graph) we can see that 32,5% of all " Newbie"  accounts are last used 10 days or less after registration.
http://i67.tinypic.com/33yn9zq.jpg

In conclusion,there is 1,16 million users which have abandoned their accounts after 10 or less days. Small fraction of all Newbie and Brand New accounts have links in their accounts , in majority of them , they are with purpose to use power of this forum to increase google positions of other web-pages , there also is notable percentage of accounts which have status " Never" in Last Activity. The biggest part of inactive accounts are still not explained, some of them could be sleeping spam accounts. In my last posts* users suggested that inactive accounts are consequences of IP bans.

*last posts:
1)https://bitcointalk.org/index.php?topic=4649770.msg41998196#msg41998196 (https://bitcointalk.org/index.php?topic=4649770.msg41998196#msg41998196)
2)https://bitcointalk.org/index.php?topic=3333773.msg34863787#msg34863787 (https://bitcointalk.org/index.php?topic=3333773.msg34863787#msg34863787)

Maybe the "Never" activity means that they registered their accounts but never logged in.

Also, one of the probable reasons why there are so many Brand New accounts is that when a Newbie gets nuked, all his posts are deleted, so they go back to the Brand New rank.

It could be true that nuked accounts contrubute in this way, but nuked "newbie" or whatever rank of accounts who get "nuked' can not amount to over a million "brand news" , i do not believe that mods are so active in banning,or are they ?

I can't say. Whoever, they tend to ban A LOT of users. And there are more Newbies spamming the forum than you probably know.

Here is a post from Loyce where he lists 567 users who need a ban (I assume most are Newbies): https://bitcointalk.org/index.php?topic=4720640.msg42653191#msg42653191
Here are more ~600: https://bitcointalk.org/index.php?topic=4720640.msg42653763#msg42653763
Here are more 2751: https://bitcointalk.org/index.php?topic=4720640.msg42773062#msg42773062

And there are many many more...

Good stats OP (a kind of  relief also that some stat are not related to merit .)

Stats from your first post.

https://bitcointalk.org/index.php?topic=3333773.msg34863787#msg34863787

Quick sorted these 900K users:

661K Brand New
190K Newbie
45K Jr. member
12K Member
11K Full Member
8K Sr. Member
2K Legendary
4K Hero member

Stats from your second  post.


https://bitcointalk.org/index.php?topic=4649770.msg41998196#msg41998196
Brand New   733503   73,35%
Newbie         212864   21,28%
Jr. Member    33395     3,34%
Member        13682     1,36%
Full Member  6334       0,63%
Sr.Member     222        0,02%
Hero               0            0%
Legendary      0            0%


OP, Can you do a favor and find how many (Member and above were become active in 2018) and from your second post , can you find that how many Member and above was active after May 2018.

Just curios , how many people are active in this community or what is active size of the community.

The Brand New phenomenon is down to the Evil IP system. When an account is banned the IP address assigned some evil points and all IP addresses close to it a smaller number of evil points. When a new account is signed up from an IP address that has a certain number of evil points they are required to pay a fee before they can post (the higher the evil score the higher the fee). The spambots sign up many accounts to find one they can post from for free.

This is an interesting find. How about deleting accounts with no activity for years ?

Regarding the website links in the newbie accounts, I suppose this can easily be solved by adding a rel="no-follow" attribute in the website URL field. I just checked and currently the link is do-followed, which is like a magnet for these spammers. If this link was made to be no-followed, google wouldn't count it as relevant. Hopefully, as a result of this change, the brand new users whose sole purpose is to obtain a backlink to their website will lose interest to join and spam will be reduced, even if it is a small fraction.

Hello and thank you.

I am plannig to make a post in near future with information regarding the ranks "Member" and above, which will include also activity.


I agree, I imagine that this could be easily implemented and good solution.

Its probably not needed. I don't see any real benefit other than space, but I don't think that's too much of a problem for theymos. I guess it does skew the forum statistics a little bit, but I don't think that justifies it. I don't really see too many benefits of doing it.

Maybe, later down the line things will change, and theymos will start pruning old accounts, but I think I remember this being discussed before, and it was ruled out for a few reasons. I can't seem to find that thread though.

Thats a really interesting find. Maybe it has something to do with the hype of bitcoin when it hit $20k, people came to this forum then suddenly lost interest when it he price dropped?

I think no, looks like brand new member got evil points when registering, like what TheQuin said above. So, they abandoned brand new that got evil points and try to find another IP for register.

@Welsh
Not sure but are you referring to the below thread?

https://bitcointalk.org/index.php?topic=782493.0

I think spammers also "stockpile" accounts, and that brings me to a question to OP: can you make an overview of the time it took for all accounts to create their first post after registration? (I realize this means scraping a couple of hundred thousand pages)

They definitely do that. It makes sense that it will take a lot of time trying different IP addresses to accumulate enough useable accounts to launch a spam attack with.

Look at the " Brand New " Stats,almost all of them have 0 post, this amounts to about 1,4 million accounts! There is nothing to srape there.
In case of " Newbies" there is about 65% which have done less than 5 post, there is information to be scraped , it would only be worth it , if in the end we will see first post made in very similar time frames or very shortly after registration which I doubt.

I bet most of the newbie accounts are from that period only.

Such a crash is very bad for crypto in general. It will not only result in huge loss for the new investors but some may get done with crypto with this first bitter experience. I mean loss of 10-20% is still fine but someone loosing half of his money or more in few days can have a very large impact. Probably many such investors never looked back at crypto or this forum.

A lot of people only come here because they have a single question to ask. It might be a general one about a wallet or way to exchange in a specific currency or they may have an issue with a service that has a support thread here.

Many thanks for publishing the very interesting topic.
I give you link to the one : Using R for Time series Analysis (http://a-little-book-of-r-for-time-series.readthedocs.io/en/latest/src/timeseries.html).
Personally, it might be better if you use time series plots for your data presentation.
Of course, you can plot time series datasets with other statistical softwares, not only R.

I do not believe that its true , look at the Newbie Registration rate, a lot of users registred their accounts before big Bull run when they could potentially lose money to such extent as you mentioned
http://i67.tinypic.com/dn1pv7.png


I believe that user TheQuin is right,  most of new users are here to ask few questions about blockchain .

Can you make a bit deeper analysis, with delayed effects of merit system and new rank requirements on newbies' activities in the forum over the last 6-7 months since the launch day of merit system in January?

Yes, I am familair with time series but in this case I believe bar plots seems more appropriate.
I am using OriginLab there ar eplenty of options. But thanks for suggestion.


As I understand you would like to see the potenital impact of merit system to forum using scraped data.
The problem is that there is other strong factors among implementation of merit sytem, as the change of whole interest in crypto space because of the market bubble , this makes me think that we wont see much. Also i speculate that data in user profile is not enough in this case , if i scraped the posts and analysed posting times , it could give us some interesting information.

I believe that there are already users who have done extensive research on merit sytem impact,
Take a look at this user and his posts, they are very interesting also.
https://bitcointalk.org/index.php?action=profile;u=1582324

wow, this is interesting.
I did know that there is a lot of Newbies and Juniors but I never expected to see so big numbers. This explains some things like the spammers in mega threads. Thanks for these statistics.