Topic: Number of Posts per Rank (Read 337 times)

I gave you a merit as well. I was starting to think that all new members were here for the bounty, thanks for being interested in the forum.

You're a really nice guy. Thank you for referencing Zendex's post, another one that has been instantly bookmarked.

You see, I really thought you were a legendary member by the way you write and behave. Anyway, even though the road is long, I'm sure it won't take much time until you reach that final rank.

Thank you again for helping everybody understand what this forum is all about.

I think that you've taken a good attitude towards this forum now and are willing to move on with the propper approach. I've also seen that your posts are getting more elaborate as days go by, which is also good (I can catch on a fair amount on your Brazilian posts to get a rough idea). For that I have given you your first sMerit.

Zendex also created a good stat on Merit here: https://bitcointalksearch.org/topic/where-the-merit-pours-3093768 to get a quick glimpse of "where the Merit Pours". And although Merit should be a secondary factor on the forum, it's importance on ranking and how to approach posting does make it a hot topic to read on.

P.D. Regardless of your Merit Source assumption being a bit off as you've come to see, it got me thinking and became a driver for my new stat which I'll try to publish today at some point ...

Unfortunately, I haven't downloaded any data on age, gender nor location. Looking through profiles, I found this data to be blank extremely often, so I've given it a pass in order to speed-up data downloading by focusing on data that is complete.
From a user's perspective, it's natural not to fill in this information, since they would help to link profiles to real-life identities, which 99% at least of us do not want for multiple reasons.

I did find this poll though: https://bitcointalksearch.org/topic/poll-the-distribution-of-gender-at-bitcointalk-172707
Of course the poll base is very small (148 voters) and rather old, so it's not too conclusive, but lacking anything else, it gives us a rough idea: nearly eight times more male than female (which may be a little short).

Not easy to say. I've looked at the detailed underying data and can't see a reason (I was looking for a large concentration of averages in the [0..1) interval nearer to the 1 superior limit and in the [1..2) interval closer to the 1 inferior limit, which would mean that the means within each group would be very close to each other, but that is not so).

If you look at all the 0/1 pairs on the second graph, there are all sorts of ratios between them.
I tried to look at it from the logical point of view, but I can't see extra motivation for Full Members to post proportionally to other sMerited groups.

Bare in mind that:
- Data is biased in a sense in this second graph, since focus is on users in a sMerit Tx (nothing is said about those that post, heavily or not not, but lack an involvement in an sMerit Tx). If they were added to the data, I'm sure things would move around a fair bit. Alas, taking into account all active profiles would take ever so long, since data is not available as a raw file and must be retrieved at a very slow pace by other means.

- Considering the overall average, Legendary has the largest average of posts per day, but really don't drive-up overall Forum average since their aggregate number as a group is small in comparison to Members and Full Members.

@DdmrDdmr your post has been more than clarifying.

You see, you've answered another thread started by this guy complaining about not being merited yesterday. I had posted something about activity thresholds drawing merit sources to your posts rather than plain simple merit awarding, and you and another more experienced user promptly corrected me.

I believe I understand a little better how the system works after reading your answer from yesterday and the post that originated this thread. I see how I was wrong, even though I really didn't mean to accuse the merit sources of being biased or anything (in a nutshell - I should have phrased that response in a different way).

But the quality of this thread is extraordinary, there's much quality data to be shared throughout other users and I myself had assumptions that were quite different from all of this data. And the choice to build that 2nd graph over merited and meriters only was pretty much accurate, at least in my understanding. One could say that these actors are the ones that make the most difference to the forum in general, so their trends are highly insightful.

Well, thanks for the lesson. I'll keep up on exploring the forums, especially that Other > Beginners & Help section (which I found by checking your other post where you explain why you merited a post) that had completely escaped my attention.

Amazing, as usual.
I`d love to see some statistics about gender here in the forum. Even knowing many can lie about their gender, I believe some kind of analysis about gender, a range of age and so can be really interesting!!

Just a suggestion!!

Thanks, anyway, for another amazing work

This is like wow!
Way to go for the great insights!

It usually always goes down to a 80-20 rule and what we're seeing here is more of the same. I would put my money on this rule being applied to the real value/meaningfulness of the frequency of posts...

One thing caught my attention, tho. And I wanted to pick your brain on it, please.


Why do you think Full Member 0 and 1 (Avg Posts per Day - users with sMerit TX) have more or less the same avg?
They, along Legendary, are the only ones managing the avg.

A Newbie asked on another thread if there was a chart showing the amount of time between posts. I thought I'd give it a pass, but then came to believe it could be an interesting question, looking at it from the posts created per Rank/person’s perspective. That is perhaps a bit more ilustrating than time between posts.
I thought it would be quick to answer. Unfortunately it isn't, but since I was on it to answer the Newbie in a sense, I thought I'd see it through.

The overall forum data cannot be accessed, but I do have data from a couple of previous analysis, which allow me to get picture from two distinct subsets of users (therefore, conclusions are limited to those subsets):


1. Users created after Merit System kickoff


Results are based on following the posting activity of users that were created during the first two weeks after Merit System Kick-off, having made at least 1 post.
Column AvgPostsPerDay is calculated based on total posts of the user, divided by number of days since user was created (up until day of dataset extraction).
A 0 represents an average of between 0 and 1 posts per day, a 1 an average between 1 and 2, and so on.

Contrary to what many of us think, Newbies are not posting every single day like mad. There are many with just 1 or 2 posts since they joined and that's it. That why the average posts per day with them (0,24) or without them (1,25) is very different. Bots may play a role in this somehow.




2. Users that are involved in a Smerit Transaction (awarder or awarded)

Same calculation criteria here, but with a different set of users (some don't favour the term “user”, but if we call them “members” well end up getting confused due to rank names – or something worse).

Users here are those that either sent or received sMerit up until the 23/03/2018 (dataset is not the most recent, but I already had it from previous rank based analysis and it's quite hell to update it all just for this right now).




With this set of users, average goes up to 1,38 posts per day, but remember this is biased, as it only involves merited/meriter users.

Note: I decided not to paste the data table here as it is too long and graph shows all the data.