Long time for this kind of update, keep in mind we can assune right now 80% of merit is given from merit source and 80% of merit source has the rank legendary, so the majority of the merit from the higher ranks is from sources!
For a perfect graph we need to remove sources.
I've removed data for admin, founder, mods, Vip and brand new since they have less than 0.0xxx % circulation of total merit.
Datas are from DdmrDdmr datas.


MERIT FROM -> TO

Just a few comments to help contextualize the data and conclusions that may be derived:

- Rank is derived from the user profile, and is in a sense a snapshot of the user’s current profile. The ideal would be to have the implied ranks at the moment of the merit being awarded. Since we do not have that detail, what is used is the next best thing.
Nevertheless, this does imply, for example, that all sMerit to and from @Jet Cash are attributed to the Legendary rank, which is his current rank. In truth, he ranked-up a couple of weeks ago, so to be precise, most of his merit TXs should be really attributed to a Hero rank.

- The graphs cover all sMerit TXs since the Merit System start. Since there was more merit at the beginning, the initial TXs wheigh more than recent TXs.

- Both the above can be mitigated largely by delimiting the data to, let’s say, last running month or so (this can be done on the Merit Dashboard by narrowing the Date Range using the slider). This would be representative of what goes on "now" in the Merit System, kill initial TX wheight, and mitigate the change of rank effect mentioned above.
If the graph were to be redone with this criteria, the percentages would move up to 10 points in some cases, although I believe @d5000’s observations apply just the same.

I agree with that, maybe I wasn't clear enough. I also give a substantial part of my merits to other Legendaries. But maybe there's an effect that we Legendaries unconsciously value other Legendaries' posts higher than those that come from low-ranked members, even if they were of similar quality.

It's only an invitation to reflect on the practice. Not to propose to merit low-quality posts.

After months and months, do you think anything is changed? Can you elaborate now a new set of datas without the sMerit spend from knowed sources to get a more truthful graph in the end?
Do you think if we only remove the knowed source merit (instead of removing them all since we do not know them all) do will have a similar graph?

@kvanko I will try to change on next update, ty for feedback!

The merits transactions are public: https://bitcointalk.org/merit.txt.xz

Then you can make your own script in any language(i don't know if there are ready tool, but i assume so) that grab the data from the forum profiles to get the rank or any other info you want.

Yes, data distribution does change overtime, although not significantly for all ranks nor combinations. For example, if you delimit the dates to last running month (25/07/2018..24/08/2018) and compare the data to the full history aggregate we can see that:

-   Legendary (From Rank) barely has a variation of over 2 points in the sMerit awarded to any rank (in an other context 2 points is massive-  i.e. sales, but not here).
-   Hero on the other hand goes from awarding 19,81% of its awarded sMerit to the Member rank (Overall historical view), to 11,69% (one month view). That’s over an eight point decrease there. Hero’s assignment to Jr. rank goes from 6,13% to 10,11%, nearly four points increase more.

The above shows that distribution shifts overtime, partly due to causes I pointed out before, and that is why I would favour a view of last running month as a more accurate view (more aligned to day to day nowadays awarding habits).


Recently I created a topic called sMerit Senders & Receivers – Weight of top 100 and 200 weekly contributors. There we saw that the weight of the top 100 sMerit senders is massive (between 60% and 65% of monthly awarded sMerit as of late). Merit sources are likely within that group for sure, so taking them off the equation only covers 35%.40% of awarded sMerit by non-Merit Sources. The weight of Merits Sources is unfortunately such, that not considering them in the from/to rank sMerit awarding scheme would be a massive reduction from reality in my opinion.
On the other hand, If there were to exists a 100% non-equivocal list of Merit Sources, we could consider creating a flag to allow them to be filtered in/out (although you cound not distinguish between personal sMerit and source sMerit being awarded). Nevertheless, I can guess from the top 100/200 some of them, but not all of them (I often appear on the list too, not being a Merit Source).

I don't think ban for new users is a good idea since statistically in 10,000 accounts created someone will create an account only because he wants to actively participate in the forum.
The problem as usual is the bounty section, if you limit the bounty section to member + or jr with merit, a lot of spam will be removed as abusers will be no more able to join with 100+ accounts the same campaing like this one, he made 560 accounts only for abusing telegram airdop campaing https://bt.irlbtc.com/view/4759763.msg42989933#msg42989933

Yeah, the merit system does nothing for bots or spam. It just stops them from rising through the ranks without getting merit. We should require some merit requirement to become Junior though because plenty of bots and spammers are still farming and spamming away on them which only requires 30 activity and anyone can get that by posting over a few periods.


This is never going to happen. We shouldn't be turning people away who actually want to come here for the forum's intended purpose to discuss or get help with things about bitcoin. We should look at tackling the issue with sig spam and bots in other ways though and with penalising or annoying regular users as little possible.

You can use cheat codes. Anywhere within the forum type:
IDDQD to become immune from ban and remove limitations
IDKFA to fill up merit and activity