Topic: [BPIP] Bitcointalk Public Information Project [Back in Action] - page 34. (Read 19631 times)

A more accurate gauge of merits earned would be to utilise the methodology of how many unique times merit is earned, not totals received (regardless of starting amount).

A person who receives fifty unique merits of one merit each has earned far more than the next person who just once receives fifty merits.

Let that be your starting point.

Yes and yes. I'm still working on some edge cases, including this one and also some scenarios where multiple events happen around the same time (e.g. if someone had included johhnyUA when you excluded him). It looks like I should switch from set operations (SQL) to processing each event sequentially (loop). There is also a delay between when Bitcointalk updates dtview and full;dt lists, which is very annoying.

I saw an opportunity for a quick test and I took it.


Shouldn't it say "johhnyUA DT1 has gained the ability to choose DT2 members" when he went from -1 to 0?

Also, shouldn't distrust come before loss?

There had been confusion with the ranks/numbers shown under "Most Merited" on the front page earlier in the thread. I explained that the tooltip had indicated the Most Merited report had been including initial merit, as it has, and that those numbers/ranks are calculated using initial merit + earned merit to reflect & rank the merit numbers that are shown on a users profile on bitcointalk.

However, the "M" above it, which was linked to the r=mostmerited report did not match the information being displayed on the front page. So I corrected this report to match the front page ranks/numbers, and then moved the prior report to r=earnedmerit, so it would clearly indicate it is showing earned merit. I've also recently noticed that the "M"(merit) area under the MATR section on the profile page is also using the incorrect rank, but displaying the correct value. This rank will be fixed as soon as I get time today/tomorrow.

Recognition is based off of the r=earnedmerit rank/calculation, which does not include initial merit. The earned merit ranking may not be shown on the front page, but it makes sense to base the Most Recognized rank off of it because a Most Recognized profile would likely be one that has more earned merit.

As soon as I can, I'm also going to be setting up little "i" icons next to each of the main page ranks to provide more info on the calculations, and will provide a link to the earned merit report under the tooltip for Most Recognized.

Thank you LoyceV, it contains useful information but we're looking into some privacy aspects of this before we commit to anything.

Thank you Timelord2067, there is an issue on the profile page also related to this, and potentially some confusion which I may have caused when trying to clear up the confusion that Quickseller's (likely) alt had previously , I'm looking into this and will have an update done soon to hopefully clear up any confusion!

I would appreciate it if you could acknowledge receipt of my report of a glitch in BPIP as reported a couple of days ago:

I responded in the extension thread, sorry about the delays, we've been discussing 2 or 3 other things with ibminer over the last few days.

His request is interesting tho, and you're not answering it


He is not talking anywhere about adding this feature to the forum.

No need to double post, both threads are active and monitored by the BPIP team
I would love to see that implemented in the forum instead made trough third party setup,  ut it won't happen.

I think the BPIP it's more like a better sorting and user friendly GUI of the already publicly available information, i do not think they will implement any "private" features like your suggestion, but that's up to the BPIP team.

I posted this in other topic, but I am doing it here also, as I am not sure if other topic is active.


One suggestion/request/question:

Is it possible to add something like personal notes feedback for members in BPIP extension?

(For example, if I want to have private feedback note for specific member, I can write short text that only I can see.)


PS
My feedback so far using new extension:
It is working great for me, and I had no bugs to report.
Only thing I noticed is slower loading information time for some members

Will think about it. I prefer a quick visual clue of who's in or out (red "x") so perhaps we'll come up with some sort of combination.


That's a good point. I'm thinking to not show users who don't have at least one DT2 inclusion.

I have another suggestion, see this:


My suggestion is to not show DT2-exclusions for users who were already not on DT2. JollyGood (nothing against him, but he's in my example) has excluded 1517 users, and considering most of them were never anywhere near DT2, their exclusion doesn't add anything to the list.

Agreed - that would be a great re-addition.

---

Don't worry, I archived it... 

---

---

How is the "most trusted" feature calculated?

My score dropped from 229 to 277 when just one of my positive trust feedbacks went from being DT1/2 to DT/??3/4??  (it's still there, just not a DT1/2 any-more). Have just noticed Hhampuz has drastically modified his trust list in the last couple of days, but still the question remains.

---

---

Another glitch in your system: Vod, LoyceV, PrimeNumber7, suchmoon, ibminer, theymos and not forgetting iasenko

https://bpip.org/report.aspx?r=mostmerited - 117 (not 287) Archive [1a], [1b]

---

(117+152+277) = 546

546/3 = 182

1000 - 182 = 818 which would be 30th place - which is where I was until a week or thereabouts.

---

if we used the 227 value the end result would be 834 or ~24th place on most recognised.

---

There's definitely something not right with the calculations. (not on my end mind)

I love love love how the new DT inclusion log shows the strength after the action - keeping a log to show what actions precipitated the inclusion.

Could you please put back the actual number instead of having to hover for it?

Would you be interested to add a link to my archived unedited posts for each user? If so, I can make a list of included userIDs. I currently only have data for the past 7 months, but it's a start.

Examples:
ibminer: http://loyce.club/archive/members/8/84866.html
suchmoon: http://loyce.club/archive/members/23/234771.html
Anara22 (a Nuked user): http://loyce.club/archive/members/276/2762486.html

I have a terrible user interface (read: no user interface ), BPIP can make accessing this data a lot easier.

I reported it a few hours ago... but it is still here.    Mods be slackin'   

Of all posts to copy you chose THE SPAMBUSTER GANG LEADER's post? LOL

Bye now.

Thank you for your reply,


Just a nit-picking bit of clarity, should that be

152, 229 then 285??

---

I think that was a part of a discussion I was reading with Vod trying to figure out formula.  I never ended up finding a similar formulae that was using square roots of values that'd been squared and added (multiplied) together.  It's a convoluted piece of formula and mostly works (never re-found it and put looking for it to one side)

The score is calculated differently, referring to this post from 2018 back then the recognition score was calculated based on this formula.

Today we have just a little change - the actual result of the above equation is extracted from 1000.

So you have the following formula:
Recognition = 1000 - ((Activity Rank + Trust Rank + Merit Rank) / 3).

In your case:  1000 - ((285 + 152 + 229) / 3) = 778

You can check wich of your scores are droping or who is above you.

What is the methodology being to used to calculate recognition?  Why I ask I was at 30th place with a score of 791.667 and then around the time my post count started dropping without explanation (as discussed in the original BPIP thread) now my recognition score is a dozen points lower while the other users with scores around mine are all fairly static.

At the time it was suggested that the posts were very old in threads that had been nuked, now I'm not so sure as the evidence suggests manipulation by mods to push my BPIP score down with very targeted deletion of posts.