Which boards are truly "under-merited" vs. "over-merited"?

[nutildah]

Q: What are you doing on a peaceful Sunday, God's holiest of days, Nutildah?

A: "Why I'm analyzing merit distribution across the forum, of course!"

In this topic, I described two key metrics that together account for roughly 16% of an account's Post Health Score (I will update these scores in the next few days BTW):

• Semantic Repetitiveness - a measure of an account's repetition of the same ideas or concepts using different words
• Unique Word Ratio - a measure of how often an account is likely to repeat the same words, which is also indicative of vocabulary size

After generating semantic repetitiveness and unique word ratio scores for over 3000 forum members, it became apparent that users with high (good) scores for these metrics tend to receive more merits than users with low scores for these metrics. That is to say, members who post about several different ideas and concepts while using a wide variety of words are generally more appreciated by the community than those who do not.

On a base level, it makes sense that such posters would earn more merits, as we are more likely to be mentally stimulated by new information and fresh takes on subjects than old information and oft-repeated takes.

Using these two metrics as a reference point, I wanted to know if certain boards were indeed "under-merited" or "over-merited" relative to the average semantic strength (an average of semantic repetitiveness + unique word ratio) of members currently posting on each board. To do this, I calculated averages for 16 of the forum's most popular boards plus the most active posters in the Wall Observer thread. Averages were generated by collecting user IDs of the 60-120 most recent posters on each board (40 members for the WO thread) and calculating the semantic strength for each user ID. I also used BitList to find the average merits per post of each board across the last year.

The results were as follows, sorted by merits per post:

Board/Thread____________	Merit/Post_	Sem. Str._	Sem. Rep._	UWR__	M/P to SS ratio
Wall Observer thread	2.35	93.7	93.7	93.7	0.0251
Dev. & Tech. Discussion	1.84	81.8	85.8	77.7	0.0214
Bitcoin Technical Support_	1.25	82.0	85.3	78.7	0.0147
Meta	1.23	76.8	80.0	73.7	0.0154
Reputation	0.91	79.0	83.0	74.9	0.0110
Project Development	0.71	84.8	84.0	85.6	0.0084
Mining	0.50	75.9	78.6	73.2	0.0064
Speculation	0.46	65.3	67.8	62.9	0.0068
Hardware	0.40	84.8	85.2	84.4	0.0047
Beginners & Help	0.28	65.1	64.6	65.5	0.0043
Bitcoin Discussion	0.21	61.0	63.8	58.3	0.0033
Trading Discussion	0.09	55.7	58.6	52.9	0.0015
Speculation (Altcoins)	0.09	67.9	67.4	68.4	0.0013
Economics	0.08	64.5	65.6	63.3	0.0012
Gambling	0.07	68.5	70.4	66.6	0.0010
Altcoin Discussion	0.07	68.4	70.2	66.6	0.0010
Gambling Discussion	0.01	64.9	65.1	64.8	0.0002
___
Average	0.62	72.9	74.7	71.2	0.0075
St. Dev.	0.69	10.3	10.3	10.5	0.0076

The data suggests a fairly strong correlation (Pearson r ≈ 0.79) between  merits/post and semantic strength, although some boards are clearly over- or under-merited compared to the average (M/P to SS ratio = 0.0075). To better visualize this phenomenon, here is a scatter plot of merits/post and semantic strength:



Boards above the line are potentially "over-merited" while boards below the line are potentially "under-merited."

Replacing a linear trend line with a quadratic trend line is a more appropriate fit for our data since it better captures the acceleration in merits per post as it corresponds to increasing semantic strength:



So... what boards are truly the most over-merited vs. under merited? We can analyze this by focusing on each board's residual, or the difference between its actual data point and where it should be on the trend line.

Most Over-Merited Boards

Rank__	Board	Residual___	Interpretation
1	Dev. & Tech. Discussion___	+0.82	Significantly over-merited
2	Meta	+0.55	Clearly over-merited
3	Speculation	+0.27	Moderately over-merited
4	Wall Observer thread	+0.22	Moderately over-merited
5	Bitcoin Technical Support	+0.22	Moderately over-merited

Most Under-Merited Boards

Rank__	Board	Residual___	Interpretation
1	Hardware	-0.86	Significantly under-merited
2	Project Development	-0.55	Clearly under-merited
3	Gambling	-0.21	Moderately under-merited
4	Altcoin Discussion	-0.21	Moderately under-merited
5	Speculation (Altcoins)_____	-0.17	Moderately under-merited

Here are the residual values for the other 7 boards:

Bitcoin Discussion: +0.11
Beginners & Help: +0.10
Reputation: +0.09
Trading Discussion: +0.01
Economics: -0.09
Mining: -0.13
Gambling Discussion: -0.16


In conclusion, some boards definitely receive less or more merits than expected.

Although Gambling Discussion does receive slightly less merits than it should, its nowhere near as unfair as the situation in Hardware or Project Development. The Gambling board is definitely under-merited. So, I ask that you hold off meriting this post and instead give some merits to your favorite gambling section posters. They'll need them more than I do!

[Mahiyammahi]

This was an amazing research you have done. I would like to watch similar research for local boards if you can. Cause I think the local section Bangla (বাংলা)  is getting neglacting same like gambling board.

[popv9]

Do you have a minimum value for the semantic strength of all boards? Even the most vapid posts should mark some floor which provides greater context to the current range. Certain boards may seem undermerited, but could you really believe that altcoin threads, or promotional gambling updates would contain that many useful posts? If you were to compare it to boards like Off-topic, or even better, Lending, where many posts will follow the same structure, I would wonder whether the strength of the correlation truly maps to reality.

[Karl_3000]

Although Gambling Discussion does receive slightly less merits than it should, its nowhere near as unfair as the situation in Hardware or Project Development. The Gambling board is definitely under-merited. So, I ask that you hold off meriting this post and instead give some merits to your favorite gambling section posters. They'll need them more than I do!

That is how just life is. I see it as something that can happen in life as people can concentrate somewhere than other places and all will begin naturally. I have a suggestion, theymos can give merit source to someone that is more merit neutral who is focus on meriting the gambling board posters. This is how this can change.

[Mia Chloe]

~snip

Most boards already have this data and I think especially local boards do. The fact that the wall observer board looks this good in stats actually points to a more different stat actually. The reason for the low semantic repetitiveness is how random discussions there are, and can be. You literally could have a dude talk about his trip during vacation.

On the other hand what is there to discuss on the wallet board aside wallets? As such you should logically expect repetitive grammatical expression and semantics it doesn't mean the board is a spam...

[nutildah]

I would like to watch similar research for local boards if you can. Cause I think the local section Bangla (বাংলা)  is getting neglacting same like gambling board.

This wouldn't make a whole lot of sense because I can only analyze English-language posts. What it would uncover is which local boards write better English-language posts than other boards, on average.

Do you have a minimum value for the semantic strength of all boards? Even the most vapid posts should mark some floor which provides greater context to the current range.

I'm not quite sure what you mean. The board with the lowest semantic strength posters is Trading Discussion, followed by Bitcoin Discussion (oh, the irony). Individual account scores for these metrics go much lower (20s-30s), but are never 0.

Certain boards may seem undermerited, but could you really believe that altcoin threads, or promotional gambling updates would contain that many useful posts?

But remember the analysis is in the context of semantic strength, in which case, they are earning about as much merits as they should be (slightly less in the case of gambling section posts).

If you were to compare it to boards like Off-topic, or even better, Lending, where many posts will follow the same structure, I would wonder whether the strength of the correlation truly maps to reality.

These two sections don't really have a personality or dedicated posters (aside from shasan and other lenders). Accounts posting in these boards shouldn't be defined by their posts in these boards, IMO.

[Karl_3000]

On the other hand what is there to discuss on the wallet board aside wallets? As such you should logically expect repetitive grammatical expression and semantics it doesn't mean the board is a spam...

If you go to the wallet side and be reading what people are posting there, other threads there are one of the most threads that posts are not repeating posts. You should differentiate between repetitive posts on those long threads and compared it with long thread on the wallet threads, most posts there are not repetitions.

You can say some people can ask similar questions but which can be very important because what the person might be dealing with is money which is bitcoin.  Anyone that gives solution are not repeating what they have already posted, they are giving solutions.

Similar threads like that on gambling boards that are new topics are not counted as spam if you post something good too, but some people like to discuss on a thread that supposed to have died and buried under many threads above it.

[popv9]

Do you have a minimum value for the semantic strength of all boards? Even the most vapid posts should mark some floor which provides greater context to the current range.

I'm not quite sure what you mean. The board with the lowest semantic strength posters is Trading Discussion, followed by Bitcoin Discussion (oh, the irony). Individual account scores for these metrics go much lower (20s-30s), but are never 0.

What I meant is that I would be curious about the lowest possible semantic strength across each board. Right now, Trading Discussion at 55.7 seems low compared to the Wall Observer Thread's 93.7, but how does the trend continue to the left?

I am also interested in the result of applying the analysis to the semantic strength of just the posts for each board instead of selecting them from recent posters: the current sampling method may skew results.

[Decimetre]

This is a very good research that also buttresses your unbiased stand as regards merit distribution. This will also go a long way to improve posts in the gambling sections too and reduce spamming because gambling section posters will not only post for campaign quotas but work harder to give merit worthy posts.

A fair judge you are!

[Supreme Donvic]

Although Gambling Discussion does receive slightly less merits than it should, its nowhere near as unfair as the situation in Hardware or Project Development. The Gambling board is definitely under-merited. So, I ask that you hold off meriting this post and instead give some merits to your favorite gambling section posters. They'll need them more than I do!

That is how just life is. I see it as something that can happen in life as people can concentrate somewhere than other places and all will begin naturally. I have a suggestion, theymos can give merit source to someone that is more merit neutral who is focus on meriting the gambling board posters. This is how this can change.

People in this forum criticize a lot, if a merit source starts giving out more merit on the gambling board, they are going to start criticizing him or her, and they will say he or she is giving more merit to a shit gambling board posts.
We have a lot of boards that lacks merit in this forum is not just gambling board, I think theymos should look at those who applied for merit source under boards that are lacking merit and make them merit source and instruct them to focus more on the boards they applied under, that way merit will be allocated equally across all boards.

[Karl_3000]

People in this forum criticize a lot, if a merit source starts giving out more merit on the gambling board, they are going to start criticizing him or her, and they will say he or she is giving more merit to a shit gambling board posts.
We have a lot of boards that lacks merit in this forum is not just gambling board, I think theymos should look at those who applied for merit source under boards that are lacking merit and make them merit source and instruct them to focus more on the boards they applied under, that way merit will be allocated equally across all boards.

It depends on the kind of post the person is sending the merits to, there are some good posters on the gambling board and they have good posts there which deserve merits. If merit source is sending merit to trash posts, I think he should be questioned in the public despite that theymos said they should complain to him instead.

I have seen merit source sent merits to some posts on the gambling board that is not really worth the merits but the merits were sent because they apply on their merit giveaway thread or because they are about to rank up to the next rank. If the person has contributed good on the forum, you will also see no people complaining about some unworthy merits. There are some posts like this also on the local board.

[libert19]

...So, I ask that you hold off meriting this post and instead give some merits to your favorite gambling section posters. They'll need them more than I do!

Ayy, I am impressed that this post of yours hasn't earned a single merit as of yet. May be 'they' took your statement seriously. Also, tells you (as in not, 'you' particularly but generic you), users here do read post content instead of blindly sending merits looking at the username as I have seen being said before.

[Cointikka]

Ayy, I am impressed that this post of yours hasn't earned a single merit as of yet. May be 'they' took your statement seriously. Also, tells you (as in not, 'you' particularly but generic you), users here do read post content instead of blindly sending merits looking at the username as I have seen being said before.

This guy is known to spam but dear @nutildah is not interested in sig ban, why? I have never seen a big spammer like him. No, one likes him yet he comes up with such a replies which does not make any sense. Please do a check and being one of the busiest guys for something, you should check his posting habits, merit given, boards posting habits, reply analysis, use of AI, original posting habits, and should I say more. I decoded your ways will decode some of your ways to come up with a sig ban. I will love to see how his posting shows up and is related to someone else posting habit.

[libert19]

...

Oh hello there. I know when I am posting for signature quota, and my above comment ain't one of it; it was based on genuine observation and I would have posted same even if I wasn't wearing a signature or was sigbanned, but in that case, you wouldn't be having problem with it.

I will love to see how his posting shows up and is related to someone else posting habit.

I am also curious tbh.

Also, you can create a thread about me in Reputation and we can continue there, instead of going off-topic here.