Mismatch between API and chess.com club member count (SOLVED)

I created a String Array with all the users under the "all_time" JSON array. Chess.com says this club has 991 members. The string array only has 753 usernames. Just to doublecheck I manually copied everything after "all_time" and pasted it here:

https://wordcounttools.com/

It also shows the word count for "username" is 753. 

Who is wrong? 

Try adding all_time, monthly and weekly together. If the club keeps its list of closed accounts the numbers will be slightly low compared to the website, which I believe counts closed members [Edit: no, see post #3] we can't see via api.chess.com (because on the website, team admins are the only ones who can see them, and via api.chess.com we are not logged in).

Example: Chess.com Developer Community

Website: 991

$ ./club-numbers chess-com-developer-community

chess-com-developer-community

weekly: 130

monthly: 108

all_time: 753

total: 991

Looks OK to me. No bug.

Banned members may count too as well as closed.  [Update: No, wrong again. Read on ...]

Asger's Great Viking Warriors, website: 1,054

$ ./club-numbers asger-s-great-viking-warriors

asger-s-great-viking-warriors

weekly: 154

monthly: 89

all_time: 811

total: 1054

... so it looks I was wrong about closed and banned members. Asger's GVW has 36 closed members (all reasons) and 15 banned members (I'm an admin, so I can see) and they are not being counted for the website or for api.chess.com per the numbers above.

When banned members of a club do and do not get treated as members is historically confusing and there is at least one outstanding bug report not related to api.chess.com.

But hey, regular members are counted Just Right.

Thanks. I thought all time meant well... all time lol. Did not know players could only be on one section at a time. I edited the title so no one freaks out. 

The divisions I don't find useful. A fair question to ask, and I learned some minutiae. Good idea changing the title.

I fixed it and now I get the following:
Week members: 130
Month members: 108
All members: 753

Just as you did. Now I just have to wait like 5 minutes for it to calculate the average club´s rating. By any chance have you calculated that before? I want to compare my result  to something else to verify it. 

I guess the divisions can be useful. A club may have 1000 members but only 200 that are active. We have 991 members but very few post. I wonder if the other 753 members read our posts.

 After 5 minutes this is what my program came up with:

Week average 842.4692307692308
Month average 724.5462962962963
All average 665.2045152722444
Averages together 2232.2200423377717
Averages together divided by 3 744.0733474459239
744.0733474459239

Seems like we suck at bullet chess. We have an average rating of 744 lol.

The trouble I find with the divisions is that not enough is counted as activity. Things known not to count:

• Making a move in a team match game
• Voting in a team match game.

I've other sources to tell me that in some groups people aren't reading notes or new match notifications!

I've not done average club rating at any time control no. Doesn't sound hard, but I'm not sure I want to do it either unless you find a discrepancy with the website. If you do, drop me a PM or start a new thread maybe? I'm about to become busy.  After I make coffee! ☕️☕️

Makes sense. That activity should count as you say. 

Can you share how to know if someone did not read a note? 

When they thank me for sending them a PM asking them to join a match after I've been nagging for more people to join for a week via news and notes. Strictly, it's a guess, but it's a pretty solid guess for a few people!

LOL. I thought it involved some weird chess.com bug in the page code

Does the website show average club rating? I dont see it anywhere. Clicking on "Player leaderboard" shows only players from the club but the average is still global since it says the number of bullet players is about one million.

The only place I know, and it doesn't say but I bet it isn't 960 and may be cached:

Oh, I see. Can you tell me what club is that? Im gonna try to put my Java library to the test and calculate the average daily rating to see if I get the same. 

Asger's Great Viking Warriors, asger-s-great-viking-warriors

Here we go. 1054 members, will take more than 5 minutes to load all that JSON lol!

I'm yet to break down and instrument anything, but I believe it's latency at api.chess.com that is the limitation I see, once HTTP/1.1 connection caching is turned on. SSL negotiation takes time. I suppose I should instrument something and stop guessing, as programmers' guesses are usually wrong. One script I have is definitely slow and my first attempt to improve it did nothing.

On the bright side I will get to polish my loading bar skills on Java GUIs