Message boards : Number crunching : Double counting?
Message board moderation

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
BellyNitpicker

Send message
Joined: 16 Jun 20
Posts: 8
Credit: 2,318,092
RAC: 0
Message 49280 - Posted: 26 Jan 2024, 22:30:49 UTC

I have just checked my stats on BOINCSTATS and discovered that my score for the last three years on LHC@Home has been replicated into an LHC-DEV score. It shoots me up the charts a little, but perhaps it should be removed.

https://www.boincstats.com/stats/-1/user/detail/96181147537/projectList
ID: 49280 · Report as offensive     Reply Quote
Harri Liljeroos
Avatar

Send message
Joined: 28 Sep 04
Posts: 677
Credit: 43,797,345
RAC: 14,426
Message 49281 - Posted: 27 Jan 2024, 16:49:28 UTC
Last modified: 27 Jan 2024, 16:50:09 UTC

I see the same and I am not even a member of lhc-dev. But there it is on my signature as well.
ID: 49281 · Report as offensive     Reply Quote
Dr Who Fan
Avatar

Send message
Joined: 26 Nov 16
Posts: 10
Credit: 171,395
RAC: 2
Message 49282 - Posted: 27 Jan 2024, 18:17:12 UTC - in response to Message 49281.  

I see the same and I am not even a member of lhc-dev. But there it is on my signature as well.

Same thing with me.

ID: 49282 · Report as offensive     Reply Quote
Bonny

Send message
Joined: 9 Oct 23
Posts: 4
Credit: 461,744
RAC: 2,751
Message 49291 - Posted: 28 Jan 2024, 14:34:24 UTC - in response to Message 49280.  

It shoots me up the charts a little, but perhaps it should be removed.

In my opinion, this should be corrected.
I only want to have points in my statistics for which my computers have calculated.
ID: 49291 · Report as offensive     Reply Quote
rob

Send message
Joined: 4 Mar 11
Posts: 22
Credit: 3,606,456
RAC: 1,130
Message 49296 - Posted: 29 Jan 2024, 9:13:44 UTC

A quick look susggests this a boincstats issue rather than a BOINC issue. As boincstats.com is a third party you may well get some satisfaction by asking this question on their support forum.
My guess is that they are using a wild card search to generate the stats. Going down ones list of active projects and search for "lhc*" which finds both "lhc" and "lhc-dev" and assumes they are both the same project and adds the credit for both projects, it then does a search for "lhc-dev*", finds only that project and so presents that data on its own.
ID: 49296 · Report as offensive     Reply Quote
Bonny

Send message
Joined: 9 Oct 23
Posts: 4
Credit: 461,744
RAC: 2,751
Message 49298 - Posted: 29 Jan 2024, 10:32:04 UTC - in response to Message 49296.  

A quick look susggests this a boincstats issue rather than a BOINC issue. As boincstats.com is a third party you may well get some satisfaction by asking this question on their support forum.
My guess is that they are using a wild card search to generate the stats. Going down ones list of active projects and search for "lhc*" which finds both "lhc" and "lhc-dev" and assumes they are both the same project and adds the credit for both projects, it then does a search for "lhc-dev*", finds only that project and so presents that data on its own.


This was also my first assumption, but the wrong data also appeared in https://stats.free-dc.org/.
Therefore, I would say that the problem results from an erroneously sent file from LHC.

There are about 320 accounts affected as far as I can guess. The only question is how this can be corrected.
ID: 49298 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2124
Credit: 159,926,969
RAC: 54,657
Message 49299 - Posted: 29 Jan 2024, 10:56:05 UTC - in response to Message 49298.  

The transfer is with a .xml file.
So, the faulty members, must be subtracted.
ID: 49299 · Report as offensive     Reply Quote
rob

Send message
Joined: 4 Mar 11
Posts: 22
Credit: 3,606,456
RAC: 1,130
Message 49300 - Posted: 29 Jan 2024, 13:11:36 UTC - in response to Message 49298.  

The transfer from the projects is, as has been stated, by each project exporting an XML file. Once on the third-party (boincstats etc.), they may well then transfer the data to something like and SQL database on which they run a query, probably not on a continuous basis, but daily and follow the same basic process I outlined earlier.
I don't think there are many active projects with the same first "word", so this running the query with the project names in the reverse order may help eliminate this double counting, also the use of a tag for "this record has been counted in today" may also work (indeed may be better?).
How to "cure" on existing data - this is a thorny one, as it would require going back through the database and finding when this started then recreating the data from that point, which is no small task.
ID: 49300 · Report as offensive     Reply Quote
gemini8

Send message
Joined: 4 Dec 15
Posts: 7
Credit: 1,076,791
RAC: 85
Message 49308 - Posted: 29 Jan 2024, 15:56:25 UTC - in response to Message 49298.  

A quick look susggests this a boincstats issue rather than a BOINC issue. As boincstats.com is a third party you may well get some satisfaction by asking this question on their support forum.

This has been asked, and Willy clearly stated that this issue is on the LHC side of things.
He can't collect data that no one provided beforehand, and that's where the projects come in.
- - - - - - - - - -
Greetings, Jens
ID: 49308 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2124
Credit: 159,926,969
RAC: 54,657
Message 49310 - Posted: 30 Jan 2024, 5:44:14 UTC - in response to Message 49298.  


This was also my first assumption, but the wrong data also appeared in https://stats.free-dc.org/.
Therefore, I would say that the problem results from an erroneously sent file from LHC.

There are about 320 accounts affected as far as I can guess. The only question is how this can be corrected.

https://stats.free-dc.org/ have now the correct counter again for LHCatHome-dev.
ID: 49310 · Report as offensive     Reply Quote
Bonny

Send message
Joined: 9 Oct 23
Posts: 4
Credit: 461,744
RAC: 2,751
Message 49313 - Posted: 30 Jan 2024, 13:10:21 UTC - in response to Message 49310.  
Last modified: 30 Jan 2024, 13:29:35 UTC

@maeax
It was disappeared, but it has reappeared at https://stats.free-dc.org/.
ID: 49313 · Report as offensive     Reply Quote
Dr Who Fan
Avatar

Send message
Joined: 26 Nov 16
Posts: 10
Credit: 171,395
RAC: 2
Message 49315 - Posted: 30 Jan 2024, 15:49:51 UTC

PM sent to LHC Administrator to look into the stats export mix-up.
ID: 49315 · Report as offensive     Reply Quote
Fardringle

Send message
Joined: 3 Mar 06
Posts: 2
Credit: 1,743,036
RAC: 0
Message 49370 - Posted: 3 Feb 2024, 2:34:26 UTC - in response to Message 49315.  

Free-DC stats did have the correct info for this -dev project again for a short time, but it has gone back to having the same info for both projects again. It's definitely an issue on the project side. Something wrong in the process that is generating the stats export files.
ID: 49370 · Report as offensive     Reply Quote
Bonny

Send message
Joined: 9 Oct 23
Posts: 4
Credit: 461,744
RAC: 2,751
Message 49460 - Posted: 8 Feb 2024, 13:02:59 UTC - in response to Message 49370.  

I'm just curious.
Is any action being taken on this matter?
ID: 49460 · Report as offensive     Reply Quote
Dr Who Fan
Avatar

Send message
Joined: 26 Nov 16
Posts: 10
Credit: 171,395
RAC: 2
Message 49463 - Posted: 8 Feb 2024, 15:26:17 UTC - in response to Message 49460.  

I'm just curious.
Is any action being taken on this matter?

I have sent another PM to different project admin hoping to get some action.
ID: 49463 · Report as offensive     Reply Quote
Dr Who Fan
Avatar

Send message
Joined: 26 Nov 16
Posts: 10
Credit: 171,395
RAC: 2
Message 49612 - Posted: 21 Feb 2024, 23:19:49 UTC
Last modified: 21 Feb 2024, 23:22:51 UTC

BUMP: Still waiting for any project Admin to reply or look into credit problem

Can a project manager or admin look into the cross entanglement of credit mentioned in the beginning of this thread/topic?
ID: 49612 · Report as offensive     Reply Quote
Profile Nils Høimyr
Volunteer moderator
Project administrator
Project developer
Project tester

Send message
Joined: 15 Jul 05
Posts: 242
Credit: 5,800,306
RAC: 0
Message 49652 - Posted: 26 Feb 2024, 10:45:35 UTC

As far as I can see, this is not on the project side. On our servers the DB-dump jobs are running and the contents of the html/stats directories of the 2 projects different. (Orders of magnitude more hosts and users for LHC@home than our dev project.)
ID: 49652 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2433
Credit: 227,746,920
RAC: 125,537
Message 49654 - Posted: 26 Feb 2024, 11:58:10 UTC

Downloading the user stats file from -dev works fine and it reports the correct credits (at least for my account).


boincstats reports this

As for LHC@home:
Last update user XML 2024-02-26 07:15:20 UTC (02:04:55 old)
Last update host XML 2024-02-25 07:14:35 UTC (1 day 02:05:40 old)
Last update team XML 2024-02-26 07:15:20 UTC (02:04:55 old)

As for LHCathome-dev:
Last update user XML 2024-01-26 08:51:04 UTC (31 days 01:54:01 old)
Last update host XML 2024-02-25 13:00:03 UTC (21:45:02 old)
Last update team XML 2024-01-26 08:51:04 UTC (31 days 01:54:01 old)
ID: 49654 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2124
Credit: 159,926,969
RAC: 54,657
Message 49655 - Posted: 26 Feb 2024, 12:03:28 UTC

Willy de Zutter.
He is the man for this stats.
So, you can inform him.
Maybe, LHCatHome and LHCdev have the first three digits the same.
ID: 49655 · Report as offensive     Reply Quote
Fardringle

Send message
Joined: 3 Mar 06
Posts: 2
Credit: 1,743,036
RAC: 0
Message 49659 - Posted: 26 Feb 2024, 19:04:02 UTC - in response to Message 49652.  

As far as I can see, this is not on the project side. On our servers the DB-dump jobs are running and the contents of the html/stats directories of the 2 projects different. (Orders of magnitude more hosts and users for LHC@home than our dev project.)


Multiple different stats reporting sites have the same problem, and they all have said that it is a problem with the stats files coming from the projects.

Bok at Free-DC stats has been manually downloading the stats for the LHC-Dev project so his site sometimes shows the correct numbers, but only because he's replacing the information from the regular export files from the project.
ID: 49659 · Report as offensive     Reply Quote
1 · 2 · Next

Message boards : Number crunching : Double counting?


©2024 CERN