buggnubg - Bugs: bug #39807, Two differents analyses to the...


bug #39807: Two differents analyses to the same position

Submitted by:  None
Submitted on:  Sun Aug 18 17:45:59 2013  
Originator Name: WILSONOriginator Email: -unavailable-
Item Group: Other bugCategory: Evaluations
Release: 1.01-ReleasePrivacy: Public
Operating System: Windows 7OS Type: 64Bit
Processor Type: AnySeverity: 3 - Normal
Open/Closed: ClosedAssigned to: Michael Petch <mdpetch>
Priority: 5 - NormalStatus: Invalid

Add a New Comment (Rich MarkupRich Markup):

You are not logged in

Please log in, so followups can be emailed to you.


Fri Mar 7 18:01:21 2014, comment #5:

This issue is because the neural network changed and previous analysis can be different with 1.00+ vs the older 0.90 Neural Network based releases.

Michael Petch <mdpetch>
Project AdministratorIn charge of this item.
Fri Aug 30 00:40:38 2013, comment #4:

The first time you stored it (weeks ago) what version did you use?
Answer: About May. Version gnubg-MAIN-20121023.

When you redid it recently what version did you use?
Answer: gnubg-release-1_02_000-20130728.

Petch, you can reinstal the old version and analise that position. After you perform the analyse with the new version.

Note: I am with difficulties to reset my account. I lost my password. My email is -unavailable-
I already asked but the system do not send it to me.

Mon Aug 19 04:39:30 2013, comment #3:

I need a bit of clarification because I may not understand. You mention two versions.

The first time you stored it (weeks ago) what version did you use?
When you redid it recently what version did you use?

The neural net was upgraded between 2012 and v1.0x+ . The new network is considerably stronger so I need to know exactly what was analysed with in both cases.


Michael Petch <mdpetch>
Project AdministratorIn charge of this item.
Mon Aug 19 04:24:45 2013, comment #2:


I use just one version, but the prior analyse was recentily stored and may be performed by gnubg-MAIN-20121023.
The last análise belongs to gnubg-release-1_02_000-20130728.
But what amazes me is the fact that the latest version display an inconsistent analyse result.

Best Regards.

Sun Aug 18 17:50:08 2013, comment #1:

Were you using different versions of GNUBG?

Michael Petch <mdpetch>
Project AdministratorIn charge of this item.
Sun Aug 18 17:45:59 2013, original submission:

I performed one analyse and stored it. The best move was B/21(2), 23/15. After two or three weeks I reanalyzed the same position again, and the results has changed. The movement B/21(2), 23/15 is hight now the 13º. See the attachments and the position. In the position, click on green line to see the previous analyse.
Note: In both analyses I analised all moves (35) at 4 ply.

Best Regards,




(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach File(s):

Attached Files
file #28870:  006- hazy - Kills 66 - 55- 44 e 33.sgf added by None (4kB - application/octet-stream)
file #28871:  Analyse 2.xlsx added by None (93kB - application/vnd.openxmlformats-officedocument.spreadsheetml.sheet)


Depends on the following items: None found

Items that depend on this one: None found


Carbon-Copy List
  • -unavailable- added by mdpetch (Posted a comment)
  • -unavailable- added by None (Submitted the item)

    Do you think this task is very important?
    If so, you can click here to add your encouragement to it.
    This task has 0 encouragements so far.

    Only logged-in users can vote.


    Please enter the title of George Orwell's famous dystopian book (it's a date):



    Follow 6 latest changes.

    Date Changed By Updated Field Previous Value => Replaced By
    Fri Mar 7 18:01:21 2014mdpetchProcessor TypePowerPC=>Any
      Assigned toNone=>mdpetch
    Sun Aug 18 17:45:59 2013NoneAttached File-=>Added 006- hazy - Kills 66 - 55- 44 e 33.sgf, #28870
      Attached File-=>Added Analyse 2.xlsx, #28871

    Back to the top

    Powered by Savane 3.1-cleanup1