buggnubg - Bugs: bug #17484, Possible problem in rollout stats...

 
 

bug #17484: Possible problem in rollout stats (gui version)

Submitted by:  Massimiliano Maini <maxmaini>
Submitted on:  Mon 21 Aug 2006 07:54:48 AM UTC  
 
Originator Name: Max MainiItem Group: Other bug
Category: RolloutsRelease: 0.14.3-devel
Privacy: PublicOperating System: Windows XP
OS Type: NoneProcessor Type: None
Severity: 3 - NormalOpen/Closed: Closed
Assigned to: NonePriority: 5 - Normal
Status: Fixed

Add a New Comment(Rich Markup)
   

You are not logged in

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

 

Sat 02 Dec 2006 10:07:18 PM UTC, comment #1:

Hi Max,

I hope I have fixed this. Closing.

Christian Anthon <c_anthon>
Project Administrator
Mon 21 Aug 2006 07:54:48 AM UTC, original submission:

GNU Backgammon Position ID: 1wwAAH4HAAAAAA
Match ID : UgngACAAAAAA
+13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
| | | O O O O | OO 2 points
| | | O O O | OO
| | | O | O
| | | | O
| | | | O
v| |BAR| | 7 point match
| | | 6 | X
| | | X | X
| | | X X | X
| | | X X | X On roll
| | | X X | XX 0 points
+12-11-10--9--8--7-------6--5--4--3--2--1-+ X: MaX (Cube: 4)

Hint says redouble take (at any ply) and I do a rollout from the hint window with the following settings (as they appear in the rollout settings window) :

648 trials, no truncation, no separate evalation, never stop (sd or jsd), cubefull, truncate at exact bearoff database, variance reduction, quasi-random dice, cube decision same as chequer play, same settings for both players, player0 settings for truncation point. First play both : world class.

[Notice that settings are irrelevant to my question but allow you to reproduce the thing exactly]

After a few seconds I get the result (redouble take, as expected) and I look at the stats : here I found somethiong I can't understand.

In the panel 'Player gnubg owns 8-cube' (double take) there are 249 single wins for gnubg (38.43% of 648) and 399 single wins for MaX (61.57% of 648) with cube at 8, for a total of 648 trials. Cube stats are empty, since no recube can take place.
On the other hand, in the panel 'Player MaX owns 4-cube' (no double) there are 28 single wins with cube at 4 and 157 with cube at 8 for gnubg (total is 185, 28.55% of 648) and 1 single win with cube at 4 and 133 with cube at 8 for MaX (total is 144, 20.68% of 648). Total is 185+134 = 319 ?! Where are the other 329 trials ?!

Further down in the cube statistics it says that at 8-cube there has been 290 double,take and 329 double,pass for player MaX ... Does this mean that in the first part (Win statistics) the double,pass are not included ? To clarify (correctme if I'm wrong), in the win stats only games (trials) played to the end (no double,pass) are shown ?!
Is this kinda standard (ie snowie does the same) ?

Another examples of stats I can't understand :

GNU Backgammon Position ID: 4Z0ZA0Armw0BCg
Match ID : MAHgATAAWAAA
+-1--2--3--4--5--6-------7--8--9-10-11-12-+ O: GnuBG
| O O O O O | | O O X | 3 points
| O O O | | O O X | On roll
| | | |
| | | |
| | | |
| |BAR| |^ 15 point match (Cube: 1)
| | | |
| X | | |
| X X | | |
| X X | | X X |
| X O O X X | X | X X O | 11 points
+24-23-22-21-20-19------18-17-16-15-14-13-+ X: BGBlitz

I do a rollout (648, full, cubefull, expert both players) on this cube decision from the hint window and I get in the rollout stats window :

Centerd 1-cube :
Win stats : all 0 (all the trials ended with a double pass ?!)
Cube stats : 2-cube 161 double,take 290 double,pass for gnubg
2-cube 38 double,take 29 double,pass for BGBlitz
4-cube 3 double,take 3 double,pass for gnubg
4-cube 0 double,take 0 double,pass for BGBlitz
Totals : 164+293 + 38+29 = 524 ??!!

BGBlitz owns 2-cube :
Win stats : all 0 ??!!
Cube stats : all 0 ??!!
Totals : all 0 ??!!

Massimiliano Maini <maxmaini>

 

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

Attach File(s):
   
   
Comment:
   

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -unavailable- added by c_anthon (Posted a comment)
  •  

    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 2 latest changes.

    Date Changed By Updated Field Previous Value => Replaced By
    Sat 02 Dec 2006 10:07:18 PM UTCc_anthonOpen/ClosedOpen=>Closed
      StatusNone=>Fixed

    Back to the top


    Powered by Savane 3.1-cleanup1