BGonline.org Forums

Gnubg odd/even effect extreme example ?

Posted By: MaX
Date: Friday, 21 August 2009, at 2:19 p.m.

In Response To: Why do bots have an even/odd effect? (Tom Keith)

X is closed out. Evaluating 0ply with O on roll gives O 76.8% GWC while evaluating with X on roll gives X 12.6% (hence 87.4% for O). As stated by Tom, chequer play may be unaffected, but cube decisions are and different plies give very different results for the doubling decision for white: swings a lot from marginal take to huge pass across odd/even plies.

    GNU Backgammon  Position ID: DwAAsN22AYgAAA
                    Match ID   : MAHgAAAAAAAA
    +-1--2--3--4--5--6-------7--8--9-10-11-12-+  O: gnubg
    | O  O  O  O  O  O |   |                  |  0 points
    | O  O  O  O  O  O |   |                  |  On roll
    |    O             |   |                  |  
    |                  |   |                  |  
    |                  |   |                  |  
    |                  |BAR|                  |^ 7 point match (Cube: 1)
  X |                  |   |                  |  
 XX | X                |   |                  |  
 XX | X                |   |                  |  
 XX | X                | X |                  |  
 XX | X              O | X |       O          |  0 points
    +24-23-22-21-20-19------18-17-16-15-14-13-+  X: MaX

O on roll:
Position ID: DwAAsN22AYgAAA
Match ID: MAHgAAAAAAAA
Evaluator:  Crashed
        Win     W(g)    W(bg)   L(g)    L(bg)   Equity    Cubeful  
static:  76.8     0.0     0.0     0.1     0.0    +0.534    +0.785   
 1 ply:  88.0     0.0     0.0     0.1     0.0    +0.759    +0.923   
 2 ply:  79.7     0.0     0.0     0.1     0.0    +0.593    +0.999   
 3 ply:  87.7     0.0     0.0     0.0     0.0    +0.754    +1.000   
 4 ply:  82.0     0.0     0.0     0.0     0.0    +0.639    +1.000   

 
X on roll:
Position ID: u20DEAEPAAAwAA
Match ID: cAngAAAAAAAA
Evaluator:  Crashed
        Win     W(g)    W(bg)   L(g)    L(bg)   Equity    Cubeful  
static:  12.6     0.1     0.0     0.0     0.0    -0.746    -0.919   
 1 ply:  23.2     0.1     0.0     0.0     0.0    -0.534    -0.930   
 2 ply:  12.0     0.1     0.0     0.0     0.0    -0.759    -1.000   
 3 ply:  20.3     0.1     0.0     0.0     0.0    -0.593    -1.000   
 4 ply:  12.3     0.0     0.0     0.0     0.0    -0.754    -1.000   

Similar, with O's board not closed, problem is still there:

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

O on roll:
Position ID: DwAAMLvtAogAAA
Match ID: MAHgAAAAAAAA
Evaluator:  Crashed
        Win     W(g)    W(bg)   L(g)    L(bg)   Equity    Cubeful  
static:  73.6     0.0     0.0     0.3     0.0    +0.469    +0.691   
 1 ply:  86.5     0.0     0.0     0.3     0.0    +0.726    +0.912   
 2 ply:  75.5     0.0     0.0     0.2     0.0    +0.507    +0.861   
 3 ply:  86.6     0.0     0.0     0.4     0.0    +0.727    +0.989   
 4 ply:  77.3     0.0     0.0     0.3     0.0    +0.543    +0.928  
 

X on roll: 
Position ID: dtsFEAEPAAAwAA
Match ID: cAngAAAAAAAA
Evaluator:  Crashed
        Win     W(g)    W(bg)   L(g)    L(bg)   Equity    Cubeful  
static:  14.9     0.5     0.0     0.0     0.0    -0.697    -0.903   
 1 ply:  27.1     0.4     0.0     0.0     0.0    -0.453    -0.744   
 2 ply:  14.2     0.5     0.0     0.0     0.0    -0.710    -0.987   
 3 ply:  25.4     0.4     0.0     0.0     0.0    -0.489    -0.829   
 4 ply:  14.0     0.5     0.0     0.0     0.0    -0.715    -0.988   

I'm wondering if other bots have the same issue at 0ply ... Anyone with Snowie/XG/BGB ?

MaX.

Messages In This Thread

• Falafel to play 41
  Stick -- Monday, 10 August 2009, at 3:33 a.m.
  • Falafel to play 41
    Havard Raddum -- Monday, 10 August 2009, at 6:31 a.m.
  • Falafel to play 41
    sonrik -- Monday, 10 August 2009, at 6:51 a.m.
  • Falafel to play 41
    mamabear -- Monday, 10 August 2009, at 1:06 p.m.
  • Falafel to play 41
    Keene -- Monday, 10 August 2009, at 3:51 p.m.
  • Falafel to play 41
    garyo -- Monday, 10 August 2009, at 4:37 p.m.
  • Falafel to play 41
    Gustavo Goldfarb -- Monday, 10 August 2009, at 4:56 p.m.
  • Falafel to play 41
    Fabrice Liardet -- Monday, 10 August 2009, at 8:01 p.m.
  • Rollout
    Stick -- Tuesday, 11 August 2009, at 1:21 a.m.
    • Rollout
      neilkaz -- Tuesday, 11 August 2009, at 3:22 a.m.
    • Rollout
      ScottyK -- Tuesday, 11 August 2009, at 1:57 p.m.
      • Re: GNU 4 ply..GV etc
        neilkaz -- Tuesday, 11 August 2009, at 3:07 p.m.
      • Plies again
        Matt Cohn-Geier -- Tuesday, 11 August 2009, at 4:45 p.m.
        • Plies again
          Chuck Bower -- Tuesday, 11 August 2009, at 8:43 p.m.
          • Plies again
            Matt Cohn-Geier -- Wednesday, 12 August 2009, at 3:26 a.m.
            • Plies again
              Chuck Bower -- Wednesday, 12 August 2009, at 1:48 p.m.
        • Plies again
          ScottyK -- Tuesday, 11 August 2009, at 9:28 p.m.
          • Plies again
            Keene -- Tuesday, 11 August 2009, at 9:59 p.m.
            • Plies again
              ScottyK -- Tuesday, 11 August 2009, at 10:23 p.m.
        • Plies again
          playBunny -- Tuesday, 11 August 2009, at 11:15 p.m.
          • Plies again
            MaX -- Wednesday, 12 August 2009, at 11:29 a.m.
            • Plies again
              playBunny -- Wednesday, 12 August 2009, at 1:15 p.m.
              • Why do bots have an even/odd effect?
                Tom Keith -- Wednesday, 12 August 2009, at 2:11 p.m.
                • Why do bots have an even/odd effect?
                  MaX -- Wednesday, 12 August 2009, at 3:27 p.m.
                • Why do bots have an even/odd effect?
                  playBunny -- Wednesday, 12 August 2009, at 7:51 p.m.
                  • Why do bots have an even/odd effect?
                    Tom Keith -- Thursday, 13 August 2009, at 1:54 p.m.
                    • Why do bots have an even/odd effect?
                      playBunny -- Thursday, 13 August 2009, at 4:50 p.m.
                  • Why do bots have an even/odd effect?
                    MaX -- Saturday, 15 August 2009, at 8:53 p.m.
                    • Why do bots have an even/odd effect?
                      Tom Keith -- Monday, 17 August 2009, at 11:16 p.m.
                      • Why do bots have an even/odd effect?
                        playBunny -- Tuesday, 18 August 2009, at 12:59 a.m.
                        • Why do bots have an even/odd effect?
                          Tom Keith -- Tuesday, 18 August 2009, at 12:44 p.m.
                          • Why do bots have an even/odd effect?
                            playBunny -- Tuesday, 18 August 2009, at 4:39 p.m.
                            • Why do bots have an even/odd effect?
                              Tom Keith -- Tuesday, 18 August 2009, at 5:45 p.m.
                      • Why do bots have an even/odd effect?
                        Maik Stiebler -- Tuesday, 18 August 2009, at 6:14 a.m.
                        • Why do bots have an even/odd effect?
                          Tom Keith -- Tuesday, 18 August 2009, at 12:46 p.m.
                      • Why do bots have an even/odd effect?
                        MaX -- Thursday, 20 August 2009, at 2:54 p.m.
                      • Gnubg odd/even effect extreme example ?
                        MaX -- Friday, 21 August 2009, at 2:19 p.m.
                        • Gnubg odd/even effect extreme example ?
                          Gregg Cattanach -- Friday, 21 August 2009, at 7:24 p.m.
                        • Gnubg odd/even effect extreme example ?
                          Frank Berger -- Saturday, 22 August 2009, at 6:02 a.m.
      • Rollout
        Rich Munitz -- Tuesday, 11 August 2009, at 7:35 p.m.
        • Rollout
          ScottyK -- Tuesday, 11 August 2009, at 9:22 p.m.
          • Rollout
            Rich Munitz -- Wednesday, 12 August 2009, at 12:14 a.m.
            • Rollout
              ScottyK -- Wednesday, 12 August 2009, at 1:23 a.m.
          • Rollout
            Chuck Bower -- Wednesday, 12 August 2009, at 1:52 p.m.
  • Falafel to play 41
    Toppace -- Wednesday, 12 August 2009, at 6:20 p.m.
    • Falafel to play 41
      Keene -- Wednesday, 12 August 2009, at 7:29 p.m.