[Bug-gnubg] More on use of noise in rollouts

[bob koca]

  I checked and the previous problem of the noise making a difference even though every move is forced is gone now.
 
  I did a RO of the opening position with RO playing as intermediate vs. expert player. I obtained that the intermediate wins about
17% of 1 point matches. So something is wrong, possibly just labeling the amount of noise given as being intermediate. Note that
2200 vs 800 ELO gives about 17% win chance which is more like world class vs. beginner.
 
  Also it seems strange that the opening roll is not worth anything. When i gave the intermediate player the opening roll (non doubles)
the intermediate won 16.6668% of the matches. When the expert got it the intermediate won 17.2225% matches. Could it really be
that the play is so bad that the intermediate does better to not win the opening roll?  Or did I mess something up? Thanks for any help,
 
Bob Koca
 
     This is with the expert having first roll:
 
 GNU Backgammon  Position ID: 4HPwATDgc/AIMA
                    Match ID   : MAElAAAAAAAA
    +-1--2--3--4--5--6-------7--8--9-10-11-12-+  O: gnubg
    | X              O |   |    O           X |  0 points
    | X              O |   |    O           X |  Rolled 21
    |                O |   |    O           X | 
    |                O |   |                X | 
    |                O |   |                X | 
    |                  |BAR|                  |^ 1 point match (Cube: 1)
    |                X |   |                  | 
    |                X |   |                O | 
    |                X |   |    X           O | 
    | O              X |   |    X           O | 
    | O              X |   |    X  O        O |  0 points
    +24-23-22-21-20-19------18-17-16-15-14-13-+  X: user
 
1. Rollout          16/13                        MWC:  17.22248%
       0.172225 0.051257 0.006252 - 0.827775 0.490876 0.128110 CL  17.22248%
      [0.006618 0.005328 0.001462 - 0.006618 0.009700 0.008089 CL   0.66179%]
        Full cubeless rollout with var.redn.
        1296 games, rollout as initial position, Mersenne Twister dice generator with seed 24000
        Player 0:
        Play: 0-ply cubeful, noise 0.04 (d) [intermediate]
        Cube: 0-ply cubeful, noise 0.04 (d)
        Player 1:
        Play: 0-ply cubeful [expert]
        Cube: 0-ply cubeful, noise 0.04 (d)
 
 
And This is with the intermediate having the first roll:
 
     GNU Backgammon  Position ID: 4HPwATDgc/AgMA
                    Match ID   : cAktAAAAAAAA
    +24-23-22-21-20-19------18-17-16-15-14-13-+  O: gnubg
    | X              O |   | X  O           X |  0 points
    | X              O |   |    O           X | 
    |                O |   |    O           X | 
    |                O |   |                X | 
    |                O |   |                  | 
    |                  |BAR|                  |v 1 point match (Cube: 1)
    |                X |   |                O | 
    |                X |   |                O | 
    |                X |   |    X           O | 
    | O              X |   |    X           O |  Rolled 23
    | O              X |   |    X           O |  0 points
    +-1--2--3--4--5--6-------7--8--9-10-11-12-+  X: user
 
 
            1. Rollout          18/13                        MWC:  83.33321%
       0.833332 0.487893 0.118926 - 0.166668 0.041701 0.003942 CL  83.33321%
      [0.006815 0.009393 0.008023 - 0.006815 0.006001 0.001675 CL   0.68155%]
        Full cubeless rollout with var.redn.
        1296 games, rollout as initial position, Mersenne Twister dice generator with seed 24000
        Player 0:
        Play: 0-ply cubeful, noise 0.04 (d) [intermediate]
        Cube: 0-ply cubeful, noise 0.04 (d)
        Player 1:
        Play: 0-ply cubeful [expert]
        Cube: 0-ply cubeful, noise 0.04 (d)
 
 
 
 

---

Windows Live™: Keep your life in sync. Check it out.