[Bug-gnubg] Ply analysis does not give equal GWC and MWC in 1 point match.

[bob koca]

   Very latest build being used. Gets down to a symmetric medium length race in one point match.
 
GNUbg ID is  2+4OAADb7gYEAA:cAknAAAAAAAA
 
  It is 1 point match so game win chance equals MWC.  0-ply and RO report that correctly but
1, 2, 3, and 4 ply are crazy.
 
 
    1. Rollout          13/6                         MWC:  36.99590%
       0.369959 0.000000 0.000000 - 0.630041 0.000000 0.000000 CL  36.99590%
      [0.000073 0.000000 0.000000 - 0.000073 0.000000 0.000000 CL   0.00733%]
        Full cubeless rollout with var.redn.
        1296 games, Mersenne Twister dice gen. with seed 1000000 and quasi-random dice
        Play: 0-ply cubeful
        Cube: 2-ply cubeful prune
    2. Cubeful 4-ply    13/7 4/3                     MWC:  94.59189% ( 57.59599%)
       0.304787 0.000000 0.000000 - 0.695213 0.000000 0.000000
        4-ply cubeful prune
    3. Cubeful 3-ply    13/7 3/2                     MWC:  11.25431% (-25.74159%)
       0.335456 0.000000 0.000000 - 0.664544 0.000000 0.000000
        3-ply cubeful prune [grandmaster]
    4. Cubeful 2-ply    13/7 5/4                     MWC:  99.26514% ( 62.26924%)
       0.335899 0.000000 0.000000 - 0.664101 0.000000 0.000000
        2-ply cubeful prune [world class]
    5. Cubeful 1-ply    13/7 2/1                     MWC:   0.12791% (-36.86799%)
       0.328583 0.000000 0.000000 - 0.671417 0.000000 0.000000
        1-ply cubeful prune
    6. Cubeful 0-ply    13/7 6/5                     MWC:  33.08157% ( -3.91434%)
       0.330816 0.000000 0.000000 - 0.669184 0.000000 0.000000
        0-ply cubeful prune [expert]


 
Bob Koca

---

Windows Live™: E-mail. Chat. Share. Get more ways to connect. Check it out.