Re: [Bug-gnubg] new Raccoon's brain

[Øystein O Johansen]

Hi,

Try to call  EvalCacheResize(0); before you call FindBestMove.

I'm working on somthing similar here, and that was the first problem I
think I solved. (solved right now!) It's trying to find the eval from the
cache which is obviously not cached. I wonder why?

The next problem is that qsort in line 3179 in eval.c fails when I'm
evaluation at higher plies. Can anyone explain that?

-Øystein



|--------+--------------------------------------------->
|        |          "Jaro" <address@hidden>        |
|        |          Sent by:                           |
|        |          bug-gnubg-bounces+ojohans=statoil.c|
|        |          address@hidden                         |
|        |                                             |
|        |                                             |
|        |          27.02.2003 03:03                   |
|        |                                             |
|--------+--------------------------------------------->
  
>------------------------------------------------------------------------------------------------------------|
  |                                                                             
                               |
  |       To:     <address@hidden>                                              
                            |
  |       cc:     (bcc: Øystein O Johansen)                                     
                               |
  |       Subject:     Re: [Bug-gnubg] new Raccoon's brain                      
                               |
  
>------------------------------------------------------------------------------------------------------------|




I still can not get correct answer from DLL.

Here is my situation: I'm trying to get best move for position calling
FindBestMove function


Board ("4HPKATDgc/ABMA") looks like this


{{ 0,0,0,0,0,5,0,3,0,1,1,0,3,0,0,0,0,0,0,0,0,0,0,2,0 },
{ 0,0,0,0,0,5,0,3,0,0,0,0,5,0,0,0,0,0,0,0,0,0,0,2,0 }};


And here ara my results:


-13.949%     Move: (6, 6) [24 -> 18] [24 -> 18] [13 -> 7] [8 -> 2]
-6.404%     Move: (5, 6) [8 -> 3] [13 -> 7]
-6.56%     Move: (5, 5) [13 -> 8] [13 -> 8] [13 -> 8] [8 -> 3]
0%     Move: (4, 6) [24 -> 20] [20 -> 14]
0%     Move: (4, 5) [24 -> 20] [20 -> 15]
0%     Move: (4, 4) [24 -> 20] [24 -> 20] [13 -> 9] [13 -> 9]
0%     Move: (3, 6) [24 -> 21] [21 -> 15]
-1.298% Move: (3, 5) [24 -> 21] [13 -> 8]
-5.327% Move: (3, 4) [8 -> 5] [6 -> 2]
-0.69%     Move: (3, 3) [24 -> 21] [24 -> 21] [21 -> 18] [18 -> 15]
-3.691%     Move: (2, 6) [8 -> 6] [13 -> 7]
-8.127%     Move: (2, 5) [8 -> 6] [8 -> 3]
-9.19%     Move: (2, 4) [6 -> 4] [6 -> 2]
0%     Move: (2, 3) [24 -> 22] [13 -> 10]
-0.486% Move: (2, 2) [24 -> 22] [24 -> 22] [13 -> 11] [13 -> 11]
-8.306% Move: (1, 6) [6 -> 5] [8 -> 2]
-7.231% Move: (1, 5) [8 -> 7] [8 -> 3]
-4.394% Move: (1, 4) [8 -> 7] [7 -> 3]
-9.281% Move: (1, 3) [6 -> 5] [6 -> 3]
-2.225% Move: (1, 2) [8 -> 7] [13 -> 11]
-11.897% Move: (1, 1) [24 -> 23] [6 -> 5] [6 -> 5] [5 -> 4]


First number is difference from the best move according to GNU backgammon.
Then dices and moves.


Cubeinfo is


ci->nCube = 1;
ci->fCubeOwner = -1;
ci->fMove = 0;
ci->nMatchTo = 1;
ci->anScore[0] = 0; ci->anScore[1] = 0;
ci->fCrawford = 0 ;
ci->fJacoby = 0;
ci->fBeavers = 0;
ci->arGammonPrice[0] = 0;
ci->arGammonPrice[1] = 0;
ci->arGammonPrice[2] = 0;
ci->arGammonPrice[3] = 0;


evalcontext


ec->bCubeful = 1;
ec->nPlies = 0;
ec->nReduced = 0;
ec->bDeterministic = 0;
ec->rNoise = 0.0;


I tryied many things, and can not get working right.


Any suggestions?


Jaro





 _______________________________________________
Bug-gnubg mailing list
address@hidden
http://mail.gnu.org/mailman/listinfo/bug-gnubg









-------------------------------------------------------------------
The information contained in this message may be CONFIDENTIAL and is
intended for the addressee only. Any unauthorised use, dissemination of the
information or copying of this message is prohibited. If you are not the
addressee, please notify the sender immediately by return e-mail and delete
this message.
Thank you.