gnugo-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gnugo-devel] regress gunnar:8


From: kevin yong
Subject: Re: [gnugo-devel] regress gunnar:8
Date: Fri, 11 Apr 2003 12:56:26 -0400 (EDT)

Hi, all:

  i had a quick look at regress case gunnar:8.

  problem: O11 wrong (completely waist a move).

  i went through the log file and found:
    
    O13: 35.38 - attack last move played, although it
      seems dead

  also i look at the source code:

      /* If the opponent just added a stone to a dead
dragon, then
       * attack it. If we are ahead, add a safety move
here, at most
       * half the margin of victory.
       *
       * This does not apply if we are doing scoring.
       */
      if (!doing_scoring
          && is_same_dragon(get_last_opponent_move(color),
aa)) {
        this_value = 1.5 * dragon[aa].effective_size;
        TRACE("  %1m: %f - attack last move played, although
it seems dead\n",
              pos, this_value);


So, in this case, gnugo is doing what it programmed to
do. the question is: should we keep or drop this rule:
'attack last move played, although it seems dead'. or
need some enhancement?
If we want to keep this rule, then should we raise
this case as a problem? maybe we can add some comments
to it, so that other readers of the case can have more
info about what's going on?

kevin.









______________________________________________________________________ 
Post your free ad now! http://personals.yahoo.ca




reply via email to

[Prev in Thread] Current Thread [Next in Thread]