Author: Tim Foden
Date: 00:49:32 12/28/04
Go up one level in this thread
On December 27, 2004 at 19:32:05, Dieter Buerssner wrote:
>On December 27, 2004 at 17:19:43, Peter Fendrich wrote:
>
>>Do you know of any program that solves that in a decent time?
>
>Tim showed the analysis by GLC. Very impressive! (Order of magnitude 1 s -
>probably GLC also calculated its internal TBs in that time)
Generating internal KP-K end-game table...
Done (0.171 secs)
>
>Yace needs 2 minutes - and gives rather significant score, then. GLC did not
>give that large score. Seems GLC makes BOC scores very drawish - this will of
>course help.
>
>Regards,
>Dieter
Yes, it does make it very drawish. Here's the code (it's basically the last
term in the eval function)...
Generating internal KP-K end-game table...
Done (0.171 secs)
//
// if we are in the endgame, and we have opposite coloured bishops,
// we move the score closer to a draw
//
if( m_stage == stageEndgame && m_nMen < 20 &&
m_simpleMat[white] == m_simpleMat[black] )
{
if( m_count[wBishop + 6] == 1 && m_count[bBishop + 6] == 1 )
{
Square wB = BitScanForward(m_bishopBoard[white]);
bool wBw = m_squareColour[wB];
if( m_colourSquares[!wBw] & m_bishopBoard[black] )
{
// got opposite coloured bishops
#ifdef _DEBUG_SCORE
COutput::Log( _T(" *B** opposite coloured bishops %d\n"),
eval2 * m_nMen / 20 - eval2 );
#endif
eval2 = eval2 * m_nMen / 20;
}
}
}
Cheers, Tim.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.