Author: Matthew White
Date: 13:37:44 04/03/03
Go up one level in this thread
On April 03, 2003 at 09:17:07, Ingo Bauer wrote: >On April 03, 2003 at 09:06:28, Uri Blass wrote: > >>On April 03, 2003 at 09:02:36, Uri Blass wrote: >> >>>On April 03, 2003 at 07:56:04, Jouni Uski wrote: >>> >>>>[D]8/8/8/5bk1/8/8/4B1PP/5K2 b - - >>>> >>>>S7 don't find -Bh3 in 10 minutes. Why???? I don't have full 5 piece TBs for >>>>sure, but most amateurs find move even without (any) TBs. >>>> >>>>Jouni >>> >>>There are amateurs that can find it(Comet,List,Crafty,Goliath Light) but >>>I think that you are wrong to assume that most amateurs find it >>>without any tablebases. >> >>The interesting thing is that after Bh3 gxh3 shredder7.04 has >>no problem to see score that is close to 0.00 >> >>Uri > >And even without these two moves S7.04 finds the right move (even without TBS). >I think SMK just assumes that EGTB are installed and invests his time in the >developing of the middlegame. I agree with your guess about investing in the development in the middlegame, however, I disagree with the idea (maybe if I repeat it enough, people will start doing it :)). Yes, one of the ideas behind the middlegame is to create a mating attack. However, there is another idea: to reach a won endgame. If an engine does not know enough about endgames to reach a winning one (I think that most engine matches end up in an endgame situation), then the whole middlegame was wasted. One of the problems with relying on tablebases is that they are only effective VERY late into the endgame. Most endgames that I have played and observed have several pawns on the board for each side. This means that most of the time, the EGTBs will never make it into play until the pawns have been traded off. However, leaving the pawns on may have resulted in a zugzwang and won immediately... I realize that the endgame is one of the most difficult parts of the game to add knowledge to (many of the rules that were true for the middlegame and opening are totally abandoned in the endgame), however, it could potentially result in one of the greatest strength increases. Sorry for the rant, but it is something that I feel could improve the current "state of the art" in chess engines. Regards, Matt <SNIP!> > >Bye Ingo
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.