Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: What is the best positional test for chess programs ???? ANYBODY KNOW??

Author: blass uri

Date: 09:23:40 04/21/99

Go up one level in this thread



On April 21, 1999 at 12:08:03, blass uri wrote:

>
>On April 21, 1999 at 10:09:52, blass uri wrote:
>
>>
>>On April 21, 1999 at 09:48:43, Harald Faber wrote:
>>
>>>On April 21, 1999 at 09:15:56, blass uri wrote:
>>>
>>>>>>I think that crafty16.6 may be the best program against humans and programs.
>>>
>>>>>MAYBE. Crafty made a big step forward!
>>>
>>>>>>It won 2.5:.5 against Rebel10C in 1 hour per move games and ed told us in the
>>>>>>rebelboard that it had a won position in the 4rd game that was lost.
>>>>>
>>>>>You can't convince me with 2.5-1.5 results or so, give me more.
>>>>
>>>>It is probably 3.5:.5 because ed said that crafty was winning the 4th game.
>>>>It is not a clear proof but I have not more games of crafty at slow time
>>>>control.
>>>
>>>Even if it would go on that way it would only be proof that Crafty beats Rebel.
>>>There would still remain the question how Crafty does against other opponents.
>>>
>>>>Ed will continue the games after WCCC and we will see.
>>>>Uri
>>>
>>>It will take MUCH time...
>>>A better Crafty auto232 support would result in more games to have a better
>>>impression.
>>>BTW are there any recommandations for an opening book for Crafty engine in
>>>Fritz-GUI? The Fritz powerbook?
>>>
>>>And BTW choose your opponent for the Deep Thought-position, you didn't answer
>>>yet. I'll try WChess and MCP as black, you choose the white player.
>>>Genius or your favourite Junior? :-)
>>
>>I prefer Junior at long time control because Genius is a root processor and can
>>changes it's evaluation after actions like trading queens.
>>
>>I think that in this tactical position it is better to play the first 10 moves
>>at slower time control (at least 10 minutes per move) because it is important
>>not to miss something tactical.
>>
>>I can tell you Junior5.4's and Genius3's moves with white at slow time
>>control(at least 10 minutes per move.
>>
>>Junior follow the game and play a different move only when it choose Kd1 or Kf1
>>instead of Kd2 when Genius3 choose 17.Ne4 Nf5 18.Nfg5 Bxc5 19.Qg4
>
>I think that in this line after 19...Qe5 white can rsign
>
>Junior5.4 shows 3.62 pawns advantage for black after the moves 19.Qg4 Qe5
>
>Genius3 failed low at depth 9 in move 19 (the evaluation at depth 8 was -0.15
>with the line Qg4 Qe5 Rb1 h6 Bb2 Bd4 Bxd4 Nxd4)
>
>The main line of Junior5.4 after the moves Qg4 Qe5 is Rb1 h6 Qxf4 Bxf4 hxg5 Rb8
>Rab8 Bxb8 Rxe4+ Kd1 Ne3+ Kd2 Nxd5 Bxd5 Rd4+ Kc3 Rxd5


Genius3 showed at depth 9 -2.21 evaluation with the line
Qg4 Qe5 Rb1 h6 Bb2 Bd4 Bxd4 Nxd4

It is clear that genius3 is not the right opponent to defend this position.

Uri



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.