Author: Bernhard Bauer
Date: 00:05:15 11/18/05
Go up one level in this thread
On November 17, 2005 at 14:57:07, Uri Blass wrote:
>On November 17, 2005 at 14:27:48, Dieter Buerssner wrote:
>
>>On November 17, 2005 at 10:13:27, Uri Blass wrote:
>>
>>>The original position is mate in 36 based on backward analysis by yace
>>
>>You might be interested to know how to find the mate score in such a position
>>with Yace even without TBs. Set Endgame Database usage to high (or when using
>>console mode, tb_u 4). With your computer, it should take less than 3 minutes,
>>to show a mate in 36 (in console mode even with a PV until the end).
>>
>>1.Ke4 Ke8 2.Ke3 Kd7 3.Kd3 Kc8 4.Kc4 Kd8 5.Kd4
>>Ke8 6.Kc5 Kd7 7.Kd5 Ke7 8.Kc6 Kf6 9.Kd6 Kf7
>>10.Kd7 Kf8 11.Ke6 Ke8 12.f6 g6 13.f7+ Kf8
>>14.Kd6 Kxf7 15.Kd7 Kf8 16.Ke6 Kg7 17.Ke7 Kh7H
>>18.Kf7H Kh6H 19.Kf6H Kh7H 20.Kxg5H {EGTB}
>>20...Kg7! 21.h4! Kh7 22.Kf6! Kh6! 23.g5+! Kh5!
>>24.Kf7 Kxh4 25.Kxg6!! Kg4 26.Kf6! Kh4 27.g6!
>>Kg4 28.g7 Kf4 29.g8=Q! Ke4 30.Qd8 Ke3 31.Kf5
>>Kf2! 32.Qd3 Kg2 33.Qe2+ Kg3 34.Kg5! Kh3 35.Kf4!
>>Kh4 36.Qh2# {1001}
>>
>>Yace thinks, that Kd4 is also a mate in 36, Kd5 mate in 37 and the pawn moves
>>are draw.
>>
>>BTW. I am not sure, if Yace was here just lucky. (José Raul Capablanca, "A good
>>player is always lucky." :-)
>>
>>Regards,
>>Dieter
>
>Fruit is a better player than yace and cannot find the mate even with tablebases
Fruit may get better results in engine-matches, but that doesn't mean Fruit is a
better player in every part of the game of chess.
There are examples suggesting that Yace is a better player than Fruit when it
comes to the endgame. That is understandable when you consider the huge
experience of the author of Yace, Dieter, in using EGTBs and Bitbases whithin
his program.
>so the question is how Capablanca can explain Fruit's failure:-)
>
Yace has not to explain the failure of Fruit.
>Of course I know that yace is good in this type of positions and hopefullty some
>improvement in fruit's hash may help it to solve the position(I even did not try
>them with Movei because when hash is not used for pruning I do not believe that
>it has chances to solve it inspite of the fact that Movei seems to be at
>similiar strength to Yace and scored half point more than yace in the following
>tournament http://www.husvankempen.de/nunn/tabellen/promo.htm).
>
The overall strength of Movei and the results of Movei against Yace have not
much to do whith the endgame behavior of both engines. I would expect Yace to be
better in this part of the game. As I do not have Movei, I can not prove my
expectation.
>Uri
Bernhard
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.