Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Junior 7 makes comeback this time.

Author: James T. Walker

Date: 07:43:51 07/14/01

Go up one level in this thread


On July 14, 2001 at 10:04:51, Uri Blass wrote:

>On July 14, 2001 at 09:52:58, Uri Blass wrote:
>
>>On July 14, 2001 at 09:24:32, James T. Walker wrote:
>>
>>>On July 14, 2001 at 04:14:04, Uri Blass wrote:
>>>
>>>>On July 14, 2001 at 01:33:45, Sune Larsson wrote:
>>>>
>>>>>On July 14, 2001 at 00:01:18, James T. Walker wrote:
>>>>>
>>>>>>I just finished a 20 game match between Chess Tiger 14.0 and Junior 7 on 2
>>>>>>Athlon 900 TB's with 512M ram.  Chess Tiger was leading by 9-5 after 14 games
>>>>>>but the match ended in a 10-10 tie.  (No playoff this time)
>>>>>>Junior seems to have a serious problem as shown in this last game which Junior 7
>>>>>>won but not without difficulty.  Junior 7 announced Mate in 7 on move 79 and
>>>>>>mated on move 105 ??  That is rediculous!
>>>>>
>>>>>
>>>>> Maybe a designer's choice?! So that, in this case, "mate in 7" should
>>>>> read "mate in 7 VLM" (Very Long Moves)...;)
>>>>>
>>>>> Sune
>>>>
>>>>No.
>>>>The mate in 7 is correct but there is no rule in chess that you have to mate in
>>>>less moves so Junior simply does not care about choosing the best moves and
>>>>choosing a move that leads to mate is enough for it.
>>>>
>>>>Uri
>>>
>>>OK Uri,
>>>Here is the very next game vs Fritz 6.  Junior announces Mate in 13 on move 75
>>>and gets a draw on move 142 ??  This is a problem that should be fixed!
>>
>>I suspect that in this case Junior has a wrong evaluation because of some
>>problem in the stalemate detection of Junior.
>>
>>I suspect that it is going to be never fixed because Amir Ban explained here
>>that doing tactical mistakes in rare situations is the way that the engine works
>>and it is not a bug.
>>
>>I suspect that the positions that Junior says mate for itself are a dead draw
>>positions(I cannot see a plan to win for black)
>>
>>Here is one of the relevant positions that only Junior can see mate for black.
>>
>>[D]4B3/2b3p1/8/8/4b3/4k3/6Pp/7K b - - 0 1
>>
>>Uri
>
>After thinking about it I believe that the only plan for black is to sacrifice
>h2.
>
>black wants to play Kf2 and Bxg2# but white can sacrifice the bishop in the log
>diagnol in order to get stalemate.
>
>Black need to go with the bishop out of the diagnol that protect h2 before
>trying the plan of Kf2 and Bxg2# and it means sacrificing the h pawn.
>
>Uri

I think you are right.  Black must give up the h2 pawn to win.  I guess the
stalemate problem is giving Junior problems.  It doesn't want to give up the
pawn but needs to do it to prevent the stalemate threat and win.  Despite this
problem with Junior 7 I am impressed with it's play.  So far after 50 games it
is 26.5-23.5 vs (Fritz 6.. 6-4)(Chess Tiger 14.0.. 10-10) & (Gambit
Tiger..10.5-9.5).  All game/60 minutes.  This puts Junior 7 at the top of my ELO
list for my "Standard" database.  It's only 50 games but it's off to a nice
start against the top programs I have.  I'm doing 10 more games vs Fritz 6 and
then trying it vs Deep Shredder.
Jim



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.