Author: Uri Blass
Date: 08:19:12 07/07/01
Go up one level in this thread
On July 07, 2001 at 09:44:08, Martin Müller wrote:
>Hi all,
>
>It is quite sad to see a famous program as DJ7 to loose that _way_. Not the loss
>is important. But how it looses. The evalutaion differed up to 3 or more pawns
>against several reputated programs. It makes me think that the evaluation of the
>program is not quite correct. Just see that:
>
>{\rtf1\ansi\deff0{\fonttbl{\f0\froman FigurineCB AriesSP;}{\f1\fnil
>DiagramTTFritz;}}{\info{\author
>Müller,Martin}{\creatim\yr2001\mo7\dy7\hr15\min38}}\fs20 {\pard \ql {\b \fs24
>(455) Deep Shredder,G - Deep Junior 7 [E48]\par}{\plain \fs20 Level=Blitz:60'.
>Z\'FCrich, 07.07.2001\par}}{\qj {\b {\plain \fs20 716MB, j7book.ctg. LINKS
> }1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 c5 5.Bd3 Nc6 6.Nge2 cxd4 7.exd4 d5 8.cxd5 Nxd5
>9.0\'960 0\'960 10.Nxd5 Qxd5 11.a3 Be7 12.Be3 Rd8 13.Nc3 Qd7 {\plain \fs20
>0.10/16 2:34 }14.Qc2 Nxd4 {\plain \fs20 0.33/18 1:47 }15.Bxh7+ Kh8 {\plain
>\fs20 0.37/19 1:58 }16.Bxd4 Qxd4 {\plain \fs20 0.48/17 51 }17.Rad1 Qh4
>{\plain \fs20 0.43/17 2:10 }18.Be4 Bd7 {\plain \fs20 0.52/17 2:16 }19.Rd3
>e5 {\plain \fs20 0.00/17 5 }20.Bxb7 Rab8 {\plain \fs20 0.36/15 49 }21.Be4
>g6 {\plain \fs20 0.39/16 1:25 }22.b4 Be6 {\plain \fs20 0.44/16 2:00
>}23.Nd5 Bd6 {\plain \fs20 0.52/14 1:08 }24.Rfd1 Rdc8 {\plain \fs20 0.57/16
>2:34 }25.Qe2 Kg7 {\plain \fs20 0.72/16 1:50 }26.g3 Qh8 {\plain \fs20
>0.87/16 40 }27.Qe3 Bg4 {\plain \fs20 1.08/16 3:24 }28.Re1 Bh3 {\plain \fs20
> 1.24/17 1:39 }29.Qxa7 Qf8 {\plain \fs20 1.30/17 1:52 }30.Qe3 Rd8 {\plain
>\fs20 1.29/16 1:55 }31.Red1 Qe8 {\plain \fs20 1.30/16 1:30 }32.Bf3 Qe6
>{\plain \fs20 1.40/16 55 }33.Nc3 Bc7 {\plain \fs20 1.11/16 48 }34.Rxd8
>Bxd8 {\plain \fs20 0.00/18 5 }35.Nd5 Qf5 {\plain \fs20 1.37/18 2:36 }36.a4
>Bg5 {\plain \fs20 1.42/17 2:34 }37.Qe2 Rc8 {\plain \fs20 1.50/17 1:01
>}38.a5 Rc2 {\plain \fs20 1.98/18 5 }39.Ne3 Bxe3 {\plain \fs20 0.00/19 5
>}40.Qxe3 Ra2 {\plain \fs20 0.00/18 5 }41.Bd5 Ra4 {\plain \fs20 0.00/18 5
>}42.Qc3 Qh5 {\plain \fs20 2.07/16 10 }43.Re1 Qf5 {\plain \fs20 0.00/18 6
>}44.Bb3 Ra3 {\plain \fs20 2.27/17 41 }45.Qb2 Qf3 {\plain \fs20 2.40/18 3:06
> }46.Qxe5+ Qf6 {\plain \fs20 #1/20 5 }47.Qxf6+ Kxf6 {\plain \fs20 #1/21 6
>}48.Bd5 Bd7 {\plain \fs20 2.48/21 5 }49.f4 Bb5 {\plain \fs20 2.32/17 30
>}50.Rc1 Ke7 {\plain \fs20 2.68/18 26 }51.Kf2 Ra4 {\plain \fs20 2.37/17 43
>}52.Re1+ Kf6 {\plain \fs20 2.37/18 44 }53.Rb1 Ke7 {\plain \fs20 0.00/21 5
>}54.Ke3 f6 {\plain \fs20 2.69/18 1:27 }55.Kd4 Kd6 {\plain \fs20 2.80/21 5
>}56.Bf7 g5 {\plain \fs20 2.62/17 33 }57.h4 gxh4 {\plain \fs20 2.62/16 29
>}58.gxh4 Ra3 {\plain \fs20 0.00/21 5 }59.Bb3 Bd7 {\plain \fs20 0.00/21 5
>}60.h5 Bf5 {\plain \fs20 3.19/21 5 }61.Rb2 Kc7 {\plain \fs20 0.00/20 5
>}62.h6 Kb8 {\plain \fs20 3.54/18 1:00 }63.Bd5 Kc7 {\plain \fs20 0.00/19 5
>}64.Rb3 Ra1 {\plain \fs20 0.00/21 5 }65.Rc3+ Kb8{\plain \fs20 4.40/18 3 }
>1\'960}\par \par}\par}
>
>Deep Shredder,G - Deep Junior 7 [E48]
>Level=Blitz:60'. Zürich, 07.07.2001
>
> 716MB, j7book.ctg. LINKS
> 1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 c5 5.Bd3 Nc6 6.Nge2 cxd4 7.exd4 d5 8.cxd5 Nxd5
>9.0-0 0-0 10.Nxd5 Qxd5 11.a3 Be7 12.Be3 Rd8 13.Nc3 Qd7 0.10/16 2:34 14.Qc2
>Nxd4 0.33/18 1:47 15.Bxh7+ Kh8 0.37/19 1:58 16.Bxd4 Qxd4 0.48/17 51
>17.Rad1 Qh4 0.43/17 2:10 18.Be4 Bd7 0.52/17 2:16 19.Rd3 e5 0.00/17 5
>20.Bxb7 Rab8 0.36/15 49 21.Be4 g6 0.39/16 1:25 22.b4 Be6 0.44/16 2:00
>23.Nd5 Bd6 0.52/14 1:08 24.Rfd1 Rdc8 0.57/16 2:34 25.Qe2 Kg7 0.72/16
>1:50 26.g3 Qh8 0.87/16 40 27.Qe3 Bg4 1.08/16 3:24 28.Re1 Bh3 1.24/17
>1:39 29.Qxa7 Qf8 1.30/17 1:52 30.Qe3 Rd8 1.29/16 1:55 31.Red1 Qe8
>1.30/16 1:30 32.Bf3 Qe6 1.40/16 55 33.Nc3 Bc7 1.11/16 48 34.Rxd8 Bxd8
>0.00/18 5 35.Nd5 Qf5 1.37/18 2:36 36.a4 Bg5 1.42/17 2:34 37.Qe2 Rc8
>1.50/17 1:01 38.a5 Rc2 1.98/18 5 39.Ne3 Bxe3 0.00/19 5 40.Qxe3 Ra2
>0.00/18 5 41.Bd5 Ra4 0.00/18 5 42.Qc3 Qh5 2.07/16 10 43.Re1 Qf5 0.00/18
> 6 44.Bb3 Ra3 2.27/17 41 45.Qb2 Qf3 2.40/18 3:06 46.Qxe5+ Qf6 #1/20 5
>47.Qxf6+ Kxf6 #1/21 6 48.Bd5 Bd7 2.48/21 5 49.f4 Bb5 2.32/17 30 50.Rc1
>Ke7 2.68/18 26 51.Kf2 Ra4 2.37/17 43 52.Re1+ Kf6 2.37/18 44 53.Rb1 Ke7
>0.00/21 5 54.Ke3 f6 2.69/18 1:27 55.Kd4 Kd6 2.80/21 5 56.Bf7 g5 2.62/17
> 33 57.h4 gxh4 2.62/16 29 58.gxh4 Ra3 0.00/21 5 59.Bb3 Bd7 0.00/21 5
>60.h5 Bf5 3.19/21 5 61.Rb2 Kc7 0.00/20 5 62.h6 Kb8 3.54/18 1:00 63.Bd5
>Kc7 0.00/19 5 64.Rb3 Ra1 0.00/21 5 65.Rc3+ Kb8 4.40/18 3 1-0
>
>Or
>
>Century 3 viz Deep Junior 7
>
>1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 c5 5.Bd3 Nc6 6.Ng-e2 cxd4 7.exd4 d5 8.cxd5 Nxd5
>9.O-O O-O 10.Nxd5 Qxd5 11.a3 Be7 12.Be3 Rd8 13.Nc3 Qd7 14.Qc2 Nxd4 15.Bxh7+ Kh8
>16.Bxd4 Qxd4 17.Ra-d1 Qh4 18.Be4 Bd7 19.Rd3 e5 20.Bxb7 Ra-b8 21.Be4 g6 22.b4 Be6
>23.Nd5 Bd6 24.Rf-d1 Rd-c8 25.Qe2 Kg7 26.g3 Qh8 27.Qe3 Bg4 28.Re1 Bh3 29.Qxa7 Qf8
>30.Qe3 Rc-d8 31.Re-d1 Qe8 32.Bf3 Qe6 33.Nc3 Bc7 34.Rd3xd8 Bxd8 35.Nd5 Qf5 36.a4
>Bg5 37.Qe2 Rc8 38.a5 Rc2 39.Ne3 Bxe3 40.Qxe3 Ra2 41.Bd5 Ra4 42.Qc3 Qh5 43.Re1
>Qf5 44.Bb3 Ra3 45.Qb2 Qf3 46.Qxe5+ Qf6 47.Qxf6+ Kxf6 48.Bd5 Bd7 49.f4 Bb5 50.Rc1
>Ke7 51.Kf2 Ra4 52.Re1+ Kf6 53.Rb1 Ke7 54.Ke3 f6 55.Kd4 Kd6 56.Bf7 g5 57.h4 gxh4
>58.gxh4 Ra3 59.Bb3 Bd7 60.h5 Bf5 61.Rb2 Kc7 62.h6 Kb8 63.Bd5 Kc7 64.Rb3 Ra1
>65.Rc3+ Kb8 66.Rc6 Resigns
>
>
>Century 3 on a mûch slower 1 processor computer.
>
>
>DJ7 does not see the problems it seems. What is wrong with DJ7?
>
>Regards
>Martin
I do not know what is wrong with Deep Junior7 and I guess that Junior7 is better
than Deep Junior7.
I saw Junior7 beating Deep Fritz 41-38 in 1 hour per game engine-engine game
in this club.
I did not see this kind of strange evaluation of 0.00 in positions like the
following position in Junior7's games
[D]8/2k5/5p1P/P2B1b2/1P1K1P2/rR6/8/8 b - - 0 1
It has to be a bug.
I do not know if the bug is in the static evaluation or in the updating of some
variables between moves.
I guess that Deep Junior7 was not tested enough because most buyers buy Junior7
and not Deep Junior7.
I did not try the game with Junior7 but I will be surprised if I see 0.00 in
that positions even if I replay the moves of the game.
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.