Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Wanted: Deep Blue vs. today's top programs recap

Author: Uri Blass

Date: 01:33:31 08/27/01

Go up one level in this thread


On August 26, 2001 at 21:50:03, Joshua Lee wrote:

>
>>I guess that you did not analyze Deeper blue logfiles.
>>I did some analysis of them and my impression based on the analysis that I did
>>is that Deeper blue is only 2 or 3 times faster than Deep Fritz on P800.
>>
>>The idea was to compare the times that Deeper blue got it's main line and the
>>time that the top programs get the same beginning of the main line when you give
>>the top programs some hours.
>>
>>In most cases it is impossible because we do not have similiar main lines but
>>there are some cases when there are similiar main lines.
>>
>>I did not anlayze most of the positions from kasparov-deeper blue match so I am
>>practically sure that you can find more cases of similiar main lines and the
>>question is if other programs are faster or slower in finding the similiar main
>>lines.
>>
>>pavel told me here that he is interested to help.
>>
>>If other people are interested to help in this task then they can send me an
>>email and I am going to tell them positions to analyze with top programs.
>>
>>Uri
>
>
>I am glad to help if i can Uri , for the past couple of months i have been
>running test sets , i would've been done faster except for a disk failure but i
>restarted and have 1000 In one file and Bewteen 500-600 in another as well as
>Test your positional Play so within a month i will be ready to help.
>I do want to mention a position in particular from the 97 match with comments
>from GM Gabriel Schwartzman
>this was in the Summer Issue in 1997 of Chess Life
>
>Game 5 Kasparov - Deep Blue
>Note to Blacks 11th Move
>
>11...h5
>Kasparov's surprise at this move was not hidden. Not only is it a very good
>move, but it is one that a computer would have a hard time making , since it
>depends on pure understanding of the position. As a matter of fact, Kasparov
>tried to see if Fritz or Hiarcs would choose such a move, even if left for
>hours, and they didn't! Once again the Deep Blue team has proven their computer
>shows some signs of intelligence.
>
>
>
>Kasparov,G - Deeper Blue
>r2qk2r/pp3ppp/2p1pn2/4n3/1b6/3P2PP/PPPN1PB1/R1BQK2R b KQkq - 0 1
>
>Analysis by Fritz 6:
>
>11...Bxd2+ 12.Bxd2
>  =  (0.06)   Depth: 1/3   00:00:00
>11...0-0!
>  =  (-0.03)   Depth: 1/7   00:00:00
>11...0-0!
>  ³  (-0.69)   Depth: 1/7   00:00:00
>11...0-0 12.0-0
>  ³  (-0.47)   Depth: 2/5   00:00:00
>11...0-0 12.Qe2 Qd6
>  ³  (-0.50)   Depth: 3/10   00:00:00
>11...0-0 12.Qe2 Qd6 13.0-0
>  ³  (-0.41)   Depth: 4/10   00:00:00  1kN
>11...0-0 12.Qe2 Qd6 13.f4 Ng6
>  ³  (-0.44)   Depth: 5/13   00:00:00  4kN
>11...0-0 12.Qe2 Qd6 13.f4 Ng6 14.c3
>  ³  (-0.31)   Depth: 6/16   00:00:00  22kN
>11...0-0 12.Qe2 Qd6 13.0-0 Bc5 14.Re1 Bd4
>  =  (-0.25)   Depth: 7/19   00:00:00  49kN
>11...0-0 12.Qe2 Qd6 13.0-0 Rfd8 14.Re1 Ng6
>  =  (-0.16)   Depth: 8/20   00:00:00  179kN
>11...0-0 12.Qe2 Qc7 13.0-0 Rfd8 14.Re1 Ng6 15.a3 Bc5
>  =  (-0.09)   Depth: 9/21   00:00:02  622kN
>11...0-0 12.0-0 Bd6 13.d4 Ng6 14.Re1
>  =  (0.00)   Depth: 10/25   00:00:06  2245kN
>11...0-0 12.0-0 Bd6 13.d4 Ng6 14.Nc4 Be7 15.Bg5 h6 16.Bxf6
>  =  (0.00)   Depth: 11/27   00:00:13  4740kN
>11...0-0 12.0-0 Bd6 13.d4 Ng6 14.Nc4 Be7 15.Be3 Qd7 16.Qe2
>  =  (0.09)   Depth: 12/27   00:00:32  11054kN
>11...0-0 12.0-0 Bd6 13.d4 Ng6 14.Nc4 Re8 15.Be3 Nd5 16.Qf3
>  =  (0.09)   Depth: 13/31   00:01:20  28859kN
>11...Ng6!
>  =  (0.06)   Depth: 13/33   00:01:29  32053kN
>11...Ng6 12.Qe2 0-0 13.0-0 Qd7 14.Ne4 Nxe4 15.dxe4 Rfd8 16.c3
>  =  (0.09)   Depth: 14/34   00:02:12  47451kN
>11...Ng6 12.Qe2 0-0 13.0-0 Qd7 14.Ne4 Nxe4 15.dxe4 Qe7 16.c3
>  =  (0.12)   Depth: 15/39   00:03:38  79481kN
>11...0-0!
>  =  (0.09)   Depth: 15/39   00:04:18  94481kN
>11...0-0 12.0-0 Ng6 13.Qe2 Qd7 14.Ne4 Nxe4 15.dxe4 Rfd8 16.a3
>  =  (0.09)   Depth: 16/36   00:06:54  152335kN
>11...0-0 12.0-0 Ng6 13.Qe2 Qd7 14.Rd1 Bc5 15.d4 Be7 16.Nc4
>  =  (0.09)   Depth: 17/36   00:11:56  270993kN
>
>Fritz may find this next ply i will let you know i am sure that it has been
>found at around 15minutes with one of the fritz versions.

Deep fritz plans to play h5 in the next move

Here is Deep Fritz's anlaysi

New position
r2qk2r/pp3ppp/2p1pn2/4n3/1b6/3P2PP/PPPN1PB1/R1BQK2R b KQkq - 0 1

Analysis by Deep Fritz:

1...0-0 2.Qe2 Qd6 3.a3 Bc5
  ³  (-0.31)   Depth: 5/12   00:00:00  4kN
1...0-0 2.Qe2 Qd6 3.a3 Bc5
  ³  (-0.31)   Depth: 5/12   00:00:00  4kN
1...0-0 2.Qe2 Qd6 3.a3 Bc5
  ³  (-0.31)   Depth: 5/12   00:00:00  4kN
1...0-0 2.Qe2 Qd6 3.a3 Bc5
  ³  (-0.31)   Depth: 5/12   00:00:00  4kN
1...0-0 2.Qe2 Qd6 3.a3 Bc5
  ³  (-0.31)   Depth: 5/12   00:00:00  4kN
1...0-0 2.Qe2 Qd6 3.a3 Bc5
  ³  (-0.31)   Depth: 5/12   00:00:00  4kN
1...0-0 2.Qe2 Qd6 3.a3 Bc5
  ³  (-0.31)   Depth: 5/12   00:00:00  4kN
1...0-0 2.Qe2 Qd6 3.a3 Bc5 4.Ne4
  =  (-0.25)   Depth: 6/17   00:00:00  24kN
1...0-0 2.Qe2 Qd6 3.0-0 Rfd8 4.Re1 Ng6
  =  (-0.19)   Depth: 7/19   00:00:00  57kN
1...0-0 2.0-0 Bc5 3.Qe2 Qd6 4.Nb3 Rfd8 5.Nxc5 Qxc5
  =  (-0.09)   Depth: 8/20   00:00:00  221kN
1...0-0 2.Qe2 Qd6 3.f4 Ned7 4.c3 Bc5 5.d4 Bb6
  =  (-0.03)   Depth: 9/23   00:00:01  624kN
1...Qc7!
  =  (-0.06)   Depth: 9/25   00:00:02  1316kN
1...Qc7! 2.Qe2 0-0-0 3.d4 Ned7 4.c3 Bd6 5.Ne4 Rhe8 6.Nxd6+ Qxd6
  =  (-0.09)   Depth: 9/25   00:00:02  1450kN
1...Qc7 2.Qe2 0-0-0 3.a3 Bd6 4.Nb3 Nd5 5.d4 Ng6 6.Bg5
  =  (-0.03)   Depth: 10/25   00:00:03  1828kN
1...Qc7 2.Qe2 0-0-0 3.0-0 h5 4.Ne4 Nxe4 5.Qxe4 Bd6 6.Be3 Kb8 7.Rfe1
  =  (-0.03)   Depth: 11/31   00:00:07  3803kN
1...Qc7 2.Qe2 0-0-0 3.d4 Ng6 4.c3 Be7 5.Ne4 Nd5 6.0-0 h5 7.h4
  =  (0.00)   Depth: 12/31   00:00:14  7875kN
1...Qc7 2.Qe2 0-0-0 3.a3 Be7 4.Ne4 Nxe4 5.Bxe4 f5 6.Bg2 g5 7.f4
  =  (0.03)   Depth: 13/36   00:00:43  25050kN
1...Qc7 2.Qe2 h5 3.c3 Be7 4.d4 Ng6 5.h4 0-0-0 6.Ne4
  =  (0.03)   Depth: 14/39   00:02:53  96516kN
1...Qc7 2.Qe2 h5 3.c3 Be7 4.d4 Ng6 5.h4 0-0-0 6.Ne4 Ng4 7.Ng5
  =  (0.06)   Depth: 15/39   00:04:50  164935kN

(Blass, Tel-aviv 27.08.2001)

Junior7 pIII800 is faster than deeper blue

Here is Junior7's analysis


New position
r2qk2r/pp3ppp/2p1pn2/4n3/1b6/3P2PP/PPPN1PB1/R1BQK2R b KQkq - 0 1

Analysis by Junior 7:

1...c5 2.Bxb7
  ³  (-0.27)   Depth: 3   00:00:00
1...Qd4 2.a3 Ba5 3.0-0
  ³  (-0.56)   Depth: 6   00:00:00  5kN
1...Qd4 2.a3 Ba5 3.0-0
  ³  (-0.56)   Depth: 6   00:00:00  5kN
1...Qd4 2.a3 Ba5 3.0-0
  ³  (-0.56)   Depth: 6   00:00:00  5kN
1...Qd4 2.a3 Ba5 3.0-0
  ³  (-0.56)   Depth: 6   00:00:00  5kN
1...Qd4 2.a3 Ba5 3.0-0
  ³  (-0.56)   Depth: 6   00:00:00  5kN
1...Qd4 2.0-0 Bxd2 3.Qxd2 0-0-0 4.Qf4
  =  (-0.08)   Depth: 9   00:00:00  67kN
1...Qb6 2.a3 Bd6 3.Ne4 Nxe4 4.dxe4 Rd8 5.Qh5
  =  (-0.16)   Depth: 9   00:00:00  137kN
1...Qd6 2.a3 Ba5 3.f4 Ned7 4.b4 Bb6 5.Bb2
  ³  (-0.32)   Depth: 9   00:00:00  539kN
1...Qd6 2.0-0 0-0-0 3.d4 Ng6 4.c3 Ba5
  =  (0.01)   Depth: 12   00:00:02  1450kN
1...0-0 2.0-0 Ng6 3.d4 Ba5 4.Nf3
  =  (-0.08)   Depth: 12   00:00:02  1837kN
1...0-0 2.0-0 Ng6 3.c3 Be7 4.Qe2 Bc5 5.Re1 e5
  =  (0.12)   Depth: 15   00:00:28  17295kN
1...h5 2.Qe2 Bd6 3.d4 Ng6 4.0-0 h4 5.Qf3 Bc7 6.c3 hxg3 7.fxg3
  =  (-0.04)   Depth: 15   00:00:44  26016kN
1...h5 2.Qe2 Ng6 3.c3 Bd6 4.h4 Ng4 5.0-0 0-0 6.f3 N4e5
  =  (0.01)   Depth: 16   00:02:10  71707kN

(Blass, Tel-aviv 27.08.2001)


for comparison Here is Deeper blue's logfile

Guessing Nd2
 7(4) #[O-O](10)##############################################[h5](18)# 18  T=1
Ph7h5 qd1e2 Qd8d6 pa2a3 Bb4a5 o-o Ba5d2n bc1d2B
 8(6) #[h5](-2)##[O-O](0)############################################ 0  T=11
O-O o-o Bb4c5 nd2f3 Ne5f3n bg2f3N
 9(6) #[O-O](-3)#########<ch> 'nd2'
[153 sec (main.c:1849)][cont]###################################### -3  T=39
O-O o-o Bb4d6 nd2b3 Nf6d5 ph3h4 Ph7h6
10(6) #[O-O](-15)######[h5](-5)######################################### -5
T=151
Ph7h5 qd1e2 Ne5g6 pc2c3 Bb4e7 nd2e4 Ph5h4 bc1e3 Qd8a5 pg3g4
11(6)[TIMEOUT] -5  T=178
Ph7h5
---------------------------------------
-->  11. ..   h5 <-- 29/101:33

you can see that deeper blue needed 151 seconds to see 11...h5 and the
main line of Junior at depth 16 is
the same as deeper blue's in the first 4 plies.

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.