Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Wouldn't the SSDF get more accurate results if a "DEEP" program .....

Author: Uri Blass

Date: 15:58:25 07/23/01

Go up one level in this thread


On July 23, 2001 at 18:33:22, Joshua Lee wrote:

>On July 22, 2001 at 15:32:08, Uri Blass wrote:
>
>>On July 22, 2001 at 14:46:06, Joshua Lee wrote:
>>
>>>On July 21, 2001 at 15:10:38, Uri Blass wrote:
>>>
>>>>On July 21, 2001 at 14:49:19, Joshua Lee wrote:
>>>>
>>>>>Were is the evidence to support Deep Fritz Being Stronger Than Deep Blue or
>>>>>Deeper Blue?
>>>>
>>>>The evidence is the logfiles of deeper blue.
>>>>A lot of top programs have no problem to see Qe3 in game 2 of the second match
>>>>against kasparov when Deeper blue could not expect this move.
>>>>
>>>>I compared times of Deep Fritz with times of Deeper blue based on the logfile in
>>>>order to see the same beginning of the main line in some cases and I found that
>>>>deeper blue was only slightly faster than Deep Fritz and it means that Deep
>>>>Fritz on better hardware could be faster than Deeper blue.
>>>>
>>>>
>>>>>
>>>>>I would agree if Deep Fritz finds a move that prevents Bxh7+ , which wins for
>>>>>white, then we can talk about Deep Fritz being better.
>>>>
>>>>I remember that Deeper blue did not see that Bxh7+ wins so I do not see the
>>>>point of this problem.
>>>>
>>>>Uri
>>>
>>>The point is is that Deep Blue could prevent this by not allowing it i think the
>>>19th move is a mistake GM seirawan gives it as questionable or dubious ?!
>>>
>>>However i was looking for the sac on move 20 not move 22 where he gives it so 19
>>>....Nge7? is a mistake possibly even losing
>>>
>>>
>>>SO if Deep Fritz picks something better than this THEN and only Then will DF or
>>>any other program would be better than Deep Blue
>>>Because it wouldn't fall for the trap that Deep made possible.
>>
>>One position proves nothing and it is better to investigate the Deeb blue
>>logfiles but  for the record Deep Fritz prefers Be7.
>>
>>New position
>
>[D] 2rr2k1/pp1q1ppp/2n1p1n1/3p4/1bPP3P/1P2RNP1/PB3P2/1BRQ2K1 b - h3 0 1
>
>>
>>Analysis by Deep Fritz:
>>
>>1...dxc4!
>>  ³  (-0.63)   Depth: 1/5   00:00:00
>>1...dxc4! 2.Bxg6
>>  µ  (-0.78)   Depth: 1/5   00:00:00
>>1...dxc4 2.Bxg6 hxg6 3.bxc4
>>  µ  (-0.78)   Depth: 2/6   00:00:00
>>1...dxc4 2.Bxg6 hxg6 3.bxc4
>>  µ  (-0.78)   Depth: 3/18   00:00:00  1kN
>>1...dxc4 2.Bxg6 hxg6 3.bxc4 Qd6
>>  µ  (-0.78)   Depth: 4/18   00:00:00  3kN
>>1...dxc4 2.bxc4 Qd6 3.h5 Nge7 4.Be4
>>  ³  (-0.56)   Depth: 5/21   00:00:00  27kN
>>1...dxc4 2.bxc4 Qc7 3.Bxg6 hxg6 4.Ba3 Bxa3 5.Rxa3 Qd6
>>  ³  (-0.47)   Depth: 6/22   00:00:00  88kN
>>1...dxc4 2.bxc4 Qc7 3.Be4 Qd6 4.h5 Nge7 5.Qe2
>>  ³  (-0.38)   Depth: 7/25   00:00:00  441kN
>>1...dxc4 2.bxc4 Nge7 3.Be4 Nf5 4.Bxf5 exf5 5.a3 Bc5 6.Rb1
>>  ³  (-0.34)   Depth: 8/30   00:00:02  1448kN
>>1...dxc4 2.bxc4 Nge7 3.Be4 Nf5 4.Rd3 Nd6 5.Ne5 Qc7 6.Nxc6
>>  =  (-0.25)   Depth: 9/29   00:00:06  4162kN
>>1...Be7!
>>  ³  (-0.28)   Depth: 9/32   00:00:08  5725kN
>>1...Be7 2.h5 Nf8 3.Qe2 Bf6 4.c5 Qc7 5.Bd3 Nd7 6.Re1
>>  ³  (-0.28)   Depth: 10/32   00:00:10  6751kN
>>1...Be7 2.h5 Nf8 3.h6 gxh6 4.Rd3 Bg5 5.Nxg5 hxg5 6.Qh5 Qe7 7.Re1
>>  ³  (-0.28)   Depth: 11/32   00:00:18  12633kN
>>1...Be7 2.h5 Nf8 3.h6 gxh6 4.cxd5 Qxd5 5.Ne5 Nxe5 6.Rxe5 Rxc1 7.Qxc1
>>  =  (-0.16)   Depth: 12/36   00:00:33  23538kN
>>1...Be7 2.h5 Nf8 3.h6 g6 4.Qd2 Bb4 5.Qd1 dxc4 6.bxc4 Qd6 7.Be4
>>  =  (-0.19)   Depth: 13/37   00:02:21  99146kN
>>
>>(Blass, Tel-aviv 22.07.2001)
>>
>>
>>Uri
>
>You're right about the log files however the only log files i can find are for
>the 1997 match.

This is the interesting match because the interesting comparison is between the
commercial programs and Deeper blue of 1997

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.