Author: Jonathan Lee
Date: 11:29:47 03/01/01
Go up one level in this thread
On February 28, 2001 at 13:49:35, Jonathan Lee wrote:
>I have found some comparisons of 6 or 7 games when the machine loses at least (4
>GHZ or greater) against a human.
>I hope this becomes an inspiration to the Computer Chess Club.
>
>1. There are no pawn or piece captures in the first ten moves. I guess that
>will be revised precise definition of a closed game.
>2. Another critical factor is the opening. With the exception of the Stonewall
>Defense there are no common denominators in the closed openings. The Stonewall
>happened only twice.
> My observation is to find all opening libraries that have closed games and
>look more closely. (Preventing closed positions) or (the fewest number of
>closed openings) as much as possible may be a high priority.
>3. With the exception of one game the game ends in less than 50 moves. The
>computer cannot tackle the complex middle game, thus losing material. The
>computer tends to understand end games, thus drawing or winning games.
> The exception is Shirov versus Fritz on June 22, 2000; the game lasted 59
>moves. The last 10 to 15 moves were lightning quick coming from Shirov.
>4. With the exception of one game the original queens are used. (This excludes
>pawn promotion turned into queens. What I mean is the original queens are a
>much greater factor than passed pawns.)
> The exception is Kasparov versus Deeper Blue game one in 1997.
> Another possible exception: Leko versus Fritz had one passed pawn which was
>something like a mate in 8.
>Jonathan (101 message)
5. Of those 7 losses humans were white 5 times. Of those 5 losses, 2 of them
were the Stonewall Defense.
Obviously, the other 2 games is when human was black.
6. Of those 7 losses it was Fritz 4 times, Junior 2 times, and Deeper Blue
once.
Of those 7 wins, one from Kasparov (Deeper Blue), two from Kramnik (Junior
and Fritz), one from Leko (Fritz), one from Shirov (Fritz), one from Piket
(Junior), and one from Alterman (Fritz).
If you know since 1997, any games with a parallel processor (4 GHZ or greater)
not mentioned that resulted with losses please show in this thread.
Jonathan (102 message)
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.