Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Interesting opening statistics for man-machine competitions

Author: Vasik Rajlich

Date: 02:28:33 06/26/05

Go up one level in this thread


On June 25, 2005 at 14:47:32, Robin Smith wrote:

>In a previous thread some people have been repeatedly stating that choosing
>1...e5 as a human’s defense against 1.e5 is a very poor choice. They said
>practice has repeatedly shown this. I posted some games to the contrary and was
>accused of "cherry picking" (which was true enough). So I decided to go back and
>look at all the recent man-machine matches between strong humans and strong
>computers. The results were a surprise, even for me. The matches I looked at
>were:
>
>Deep Blue-Kasparov 1996, Deep Blue-Kasparov 1997, Deep Fritz-Kramnik 2002,
>Rebel-van Wely 2002, X3D Fritz-Kasparov 2003, Hiarcs-Bareev 2003 and
>Junior-Kasparov 2003
>
>I may have missed some matches, but if I did it was not my intention or cherry
>picking. I deliberately excluded Hydra-Adams, since the discussions have been
>about Adam's supposedly poor decision making process in choosing 1...e5 as his
>defense of choice, and this decision had to be made prior to the match. I also
>only looked at games where the computer played white, since that is what the
>ruckus has all been about.
>
>The results from the above matches were:
>
>1.e4 e5    white (computer) scored 50% in 5 games
>1.e4 other white (computer) scored 68% in 11 games
>not 1.e4   white (computer) scored 67% in 3 games
>
>If I were Adams, and 1...e5 was the mainstay of my opening repertior, I would
>not see anything in these statistics to change my mind. Perhaps people who have
>been vociferous about Adams "poor opening choice" may have been coloring their
>thinking based on Crafty blitz games on ICC. These are hardly the games upon
>which to make ones opening choices for a serious tournament time control match
>with money on the line.
>
>Below is a table of results for the games. Scores are from the white (computer)
>perspective.
>
>               opening:	1.e4 e5		  1.e4 other	   not 1.e4
>	event		score  # games	  score	 # games  score	# games
>Deep Blue-Kasparov 1996	0       1	  1.5	 2	   0	 0
>Deep Blue-Kasparov 1997	1	1	  1.5	 2	   0	 0
>Deep Fritz-Kramnik 2002	0.5	2	  0	 0	   1.5	 2
>Rebel-van Wely 2002	0	0	  2	 2	   0	 0
>X3D Fritz-Kasparov 2003	1	1	  0	 0	   0.5	 1
>Hiarcs-Bareev 2003	0	0	  1	 2	   0	 0
>Junior-Kasparov 2003	0	0	  1.5	 3	   0	 0
>	total		2.5	5	  7.5	 11	   2	 3
>	score		50%		  68%		   67%
>
>-Robin

Interesting. (Although obviously it's hard to make any conclusions.)

I think the criticism of Adams' openings is based on two factors:

1) People just don't understand how much a player like Adams knows about chess.
(And yes, chess software.) It's the armchair quarterback syndrome.

2) When you lose, everything you do will be criticized. This is not just in
chess. Often you'll see a game annotated where the loser in a bad position
played move X. The commentator will say: "X is not the best, the loser should
have tried Y". If the loser had played Y, the commentator would be saying: "Y is
not the best, he should have played X". The only way to avoid this criticism is
to not lose.

Vas



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.