Author: Howard Exner
Date: 08:34:17 11/14/00
Go up one level in this thread
On November 14, 2000 at 04:49:02, Ricardo Gibert wrote: >On November 13, 2000 at 14:35:50, James T. Walker wrote: > >>Seems the GM/IM players can not hold the computers at Game/hour an more. Deep >>Junior & Deep Fritz both won their game today. I took Jame's comment as referring to the sporting competition we are witnessing between Deep Fritz and Junior. So far it's lopsided and like a sporting event we wish to see a closer match, therefore the time boost for the human team, mostly GMs. Average rating of the human team (2450 -2550 ?) > > >I would not be so sure. The trouble with that assessment is the players involved >do not know how to play effectvely against computers. Some have disclosed that they are very familiar with computers, while others not. Still, I'm waiting for that time when the "anti computer play" will be effectively and consistently demonstrated in these public Human/Computer events. Sure we've witnessed games where a computer makes terrible moves, trapping their pieces or mis reading an endgame, but how frequent are these? And given the tally of computer shortcomings they must have compensating strengths to play so well against strong opposition. This makes them more of a volatile player as your examples below indicate. >I would not be surprised >if there are savvy players below master level that can beat "strong" computer >programs at game/60. Take for example the following player on ICC: > > > >Information about rabid (Last disconnected Mon Nov 13 2000 12:28): > > rating [need] win loss draw total best >Wild 1717 [6] 12 27 0 39 1740 (13-Dec-1997) >Bullet 2722 [8] 232 2564 188 2984 2739 (06-Oct-2000) >Blitz 2346 336 3159 265 3760 2760 (05-Aug-2000) >Standard 2157 [6] 10 49 6 65 2231 (28-Feb-1998) > > 1: I like to play computers especially crafty clones > 2: This inflates my rating > > > >As you can see, this guy specializes against computers and has been able to pump >up his rating to over 2700 at bullet! I believe his true rating is 2100-2200. Or >how about this guy: > > > >Statistics for MLBernstein On for: 2:30 Idle: 3 >MLBernstein is observing game 175. > > rating [need] win loss draw total best >Wild 1838 [6] 59 12 12 83 1851 (22-Jan-1999) >Blitz 2609 [8] 143 168 120 431 2609 (08-Aug-1999) >Standard 2361 [6] 34 97 82 213 2407 (24-Jul-1998) > > 1: Marc L. Bernstein from Laguna Beach in Southern California > 2: USCF 2236 as of November 2000. Currently inactive in tournament play. > 3: Normally I don't play on the ICC without at least a 10 second increment. > 4: I am currently employed by ACADEMIC CHESS, an organization that teaches >chess to elementary school children and occasionally to middle school and high >school kids. > 5: I enjoy opening theory and like to invent novelties. I also like endgames. >My middlegame play could use some improvement. > 6: I am an experienced teacher of chess, mathematics and to a certain extent >the physical sciences. > 7: My ICC blitz rating is much higher than my USCF due to my ability to draw >against high-rated computers. > > >I happen to know this guy. He is actually a fairly ordinary Blitz player for his >rating of 2236. His distinguishing feature as a player is he is an opening >theoretician. He knows more about some openings than many professional players. Human's too have stronger aspects to their game as this person's opening knowledge shows. > >If these guys can pump up their ratings by 400-500 ratings points, then I think >the strength of computer programs are something of an illusion. Programs seem to >have terrible weaknesses that can ruthlessly exploited by only moderately strong >players that know what they are doing. Agreed but 10-4 is no illusion > >The problem with computer programs is they cannot adapt to their opponent like >humans can. Apparently, adaptability is an important component of playing >strength. I agree on all your points about computer weaknesses and how they can make foolish choices. But why are they consistently doing so well in public showdowns with strong opposition? Like an anticipated football game, if at half time the score is a blowout, I'm gone or at least searching for a closer contest. "40 and 2, 40 and 2 ... !"
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.