Author: Chris Carson
Date: 04:01:27 09/21/02
Go up one level in this thread
On September 20, 2002 at 23:12:35, martin fierz wrote: >On September 20, 2002 at 18:44:59, Chris Carson wrote: > >>On September 20, 2002 at 16:52:46, martin fierz wrote: >> >>> >>>it was an even result, but not an even match. there's a big difference, and as >>>long as you don't look at the games, you will never see it... >>> >>> >> >>I looked at the games and the results, however, what is your assessment? > >van wely self-destructed in game 1. to attempt to win an ending where you don't >really have winning chances and have very little time on your clock against a >computer is pure suicide (against a fellow human, why not - he is bound to make >mistakes too...). when you see a human do that kind of stuff, you know he is not >*really* prepared for a computer match, no matter how many training games he >played. i guess he never played really serious games, with time trouble and all, >else he would certainly have realized how dangerous this winning attempt was, >because he would have lost some of his training games in this manner. > >the other 3 games seem fair to me. if you ask me what i see in these 4 games, it >is that van wely should have won the match IF he had been in the right "frame of >mind" to play against a computer. > >you might say that van wely just blundered, as humans do, and that my argument >is wrong. but my argument is that van wely "blundered" long before his real >chess blunder by not taking an easy draw in the position with the 2B-Q by doing >nothing (i think he even declined a draw offer?!) - instead he went straight >into a situation which favors the machine: little time & tactics. and it's only >natural that he loses the game in this situation. > >my belief is that man-machine matches are all about who can force who to play on >his territory. which is IMO why white has had such a high winning %-age in >recent computer matches: van wely (100%! 4 games), gulko (75%! 8 games), smirin >(62.5% 8 games), same pattern in all, white is doing much better on average than >in "normal" computer-computer or human-human competition (i think about 55% is >normal). the extra tempo allows humans to play cautious setups as white, and >stay clear from tactics, while their attempts to do the same as black have been >unsuccessful to put it mildly :-) > >aloha > martin Nice analysis, to sum it up, "To Err is humam, to really mess things up requires a computer". ;)
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.