Author: Alastair Scott
Date: 04:37:36 10/16/02
Go up one level in this thread
On October 16, 2002 at 05:44:04, Uri Blass wrote: >I think that it is better to do a sweningen tournament of the top humans against >the top programs to test the level of programs when the prize of the winners is >going to be a similiar match to the match kasparov-Junior when players also get >money that is proportional to their points. > >I think that in that case people are going to be less suspecious about things. > >People can suspect that one player is a cheater but they will not believe that >many of the top players are cheaters so if programs can get surprising good >results in these conditions people are going to say nothing. > >The difference relative to previous events is that the humans should be >motivated to win like smirin was in his games against computers and the games >should be 120/40+60/20 time control. > >In previous tournaments the result is not a surprise because the humans were not >prepared to play anti computer and had no big motivation of something like 1000$ >for win and 500$ for draw against the computer. > >In the kramnik case it is a surprise because I expect even IM's to have good >chances to beat Fritz in similiar conditions and I expect kramnik to take no >risks when he is leading 3-2 because 3-3 is a dangerous situation and suddenly >the win does not seem sure because one mistake is enough for Kramnik not to win >the match. Yes, this is a good set of suggestions - the artificiality of the current situation, with the world stopping for a set-piece match, doesn't help, and something based on the Swiss or Scheveningen systems (to add a bit of unpredictability to the proceedings) would quieten down the conspiracy theorists. The remaining problem is the integrity of the computers. Before each (UK) National Lottery draw there is a fuss made of the set of balls to be used and the machine to be used to pick them (all of which have names). All a bit of theatre really, but it gives a reassuring feeling. Something like this needs to be done to convince everyone that all the computers are identical, all the software is production versions etc. etc. (I feel that anything which allows custom or programming-team-supplied hardware is probably too open to suspicion). Alastair
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.