Author: Uri Blass
Date: 04:48:16 06/15/02
Go up one level in this thread
On June 14, 2002 at 11:01:08, Kurt Utzinger wrote: >On June 14, 2002 at 10:28:24, stuart taylor wrote: > >>On June 14, 2002 at 07:42:21, Kurt Utzinger wrote: >> >>>I can only say once again that I have seen match results of 9-1 followed by 1-9 >>>and that for this reason quite a large number of games are needed to say for >>>sure that program A is stronger than program B. >>>Kurt >> >>It is a bit interesting why it goes in blocks. >>Maybe when Fisher beat Larsen and Taimanov 6-0 each, maybe Larsen too, that >>didn't mean ANYTHING in the world either? >>S.Taylor > >There is perhaps a big difference between human vs human and computer vs >computer matches. A player of high top level will hardly ever lose 3 or 4 games >over a match of 20 games against a player having 200-300 ELO less. Such things >however often happen in computer matches. >Kurt I disagree about it. 1)I remember that sofia polgar won a tournament in Rome with 8.5 out of 9 when she scored 6.5 out of 7 against average rating of more than 2500. Her rating at that time was less than 2400 and the results were only a statistical noise(She could not reproduce similiar results later). I believe that computers play more often than humans and this is the main reason that you can see more often surprises in comp-comp games. 2)I believe that 200-300 elo difference in computer chess means usually that the best player win more relative to the case with humans. one of the things that I try to improve in Movei is to teach it to have better chances against strong players and I am going to be happy if it can get 4 out of 20 against programs that are 200-300 elo better(My tests show that list4.61 beated it 33.5-6.5 at 5 minutes/40 moves when movei tried every possibility in the first move but at least movei could win 4 games and I posted one of the wins of movei by 1.e4 b5 in the winboard forum). You can also look at the 5th division results and find that movei beated all the programs that were weaker than Nonnian. Other programs that promoted also won almost all of the weak opponents. You can learn from it that there are not a lot of surprises when the difference in strength is big. Uri
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.