Author: martin fierz
Date: 06:14:51 04/30/04
Go up one level in this thread
On April 29, 2004 at 15:20:03, Kurt Utzinger wrote: >On April 29, 2004 at 15:04:11, Pablo Rodriguez wrote: > >>On April 29, 2004 at 14:47:16, Kurt Utzinger wrote: >> >>>The so called "boring" playing style can be applied by following the rules >>>stated below, but I'm warning you. Such games are indeed very very boring and >>>not only strong players may ruine their style: >>> >>>1) never play for a win, even if the own position is slightly better >>>2) avoid time trouble (most important) >>>3) style: quiet, cautious, prophylactical (very important) but not passive >>> (important) >>>4) positional: proper pawn structure, control open lines, try to get statically >>> balanced positions and avoid dynamically balanced positions, avoid whenever >>> possible pawn majorities >> >> If you play a few games versus junior 8 or djunior 8.zx you know this is >>clearly impossible since the program always goes for unbalanced positions. >> >>>5) trade (all) pieces - in particular the queens - whenever this is possible >>> without giving the slightest advantage to the opponent >> >> I agree, this strategy has being used in the last years by strong gms versus >>computers. Super GM Vladimir Kramnik used it several times in his match vs fritz >>in 2002, and was a full success. >> >> >>>5) avoid tactics and unclear lines >> >> Avoid unclear lines: this is difficult, but with proper training in the opening >>i don't see why this isn't working. >> >> Avoid tactics: you just can't avoid tactics, but you can avoid heavy tactics >>which is quite different. First time Kramnik went into tactics vs fritz was >>demolished by the computer when he sacrificed a knight for compensation. He >>couldn't find it, and which is more lost because fritz defended perfectly. >>However, Kramnik showed even world champions can make miserable failures or >>blunders, and hunged a knight in game 6. >> >>>7) go for endgames with bishops of opposite colours (in particular with >>> a pawn down) >>>6) choose suitable playing level [at least 120'/60, better 120'/40] and >>> avoid blitz games >> >> Of course humans don't have a chance vs strong computers at blitz. As i read >>in certain interview, Garry Kasparov was able to score near 50% vs Junior 7 on >>5' + 2'', but since this program is old i think he wouldn't score more than 30% >>vs ssdf nÂș1 shredder 8, and remember he is one of the best blitz players in the >>world, if not the best. >>> >>>On the other hand I can assure you that even 1900-2000 Elo rated human players >>>will be able to get a score of 25 % to 30 % against all top engines. >> >> No way! a 2000 elo would not do more than 10% against a computer, even at >>classical time controls. Of course, this is my opinion... >> >> Pablo > > Hi Paplo > You very much overestimate the real strength of computer programs > if humans are fighting this way. Please also have a look at some > games at http://www.utzingerkurt.com/man_machine.htm > I would even say that to get 20 % is easier vs the top five engines > with their fine books than vs a strong amateur with a less good book!! > And as to Kramnik: he did not play the boring playing style as he was > always looking for a chance to get the full point and this needs a lot > of power and makes humans tired. > Kurt hi kurt, i looked at some of the games - and i think that a simple change of opening book and perhaps some evaluation parameters would put an end to this. when i play a game against a 1600 or a 2400 player i also make different decisions based on the opponent - e.g. if i see i can force a drawn endgame in an otherwise unclear position, i will do it against a 2400, but not against a 1600, hoping to outplay him in the more complicated position. against the 2400 player the roles are reversed, and i'm happy to exchange into the drawn endgame. your top programs have been tuned to perform well against other top programs. it's a shame that they can't decide to change their style to beat weaker players who want to bore them to death. i think they would easily beat you if they could do this. cheers martin
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.