Author: Jorge Pichard
Date: 05:20:53 06/05/04
Go up one level in this thread
On June 05, 2004 at 06:40:48, Tord Romstad wrote: >On June 04, 2004 at 18:39:13, Jorge Pichard wrote: > >>Since computer can hold and remember more Opening than any Human and they are at >>the level of the very best human players such as Kasparov, Anand and Kramnik' >>the need for Fischer Random Chess will become more popular in the next 5 years. > >I don't see why the computer's perfect memory would contribute to the popularity >of FRC. Of course humans and computers have different skills. The computer is >obviously superior at remembering concrete and exact information, and at >calculating >quickly. The human is superior at pattern recognition and long-range planning. Frederic Friedel is one of the key mem at Chessbase, and produces a sucessful magazine in Germany, (Computer Schac und Spiele = Computer Chess and Games). He put forward an interesting view of the computers versus humans battle. His feeling is that in certain percentage of chess positions that are liable to arise in practice - maybe 20%- the computer is far stronger than the best humans, and will win practically the whole time. Equally, there is a percentage of positions where the computers will stand no chance against the top players- also perhaps 20% at the moment. While the game is in the 60% no-man's-land in the middle, things are very finely balanced, and the human's task is to reach the good 20% rather than stumble into the bad 20%. The computer, of course, is oblivious to this struggle, though the programmers may try to bias it toward playing human-hostile chess. Frederic's view, therefore, is that while these percentages will become worse for the humans as computers get fasters, there will still be scope for humans to steer the game into the, say, 5% of positions where the computer has no chance, while avoiding the 40% where the computer rules supreme. At some point, though, this will cease to be possible, as the slightest, most imperceptible inaccuracy (in the sense of allowing a computer-friendly position) will throw the player into the abyss. Jorge >There are some games where the computer's strengths are the more important >(like othello), some where the human's skills are more important (like go), and >some which are somewhere in between (like chess). Why does the fact that >computer >players are competitive in chess make the game less attractive for humans? > >And by the way, I don't think FRC is any more difficult to play for computers >than >classical chess. If some of the top programmers spent some time implementing >FRC, the top engines would be just as competitive there as in classical chess. > >I personally find FRC to be one of the least interesting chess variants I have >ever >seen. If you want to abandon classical chess, why not switch to some of the >many more complicated chess variants which really add something new to the >game? There are lots of such variants, including Chess with Different Armies, >shogi, hexagonal chess and Gothic chess. And unlike FRC, all of these chess >variants really *are* more difficult for computers than classical chess. > >>Even a player such as former world champion Garry Kasparov who has incredible >>memorization capabilities, complained that he could not always remember his >>opening preparation. Therefore, it will become justifiable to match the very >>best human against the very vest FRC program. > >Neither Kasparov nor Kramnik would be very interested in such a match, I >think. Leko would probably be willing to play, though. > >>Probably very soon Shredder and Hiarcs will also be available in FRC. > >Why do you think so? There is currently no market demand for a professional >FRC engine. Right now, there are several hundred engines which play classical >chess, and less than ten which play FRC. I happen to be one of the few engine >authors which have written engines for both games. Every week, I get about >50 e-mails from users with feedback about my classical chess engine. I get >almost no feedback at all about the FRC engine. Richard Pijl and Volker Anuss, >who have also written FRC engines, have been kind enough to play a few >games and send them to me, and you played a few games which you posted >here, but that's all I have received so far. It took more than a week after the >release of my FRC engine before anybody could even confirm that it worked >(I couldn't test it, because I don't run Windows). > >The truth is that there is almost zero interest in FRC. From a commercial >point of view, adding FRC support to Shredder or Hiarcs would be a complete >waste of time. > >Tord
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.