Author: ALI MIRAFZALI
Date: 14:28:19 01/25/02
Go up one level in this thread
On January 25, 2002 at 17:07:59, Uri Blass wrote: >On January 25, 2002 at 16:52:38, Dann Corbit wrote: > >>On January 25, 2002 at 16:46:38, ALI MIRAFZALI wrote: >> >>>I would like to know the opinion of the readers of this forum on the following >>>questions. >>>1.What would be the Elo rating of the perfect chessplayer? >> >>Depends on whether or not he's playing against another perfect player. If he >>wins 100% of his games, his ELO is infinite. If he plays another perfect player >>and draws 100%, then his ELO is anywhere from 1 to any arbitrary huge number >>(same as his opponent) depending on how you seeded the calculation. > >It is possible to ask the question in another way >suppose that a computer can solve chess by search and has an opening book of >1000000 correct moves > >Suppose we do a tournament when the same program play at different depthes and >decide that the depth that the program draws with kasparov gets rating of 2800. > >What is the rating of the perfect player? > >Note that I expect to see the perfect player gets something between 50% and 100% >against depth of 30 plies because the question if depth 30 plies play perfect >chess is dependent in the opening. > > >> >>>2.Are there natural limits to the strenght that can be achieved in chess >>>for a computerchess player? (Not the present centuary but in any future;that >>>is what I mean by natural limits). >> >>If you calculate to the end of the game, you cannot improve on that. I don't >>think we'll see that happen real soon. >> >>>3.If the rating of perfect player is say x ;what would be the rating of >>>the stongest computer player ever(that is the best chessprogram that can be >>>ever contructed useing computer technology) .It would be x-?.Or would it be x? >> >>It would be zero, unless it was perfect also. The perfect player would win >>every game and get all the ELO points. The imperfect player would lose all the >>games and get an ELO of zero. > >No > >It is possible that the imperfect player is going to draw part of the games. > >Saying that someone is not perfect does not mean that it is the case in every >game. > >Even the player who play random moves is going to get more than 0% against the >perfect player if you play enough games. > >Uri Interesting remarks by Uri.When Elo himself devised the scale he proposed 3300 USCF as the rating of the perfect chess player.Actually in 1997 Deep Blue searching to a general depth of 12 plys drew Kasparov 3 times.A 30 ply search will not make much difference because most GM loses to each other and to Computer programs are due to tactical errors.A very strong human player such as Kasparov or Kramnik should draw many games against the perfect player but will Ultimely lose a match because of minor inaccuracies.
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.