Author: stuart taylor
Date: 20:23:41 01/13/00
Go up one level in this thread
On January 13, 2000 at 06:19:55, Tom Kerrigan wrote: >On January 12, 2000 at 01:38:14, stuart taylor wrote: > >> But the greatest players have to make the greatest moves, and when the >>machine is already programed with almost any possible 20 greatest first >>moves in any game, that's already half the game in the machines pocket. > >No. Nobody knows what the greatest moves are. > >I have seen a million posts here saying, "this game was lost due to a bad line >in program X's opening book." If you go to a computer chess tournament, you hear >this complaint frequently. > >The fact is, you don't simply download a few gigabytes of PGN files and all of a >sudden your program plays openings perfectly. It is very difficult to keep your >program from playing the s****y moves. It is very difficult to make your program >play openings that it is good at. It is very difficult to keep people from >exploiting holes in your opening book and opening book algorithms. > >IMHO, the opening book is a necessary evil. It is a crutch to keep the program >from playing the same moves, and therefore losing every single game it plays. >None of the programmers I know like working on their opening books. None of them >get really excited about terrific new opening book algorithms. Basically, it's a >feature that we would not have unless it's necessary. > >-Tom O.K. But surely it gives a great advantage to machine, unless human spends his life trying to find holes in all known theory. And there are less and less holes left to find, and when they are found, they will be corrected very quickly in databases, whereas poor human has unending burdens which are hardley to do with chess-but a regrettable necesity-esp. for the human. S.Taylor
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.