Author: Howard Exner
Date: 19:13:12 01/15/00
Go up one level in this thread
On January 15, 2000 at 08:32:10, Albert Silver wrote: >On January 15, 2000 at 05:52:27, Howard Exner wrote: > >>In chess a major challenge is the transition >>between opening and early middle game. Different openings call for different >>approaches once out of book. Will future chess programs consider having modified >>versions of the main engine take over the next 5 - 10 moves out of book >>based on what opening is played (for both white and black pieces)? >> >>Some configurations may play better in Ruy Lopez openings while others perform >>better at Sicilians. Some early middlegames need more king safety >>while other positions should concentrate on solid, non aggressive play. >> >>Humans put on these different hats depending >>on what opening is played. Any future for this in computer programs? > >I believe this is less significant than the transition from middlegame to >endgame which is usually decisive. Knowing when to liquidate to a better ending >where the exploitation of one's advantage is often easier and less risky is a >huge advantage. Determining certain middlegame strategies for specific openings >is a risky enterprise as many of these strategies change and evolve with time, >while others have become a part of general strategic/positional knowledge. It >would probably be risky to impose certain strategic directives even if it was >possible, which I don't see how as just a move may invalidate a certain >positional approach. I envisioned this as a complex undertaking, with elements of risk of course. But there is as you say a body of knowledge that in general describes some characteristics of an opening. The idea would be to eventually revert back to the main engine after say about 5 moves after the opening book ended. If the elements of an opening call for added aggressive parameters then that would be the interim engine. Programmers, opening specialists and we who play over these games see trends in games and ask, "why is the program not following the spirit of the opening now that it has left book". It could be implemented in a basic way with a small number of side engines (closed.engine, semi-open.engine, open.engine ...) or full scale with more side engines being assigned the task of taking over from the main engine. The main engine then takes over after 5 - 10 moves later. Maybe some programs already do this to some degree? Some commercial programs do have opening specialists. They could advise the programmer, "when X comes out of the black side of a Sicillian Yugoslav Dragon then increase the attack and king safety parameters", as one example.
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.