Author: Paul Richards
Date: 15:16:38 06/04/99
Go up one level in this thread
On June 03, 1999 at 18:23:28, Stephen Ham wrote: > >time available once their adversary's opening lines were completed, yet the CM >programs scored highly. Am I missing something here or is it not logical to >assume that the CM programs are superior performing programs than Junior, >Hiarcs, Fritz, MCP, etc? Finally, since program opening books confer an >advantage/handicap depending upon their completeness/accuracy, would not the >best test of software strength be to either have programs compete without their >opening books or from the same middlegame position? Please let me know your >thoughts and whether such a tournament has been done. Shep, does this sound >like something you'd be willing to try? Hello Stephen, I'm going to rephrase the subject here. This may not be exactly what you are asking, but this concept is something I have been mulling over as well. If you take a number of engines and have them analyze the starting position, most will choose 1.e4 after a time. Some will then suggest 1...e5, others the Scandinavian with 1...d5, some the French, etc. As you progress this way you will clearly see that some programs are better than others at following chess theory and playing main line variations on their own. Others get into off the wall ideas and miss moves that are firmly established as best. Then sometimes if you let them think to 16 ply or so they will see why such a move is best. Sometimes the move is a bit too deep. But in any case it seems logical that a program that has a thorough positional understanding of the game should be able to wind its way through the opening, and by following basic principles such as not losing a tempo moving the same piece twice, trying to control open files, not creating pawn islands unnecessarily or what have you, come up with moves that fit well with established opening theory. Fritz 5.32 seems quite good at this, and it will typically play a Spanish or Queen's Gambit line without ever loading the opening book. Some programs are not as good (I would like to try this with Hiarcs but don't own it yet). There seems to be a correlation, and it seems like there should be a correlation. *shrug*
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.