Author: Bob Durrett
Date: 16:50:21 12/09/02
Go up one level in this thread
On December 09, 2002 at 18:03:52, Matthew White wrote: >On December 08, 2002 at 08:34:16, Bob Durrett wrote: > >> >>I am interested in knowing whether or not modern chess engines recognize >>advantages of these types. >> >>(1) If the White side has a substantial space advantage, but nothing else, will >>the chess engine evaluate the positions with that advantage and give it a high >>score? >> >>(2) If the White side has a substantial time (or development) advantage, but >>nothing else, will the chess engine evaluate the positions with that advantage >>and give it a high score? >> >>(3) If the White side has a substantial pawn structure advantage, but nothing >>else, will the chess engine evaluate the positions with that advantage and give >>it a high score? >> >>(4) If the White side has a substantial material advantage, but nothing else, >>will the chess engine evaluate the positions with that advantage and give it a >>high score? >> >>(5) If the White side has a substantial initiative advantage, but nothing else, >>will the chess engine evaluate the positions with that advantage and give it a >>high score? >> >>Bob D. >One of the most important ways that space advantage is used in an evaluation is >that generally helps a player which side to mount an attack on (more space = >more places to put attackers). A good test for this might be one where the >position is very quiet (i.e. no known tactical shots for many moves deep) and >very well known/analyzed (by humans). > >One thing that I have noticed in GM analyses of computer games is the absence of >brilliant moves (!!). One reason for this is that a GM probably would not give a >computer credit for a tactial shot, since computers are "expected" to see these. >Another reason that I can conceive of is that computers don't seem to do well in >quiet positions. I have watched many computer games where the computer seemed >lost as soon as the tactical possibilities dried up. I haven't often seen a >computer press a space advantage or defend well against a competitor who seems >to have an overwhelming space advantage. > >This (IMO) is how many GM vs. Computer matches are won by GM's: the GM uses a >subtle positional technique to achieve a space advantage which eventually turns >into a pawn storm, or a successful attack. Meanwhile, the computer is >arbitrarily trying to place pieces in an effort to snatch up material. > >The answers to 1,2,3 and 5 are uncertain to me. But I know that for every >program I've ever seen, the answer to 4 is yes, unless the computer has reason >to believe otherwise (i.e. it can grab more material within its horizon, or it >can end the game through a mate or a draw - though not all programs assign a 0 >for a "forced" draw). > >There are three more questions that could also be asked. I'll use your format >for consistency. > >(6) If the White side has substantially better control of the center squares >(d4, e4, d5, e5), but nothing else, will the chess engine evaluate the positions >with that advantage and give it a high score? > >(7) If the White side has the disadvantage of a weak color complex, but nothing >else, will the chess engine evaluate the positions with that disadvantage and >give it a low score? > >(8) If the White side has a piece with a superior position (like an "octopus" - >a knight on a powerful outpost square), but nothing else, will the chess engine >evaluate the positions with that advantage and give it a high score, and avoid >moving/trading that piece? > >Matt Adding to the list in that manner seems a very good idea. If all known significant positional advantages were studied in that way, then a more comprehensive evaluation of the positional chess abilities of a chess engine could be accomplished. What I am finding is that the very high quality GM games annotated by GMs have a MIXTURE of advantages, some positional and some tactical. Every game has positions in which tactics dominate, but some of the positions have the character that the GM analyses talk only about positional considerations for those positions. I call the latter type "positional positions." Even those "positional" positions also seem to contain a mix of different kinds of positional advantages/disadvantages. It may be extremely unlikely that any position exists with only one type of advantage/disadvantage. Bob D.
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.