Author: Uri Blass
Date: 09:04:51 02/02/02
Go up one level in this thread
On February 02, 2002 at 10:36:24, Oliver Roese wrote: >On February 01, 2002 at 20:40:50, Frank Nelson wrote: > >>On February 01, 2002 at 14:06:50, John D. Conte, Jr. wrote: >> >>>I know there must be a web site regarding tips on playing against chess computer >>>engines/programs. I only know from experience to play closed positions, wait til >>>the computer castles; and also to pawn storm the castled side...depending upon >>>the position. Also play positional chess, since a chess program can calculate >>>much faster than a human. I realize a lot depends on which level of play it is >>>on. I would like to print this info/site for my chess club. >>>Thanks. >> >>Gee your idea sounds smart! Keep the position closed and pawn storm the side the >>computer castles on. Usually when I blitz Hiarcs 7.32 I get burned. >>Well it's worth a try I think. If you come up with other strategical ideas for >>anti-computer play please post to this newsgroup. >> >>Ljubomir > >Hi! >Some Hints: >(1)The Golden Rule: >In opening and middlegame keep the position closed under any circumstances! > >This rule is the most important one, if you fail to obey it your moves are >counted. >Check out your opening "before" you use it. >A good pawnstructure to achieve is for example c3-d4-e3-f4. >Look also for point (3) below. > >(1) "Simplify" the position! >"Simplify" is quoted here, because simplified is not simplified for >computers! Look at the branching factor of casual chessengines and you will see >that they have a hard time to ponder even with very few stones on the board. >If you trade two stones on the board, the remaining stones gain usually in >power, so that the branching factor does not decrease by much. Thus the >complexity for computers remains roughly the same. >Not so for the human! Every dead piece simplifies his reasoning. >I knew that a good human (> 2350 ELO) can easily calculate 20 moves (=40 plies) >and more within some seconds in some pawnendings. Recently i posted such a >position here. Such a depth is out of reach for moderate current hardware. >In my games, my few wins were very often caused by reaching an endgame. > >(2) Go for static advantages, don't go for dynamic ones! >This is easily overlooked, since is not feasible in play against humans. >Nevertheless it is absolutely logically: Computers excel with their dynamic >capabilities, humans are abel to see important concepts beyond moves. >Some examples: >Dynamic advantages: >-Superior development, placement of pieces. >-Pair of bishops. >-Attacking chances on some wing. >... >Static advantages: >-Material. >-Better pawnstructure. >-Better pieces (e.g. knight against bad bishop). >... >From that follows, for example: >-Take material, don't sacrifice it. Tell nemeth not to use the trojan horse sacrifice against Fritz. I do not think that it is going to help him to score better. >-Give your pair of bishops to double the opponent pawns, avoid weak pawns. >-Dont attack prematurly. Take your static advantage and hope for the best. >... > >(3) Keep the game short! >This is an obvious concept, since the more moves you make, the higher the >probability to make a blunder in the end. >I mention it here, since it distracts point (1). You have to outweigh >this factors somehow. Good luck. > >(4) Avoid timetrouble! >Timetrouble against machines is not fun and should be avoided. Make sure >to use the right timecontrol. Obey rule (3). 4 is usually good against everyone and there are humans who get into serious time trouble also against humans. > >(5) Dont give up prematurly, if you have lost material! >This is an interesting observation: In general computers are programmed >to keep the position alive, to play out their strenghtes. But not so with >material-advantage! They cooperate with you to simplify the position. >Unfortunately, due to their lack of understanding of most basic endgames, >you are able to hold a draw sometimes. I personally drawed many games >in that manner. Dont give up pramaturaly is even better against humans. Even if you lose material the human opponent may do a tactical mistake when the chances that it is going to happen with computers is smaller. Most top programs know about most practical draws when you are a piece down(like KRN vs KR) even without tablebases. If you are 1 pawn or 2 pawns down it is usually a bad advise to resign against humans and I drew 2 games in the past when I was 2 pawns down even in the endgame(in one case it was rook endgame and in one case it was queen endgame and in the queen endgame my opponent had Fide rating of slightly more than 2000). I remember that kasparov blundered when he had 2 pawns advantage in a rook endgame in one of his match(against short). Kasparov won the game only because of the fact that short also blundered and did not find the draw. It proves that it is a bad idea to resign even against kasparov when you are 2 pawns down. Uri
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.