Author: William H Rogers
Date: 14:51:18 01/31/03
Back in the 70's I read an article that compared the differences between how computers examined a chess board and how grand masters did. The noted as we all know that a computer examines every square on the board while grand master recognize positions of power. GM's would see a collection of pieces that were controlling different areas of the board and only examined those areas for further evaluation. I was taught in the begining that control of the center square was one of the most important things that a program should do and today most programs do that to different extents, but that does not locate positions of power as a whole. I, being new to chess programming at time designed a routine that I thought would find these positions of power and try to upset them by attacking them in the strongest way possible. I hand coded the process and then set out to prove that it worked. I took a game between Fisher and Spasky where it was noted that Fisher had made a move that even the masters who were watching did not undrerstand at that time but it was a winning move which revealed later in the game. In the first ply, my simulation found that move above all others as the best, not counting anything else. I did not try to program this into any chess program at that time as I thought it would take to much time to run during a game. I was programming at time in good old 'basic'. Now I am offering an explanation of how it worked whether it turns out to be good or not remains to be seen. The first thing that I did was to create a couple more blank board then I calculated all of the squares that a color could attack including its own pieces. I did not use any xray attacks though. In this way you could see every square that was attacked, but it still did not represent constrations of power. The next thing I did was to average those attacked points on another board by counting all of the points of attack by its surrounding squares and putting the total in the new board. I limited the total to only the eight square surrounding each square in question. The numbers showed constrations of attacking power of the sides. By taking the highest number of the enemies power and taking the lowest score of one of the defending pieces, I not only help to counter the build up of power for the emeny, but help to develop the defending side. Diagrams could point out this better than I am trying to describe it here. If anyone is interested in playing with this idea, I will be glad to go into it in more detail. The question will be whether it adds to the performance of a program or not. I do not know and I still program in 'basic', but you that program in 'C' could evalutate it much better as it would run 10 times faster. I hope that this does not sound crazy but it did work in the few places that I hand coded it and can only hope that it could work even better with newer programs. My email address has changed since I first joind this group and I can not find the place to update the files, so here is my new address rusty819@netzero.net Thanks to all and I hope that I don't sound silly. Bill
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.