ricks@tekcad.UUCP (10/15/83)
#N:tekcad:9700001:000:893 tekcad!franka Oct 14 21:17:00 1983 The articles on the upcoming world computer chess championship have got me thinking. The nature of all chess programs and special purpose hardware therefor is concentrated on the tactical aspects of chess (i.e., if I move there can I pin one of his pieces, can he return the move with a check, etc.). What I was wondering was if anyone was doing work on the planning process in chess (should I engage in a minority attack or go directly for a king-side attack or play a positional game until my opponent makes a mistake). It is pretty well established that chess programs can play a pretty good game of chess tactically, but get hung up on the planning aspects of the game. Has anyone done any work on coupling an AI planning program with a tactical engine for chess and if so, has anyone gotten any useful results? Just wondering, Frank Adrian (tektronix!tekcad!franka)
bstempleton@watmath.UUCP (Brad Templeton) (10/20/83)
It seems that every year some programmer looks at the situation in chess and says, "It's terrible how hardware is all that seems to make the difference. Fast hardware means winning program. Surely somebody can sit down and write a smart program to win." They then proceed to write such a program and run it in tournaments. It usually doesn't do all that well, so in desperation they start taking chess knowledge out of the program and are very disillusioned to find the program now plays better. Someday this trend may stop, but it doesn't seem to have done so yet. -- Brad Templeton - Waterloo, Ont. (519) 886-7304