Author: Bob Durrett
Date: 08:06:22 10/31/02
Most engine design is still focussed on making the engines perform at a higher performance rating. That is good. Strong engines are essential for accurate post-mortem chess analysis. But there is a marketing motive for chess programmers who wish to see their engines widely used by chessplayers everywhere. Some people actually do play chess against their chess programs. Those who don't probably dislike getting slaughtered every game. Those who use "dumbed down" engines probably don't care for the non-human nature of such dumbed-down engines. A chess engine programmer would feel really good if he/she knew that many people were using his/her engine. Especially if there was lots of positive feedback from the users. Chessplayers vary greatly in their playing strength from absolute beginners to the top GMs. Most are at the amateur level. Nothing new here. It follows logically, however, that engines intended for training purposes should take into account the wide range of playing strengths of the engine users. Hypothesis: A training engine suitable for a GM would be unsuitable for a chess amateur and visa versa. One idea for a training engine would be for it to provide ample positive feedback [verbally, preferably via a speaker] for good moves and negative feedback for bad moves. But the "good moves" of a chess amateur might be regarded as "bad moves" if played in a GM game. Similarly, the "good moves" of a GM might not be seen at all by the chess amateur. If a chess amateur is playing against the training engine, then the engine should provide positive feedback for good moves which would be fairly hard for an amateur to find. Essentially, the engine needs to know the approximate playing strength of the engine's user and adjust the feedback comments to be appropriate for that chessplayer. As a refinement, one might make adjustments for the sex of the chessplayer. Similarly, one might tone down the negative comments and strengthen the positive comments for especially sensitive or depressed chessplayers. The use of psychology in a talking chess engine is in it's infancy. There is plenty of room for innovation here. 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.