Author: Robert Henry Durrett
Date: 05:26:58 08/17/98
Go up one level in this thread
On August 16, 1998 at 19:34:10, Shaun Graham wrote: <snip> >>I think it is legitimate to use computers in postal chess >>I think a team of computer and human can be better than one of them. >Perhaps it is legitimate, i just see it as a problem in the sense, that it's >hard to say "my computer suggested a move, that i hadn't thought of previously, >i analysed it, thought it good then played it" The problem is that it seems a >little in the grey area to say that the move played wasn't totally generated by >the computer. <snip> The above expresses an overly simplified concept as to how a correspondence chess player utilizes chess engines for correspondence chess. For me, at least, it was not at all that simple! I did compete in a tournament where everybody was allowed to use computers and I used both chess engines and chess database software. These two kinds of chess software complimented each other. With the MCP5 software, I found that it was a waste of time to let the program think too long. It seemed to bog down after an average ply of 7 or 8. To avoid this loss of time, I would make a human judgement call as soon as the program seemed to start bogging down and select the position to be looked at furthur. Then another run from that position would be run. Sometimes this was repeated several times. This gave me the confidence that the move being examined was not improperly evaluated due to horizon effects. Generally, when I received the opponent's move, I had already spent several days in evaluating moves which I felt he/she might possibly make. To furthur complicate the process, for each position I regarded as being "key," I used the idea of creating fantasy positions to try to reach, based on different candidate plans [which I dreamed up by myself] for me, and so looked at as many moves and ideas as the time allowed. I also examined alternative plans for my opponent. [I was averaging better than 70 hours per week on chess, mostly on my correspondence chess games.] The database software was used to find out what others had played in the same position. Sometimes I "latched onto" one of these ideas, if they looked better that the ones I had come up with, and began using my chess engine to look at these ideas, to supplement my prior analyses. So, it really didn't matter that the computer looks deeper if you let it analyze for a long time, since I didn't let it take forever. The player can devote only so much time to "messing with the chess software." "Time is of the essence" in serious correspondence chess. One thing I hated about MCP5 was that it took too long to think. This was partly because I was running it on a 100MHz pentium machine, of course. I hope this helps to clarify the situation regarding how the software is used in correspondence chess, and maybe point to the software needs of the correspondence chess player. Incidentally, playing correspondence chess [in the opening phase] and opening repertoire development activities were mutually complementary, and for me at least, I was doing both at the same time. This caused me to do more work than I might have done had I been just playing chess.
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.