Author: Robert Henry Durrett
Date: 04:47:01 08/10/98
Go up one level in this thread
On August 10, 1998 at 04:46:43, Steffen Jakob wrote: >Hi all! > >I would like to tell you something about an idea which I had this >weekend. I talked a bit about it with Bruce at ICC and he said that it >might already be a known idea. I would appreciate some feedback. > >Some time ago someone (KK?) posted statistics about the first moves of >white. This seemed rather useless for me. More interesting would be >the the likelihood of _any_ move played in a won game by a strong >player. This information e.g. could be used for a trivial move >ordering. My plan is to do the following: > >1. Use a precalculated list of all possible moves for the move > generator in your chess engine. >2. Add a counter to the move structure >3. scan a large database with only high quality games and > increment the counter for each game where the winner made this > move. > >The counter could be used to perform a very simple but also cheap move >ordering or even for the evaluation itself. This idea could be improved >if you don't store only a counter but a set of (pawnstruct, counter) >pairs. > >What do you think? > >Greetings, >Steffen. This reminds me of an observation I made several years ago when I was checking out the lines in GM Soltis's book "Winning with the English Opening" 2nd Edition, using MCP (I can't remember MCP's version number). In a typical run, I would key in the first move [for White] in Soltis's line and then let MCP select it's best move for Black. Then I would let MCP find it's best move for White, but NOT key it in. Instead, I would look at Soltis's book to see what Soltis recommended [for White] in the new position, and key it in [instead of what MCP wanted to play for White]. I did this for all lines in his book. It took awhile! The observation was that in some cases, MCP "held Soltis's move in contempt" and gave his moves a low rating compared to MCP's preferred move. However, and this was the surprising part, after a few moves in the line, MCP started rating the resulting position for White as being better than any position which MCP had found previously in that line. In other words, "GM Soltis was smarter than MCP." Another observation was that, in these cases, the position at the end of Soltis's line would never have occurred if MCP had been playing against itself. The relevance of the above to this thread is: Maybe Steffen Jakob's idea would help to prevent this "MCP blindness" and would prevent similar situations in chess engines using his ideas. I guess, from a computer user's point of view, producing a chess engine which took existing theory into account would result in less "blindness" and hence better moves.
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.