Author: Alex Shalamanov
Date: 02:02:58 09/08/05
Go up one level in this thread
On September 08, 2005 at 04:07:53, Odd Gunnar Malin wrote: > >I don't play corr. chess but the work is very similar to the analysis work after >a game. > >I do 'in dept' analyses of the moves where my opponent got me 'out of book' if I >can't find any other sources for suggestions. Here I have notice that it is very >often the engines suggestion of move is not what I end up with as 'best' move, I >rather convince the engine that my suggestion is the best. > >To manage to convince the engine that your analysis is better there have to be >feature in the engine that can handle this. My two favourite here are Shredder >and Gandalf, they often gives me different view of a possition and can handle >that when you go deep into a position and then backup they remember their >analyses and thereby change their view of bestmove at the root position. > >This metode of analysis is tried to make automatic in Chess Assistant but my >experience is that you get better results by manualy analyse the position and >just have the engine as a follower, plus that you keep the creative part of your >carbon-computer active. > >Maybe more engine-authors should be thinking in this lines, it is also important >that it manage to 'keep' some of its analyses the next time you look at the >position, even more important for a coor. player. > >Odd Gunnar It happens very often that I have to "correct" the chess engines I use (mostly, Shredder 9UCI or Fritz 8) when I don't quite agree with the evaluation or the move proposed leads to game positions I don't like. Instead, I think of more human-like moves that agrees with my chess experience and knowledge and start analysing it thoroughly. Sometimes the chess engine changes its evaluation and starts rating the move higher. It's usually the case with the strategic game positions. Sometimes the games I play in CC become unfinished because the tourney comes to an end and then they would be awarded by the tournament committee, so I have to send my analyses as a proof of a win or a draw, so I have to analyse the positions in-depth. Fritz 8 has a very convenient deep analysis tool and it takes less time than you do it yourself, but, undoubtedly, one has to check the Fritz's analysis output thoroughly. Frankly speaking,all correspondence chess players use chess programmes and chess databases but a human factor is, nevertheless, very big, especially, with the high-rated ICCF players or GMs. I've only listed those options that would be beneficial while playing correspondence chess. Regards, Alex
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.