Author: Bob Durrett
Date: 16:37:17 12/09/02
Go up one level in this thread
On December 09, 2002 at 17:19:19, Mike S. wrote: <snip> >To make it short, a perfect test position is where you only need to look at the >first move and bingo! that's it. > >In suboptimal and bad positions, which lack the (c)+(d) requirements, you'll >have to look at the evaluation, main variant etc. too, or even see how it would >acutally continue to play (if the engine really plays like inteded...) IOW, you >have to create an in-depth expert witness'es opinion for each engine's result on >each position. This is very uncomfortable and will probably take so much time, >you won't be through before the program ist outdated. :o) It is better to have a >suite which can produce reliable results quickly. > >I completely quit searching for good *positional* tests some time ago, and >focuss on tactics mainly. I am saddened, but not surprised, to hear that. I am finding it extremely difficult to find "positional" test positions that meet your criteria. In fact, I have found none so far. : ( >Some endgame things (í.e. specific knowledge cases) >should also be fairly easy to test in that way, too. > >But for positional judgments, I would recommend to get the impression from >games the engine has played, But does that not also require the expert analysis? >rather than from positional suites (where you'll be >probably too busy checking if you can trust the positions or not, before it >makes sense start to tests...). You may be right about that, but I'm not willing to give up yet. Unfortunately, finding LARGE suites of test positions may require more time and effort than I'm willing to give. Study of heavily annotated top-GM games which are said, by the top-GM annotator, to be of a positional character may be the best way for a human to learn positional chess. Of course, nothing is really learned until put into practice. Actually playing games is ultimately the way to learn chess, IMHO. [But it would be really nice to have a GM for a coach.] My thinking now is that it may be good to let a chess engine analyze the entire GM game. The average analysis time per move could be set to approximate tournament time controls. The key to success of this method would be to use extremely high quality games, with equally extremely high quality GM analyses. Then the computer's understanding of the positions of the game could be compared to the GM analyses. This idea could be carried to the next step: The chess engine could be given a short time to evaluate each position of such a game but require that the engine evaluate the positions in isolation. Again, the chess engine's evaluations could be compared to the GM analysis. The final step might be to purge out all positions for which the solution is clearly of a tactical nature. It seems that we would come close to a suite of "positional" test positions by the above procedure. [Sort of like coming in the back door. : ) ] Bob D. > >Regards, >M.Scheidl
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.