Author: Matthew White
Date: 16:04:17 12/09/02
Go up one level in this thread
On December 09, 2002 at 00:11:41, Mike S. wrote:
{Lots of snipping}
>(c) The first move of the solution should have a "testing character". This
>means: The first move should be unusual (would commonly bad), for that it's >only chosen when the engine has seen the idea behind the move. Sacrifices are >easy examples for that.
Thematic moves might also be very useful in a test suite. For example, I am not
sure how many engines successfully recognize when they have a minority attack,
or when it is a good idea to trade bishops if they are trying to give their
opponent a weak color complex, but these are concepts known by stronger players,
and their strategic implications should be well understood.
>(d) The position should not allow to *delay* the idea with in-between-moves,
>i.e. by playing a primitive but strong threat the opponent must respond to
>first, before playing the intended idea. This can destroy the quality of a test
>position which would be ok otherwise, and is a big problem in positional tests
>(where things, i.e. move order, are most often not so "forced" like in
>combinations).
I agree with the above, but don't zwischenzug moves often refute combinations?
If so, finding the "in-between" move might be a decent test.
Another type of position that I would consider including, which is similar to
Mike's discussion of (b), would be a position in which one MUST NOT make a
specific move (sorry, I'm a QA guy, negative tests are useful too), since that
move would prove to be a positional blunder. Computers often push pawns if they
can't see another useful move, and in some positions this might prove deadly
(though it will be subtle).
One final type of position is one where the best move is the best because it
results in a theoretically won endgame (though not a tablebase endgame). Yes,
this is almost exactly like (b), but with one subtle difference: it won't win
outright, but it shows that an engine is capable of playing for a favorable
endgame.
I guess that the theme that I am expressing here is that there are many
different ways to win. True, they are all wins, but winning because you did so
by a specific path will cause engines to rely less on brute speed and more on
"understanding," which I believe with help iron out some apparent computer chess
weaknesses.
Matt
This page took 0.01 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.