Author: Komputer Korner
Date: 03:56:12 07/31/98
Go up one level in this thread
On July 31, 1998 at 01:51:57, blass uri wrote:
>The programs I know give me evaluation in pawns and I prefer to see
>in the evaluation function the predicted result of the game(number between 0
>and 1) and not an evaluation in pawns.
>
>Uri
You can always do the translation in your head if you memorize and prorate my
Komputer Korner Annotation table.
KOMPUTER KORNER ANNOTATION TABLE
The following chart won't improve your chess but it will give you an idea of
what exactly is meant by the annotation symbols that GM's and chess programs
use.
Komputer Korner Annotation Table.
Evaluation symbol % score for white % of a pawn ahead
No. of tempos ahead
--------------------------------------------------------------------------------------------------------------------------------
= or unclear/= 50 .00 0
+/=/= or unclear 56 .16 0.5
+/= 63 .33 1
+/- 76 .67 2
+ - 90 1.0 3
+/-+/- >99 > 1.33 4 or more
This variables can easily be graphed in a linear fashion to achieve a full set
of numbers.
The difference between the = sign and the unclear/= sign in row 1 is that the
unclear/= symbol represents a more dynamic position with much more winning
chances for both sides. In row 2, the unclear symbol means the same thing as
per the above with the exception that white has slightly more winning chances
than black. Unfortunately, Sahovski Informator insists on using the =/unclear
symbol for a material inequality position where there is sufficient
compensation. Unfortunately that does not tell us whether there are lots of
equal winning chances or if the logical outcome should be a draw. In practice,
it has come to mean that the logical outcome is a draw which leaves a hole in
the symbol annotations because there is no generally accepted symbol for the
saying" both sides have equal chances". The unclear symbol can't be used for
this because it does not represent an equality ( See Axiom No. 3 of Komputer
Korner's 10 Commandments of Opening Theory). I propose that the =/unclear symbol
be used in these cases. This then differentiates between a dynamic equality
where both sides have equal chances to win, and the = symbol by itself which
should mean that a draw is the likely outcome. The interpretation used by
Sahovski Informator is thus misleading and not needed, because we already know
that there is a materiel inequality just by looking at the position. The
annotation symbols should restrict themselves to telling us who is better and if
the position is equal, is it because both sides have lots of winning chances or
is it because the position is a draw?
I have found only 6 positions in chess opening theory where one side has 3
pawns or greater material difference but the other side has enough positional
compensation so that the position is stable and unclear. In almost all other
positions, positional compensation can at most amount to 2 pawns in the opening.
Thus the contempt factor of chess software should be increased to allow
compensation for at least 2 pawns. Of course in the late middlegame and endgame
it is possible to dream up positions where a queen might be equal to a minor
piece especially when advanced passed pawns are concerned, but that is not our
concern here. The above chart assumes that the positional compensation for the
inferior side's pawn/ material disadvantage has already been factored in.
--
Komputer Korner
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.