Author: Jeff Lischer
Date: 13:43:16 05/05/01
Go up one level in this thread
On May 05, 2001 at 12:01:08, Dana Turnmire wrote:
> Here is an interesting article found in a 1989 CCR article.
>
> "Most elementary chess textbooks assign relative values to the pieces, based
>on pawn=1, as follows: N=3, B=3, (or 3+), R=5, Q=9 (or 9.5 or 10). Most chess
>computers use these numbers in their programs; in fact they play a critical
>role. But there are serious problems arising from relying on these numbers."
>
Larry wrote an interesting article on this more recently ("The Evaluation of
Material Imbalances", Chess Life, March 1999), which was based on database
statistics. He ended up with the following basic piece values:
pawn = 1,
knight = 3 1/4,
bishop = 3 1/4
rook = 5, and
queen = 9 3/4.
He also makes the following adjustments:
bishop pair = +1/2,
rook adjustment = -1/8 for each pawn above 5 of the side being evaluated
knight adjustment = +1/16 for each pawn above 5 of the side being evaluated
The rook and knight adjustments change signs for each pawn less than 5.
In this way, the valuations naturally change as you move from the
opening/middlegame to the endgame. Let's look at the trade of 2 minor pieces for
a rook and pawn(s). At the start of the game (8 pawns each and the bishop pair):
bishop + knight = 3.25 + (3.25 + 3*0.0625) + 0.5 = 7.1875
rook = 5 - 3*0.125 = 4.625
For this case, the 2 minor pieces are worth about a rook and 2.5 pawns. At the
end of the game (say 2 pawns each, no bishop pair):
bishop + knight = 3.25 + (3.25 - 3*0.0625) = 6.3125
rook = 5 + 3*0.125 = 5.375
Now the 2 minor pieces are only worth about a rook and 1 pawn.
It seemed like a pretty logical evaluation system, particularly geared towards
computer evaluation. Has anyone experimented with this approach?
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.