Author: Mike S.
Date: 23:24:38 01/30/06
When chess endgame tables had been invented, the basic message was:
"They play perfect chess." (in the endgame positions they cover)
Examples have been given which were impressive, and at least some smaller
examples like KBN-K were easy to understand for anybody. Sofa so good... :)
But the different so called "metrics" such tables are generated with, are a
topic which is not so easy to understand, especially in it's consequences of
creating different "best" variations for the same position.
[D]7k/4Q3/8/8/4K3/8/8/7n w - - 0 1
DTM=3; Nalimov tables will show #3 in this position: White can more or less
ignore Black's knight, and cannot be prevented from playing 1.Kf5/2.Kg6/3.Qf8#
(or 3.Qg7# depending on Black's moves).
DTC=2: Following the DTC metric, White wants to win the knight. Best DTC move is
1.Qh4+ followed by 2.Qxh1. White is "DTC happy" but of course that moves are far
from perfect, from a chess player's viewpoint :-) The mate will now take much
more moves. After 1.Qh4+ Kg7 2.Qxh1 Kf6, White will need 6 more moves to
checkmate, which means a total of 8 moves after the "perfect" DTC tables
recommendation :) so to speak, than only 3 moves which the Nalimovs need, by
following the DTM logic.
(Of course, 5- and 6-piece endings will include much more complicated cases of
this type.)
For an overview of 4 metrics for a single position (paste FEN):
http://chess.jaet.org/cgi-bin/dtx?
Regards,
M.Scheidl
[Event "endgame metrics example"]
[Site "?"]
[Date "2006.01.31"]
[Round "?"]
[White "DTM"]
[Black "DTC"]
[Result "1-0"]
[Annotator "Mike Scheidl"]
[SetUp "1"]
[FEN "7k/4Q3/8/8/4K3/8/8/7n w - - 0 1"]
[PlyCount "5"]
1. Kf5 {by DTM} (1. Qh4+ {by DTC} Kg7 2. Qxh1 Kf6 {#6}) 1... Ng3+ 2. Kg6 Nf5 3.
Qf8# 1-0
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.