Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

retrograde analysis

36 views
Skip to first unread message

Stefan Schroedl

unread,
Aug 15, 1995, 3:00:00 AM8/15/95
to
As a PhD student in computer science, I am
studying learning algorithms for two-player games. Currently, I am trying to
get or compute a complete database for the endgame
king&rook vs king&knight by retrograge analysis,, classifying positions
according to the number of moves to win or lose. It should be accessible by
my program, so I cannot use e.g.\ Ken Thompson's database on the internet.
Is anyone dealing with this question or who possesses a program
(e.g. written in C) for retrograde analysis which he kindly could give me?
I would be grateful for any kind of hint.

Stefan

Steven J. Edwards

unread,
Aug 16, 1995, 3:00:00 AM8/16/95
to
schr...@malta.informatik.uni-freiburg.de (Stefan Schroedl) writes:

A complete chess endgame position lookup table database (= tablebase)
for KRKN (among 31 other classes) can be had via ftp from the
ics.onenet.net site. The data is in the pub/chess/TB directory and
instructions for use can be located in the ANSI C source program tbt.c
also in the same directory.

The KRKN data, like all the other tablebases, is organized into two
equal sized files; one for white to move, the other for black to move.
There are stored as gzip binaries; the names are KRKN.tbw.gz and
KRKN.tbb.gz and they should be transferred as binary (not text) files.
After gunzip is applied, each of the KRKN files is exactly 2621440
bytes long. These files and all the others in the TB directory may be
freely copied and redistributed.

-- Steven (s...@mv.mv.com)

0 new messages