Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: KBBPKR EGDB -- Regarding FEG

Author: John Merlino

Date: 09:07:11 12/14/02

Go up one level in this thread


On December 14, 2002 at 11:20:26, Marc Bourzutschky wrote:

>Even more interesting is the fact that with a g-pawn there are positions that
>take very long to win (up to 180 moves to mate), with sometimes over 100 moves
>required before white even moves the pawn from the starting square.  Below
>is such a position, a mate in 150.  These positions were analyzed with an
>EGDB generated with Johan de Koning's FEG program, freely downloadable from
>www.chessmaster.com.

Regarding FEG, an update to this program was just released yesterday that fixes
an obscure bug in the data creation process. In a nutshell, here is how Johan
described the bug:

------------
Endgames with mates no longer than 1 (eg KNNK) did not contain the
black-is-checkmated positions, or rather their wrong value being
"not_white_win". This affected also some positions leading to
black-is-checkmated by capture (eg KNNKR with the mating move being NxR).
------------

Johan sent me a detailed list of all of the file sets that are affected by this
bug, but unfortunately I seem to have misplaced it. Marc, he might have sent
this list to you. If so, could you please post it, so that people who have
generated these file sets with FEG can do so again if they want to?

This bug only affects very very few positions in the data, but it can lead to
incorrect data if you happen to be testing with such a position. Johan and I
very much apologize for the inconvenience....

jm

p.s. Marc Bourzutschky deserves the credit for discovering this bug, and doing a
lot of testing and research to track it down before reporting it. He also,
obviously, has devoted quite a bit of CPU time and hard drive space to
developing a large amount of 6-man files. As of his last report to me, which was
about six weeks ago, he had used 3 CPU months, 100G of hard drive space, and 20
backup DVDs! We are very much appreciative of his (and his computer's) hard
work.



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.