Computer Chess Club Archives


Search

Terms

Messages

Subject: drawn endgame situations assessed incorrectly

Author: Jeroen van Dorp

Date: 17:22:22 02/15/00


This position is taken from a game be me against Fritz 5.32, a year ago, and
it's a position I often see here posted.
I know a few from literature, e.g. the one with all the entangled pawns and
black having rooks and all-with Ra5 and a white pawn on b4 -if white takes, he
loses, if not, it's draw. But comps don't see it until 50 moves later. Duh.

I played black in this game and went for a draw by threefold repetition as I
couldn't find out how to win, and Fritz refused to draw. Fritz was playing Elo
1600 at 40/120 time controls on my PII@400MHz.


[D] 8/8/8/1k3p1p/p2K1Pb1/P6p/7B/8 b - - 0 65


As the patzer I am it took me some time to see this is a draw basically. There's
no going to the queenside for both, and if the black king goes to the kingside
h5-g4 after making space with h5-h4, white stays put with its bishop at h2 and
makes the magic triangle Ke3/f2 Bh2 pawn f4.  And it's all locked up.

Fritz however tells me that from the point of view of black it's a nice win with
a score of around -1.50 or so. Even tablebases won't help Fritz'judgement. It
analyzes dead draw moves, but rates it favourable for black. It's the old story.

My question:
Some time ago I read a post regarding these kind of positions, *especially* with
Kings, pawns only and one bishop or knight at each side, and that programs had
special problems assessing these kind of situations.

Is that true, and can someone clarify WHY it's such a problem for all? Is it
just the general principle causing the problem, or a specific ending problem
with these kind of situations?

Thanks for your input,

Jeroen ;-}

http://zip.to/jeroen



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.