| |
Date: 13 Jan 2009 23:51:51
From: help bot
Subject: My Mistakeless Game
|
On Jan 14, 12:20=A0am, Sanny <softtank...@hotmail.com > wrote:
> Here is a game between Help Bot and Easy Level. The game was equal
> till 37th move where Help Bot took the pawn of Easy Level.
I ran this through a quick computer analysis, which
indicates that although much of the game was even,
I had a big advantage from moves 14 through 19, at
which point I blundered horribly.
> Then Help Bot was a pawn up and was able to win in 50 moves.
>
> Was there any way to save the Pawn. Since GetClub do not know all end
> games rules. What was the mistake in End game that GetClub loose that
> Pawn.
>
> Till 35th move the game was quite equal. But suddenly Help Bot got 1
> pawn up. Was there any Tactical / Strategy mistake by GetClub?
>
> Game Played between help bot and easy at GetClub.com
>
> -------------------------------------------------------------------------=
--=AD-----
> help bot: (Black)
> easy: (White)
> Game Played at:http://www.GetClub.com/Chess.html
> View Recorded Game:http://www.getclub.com/playgame.php?id=3DDM37417&game=
=3DChess
Sanny, many chess programs have serious
difficulties in handling Rook and pawn endings
well. Even Rybka, in one of its recent odds
matches against GM Dzindzi., messed up a
fairly simple Rook ending!
Chess books are filled with examples, culled
from grandmaster play, in which very simple
winning or drawing ideas were rejected in
favor of some hare-brained idea. In fact, at
one time certain drawable endings were
mistakenly believed by the talking-heads to be
wins, simply because Mr. Capablanca did win
them, and fairly routinely, against "masters".
Let's jump to move 34, where I allowed a
"Rook to the seventh rank", along with other
active tries like R-h4 or R-g4:
34. Kf3
This prepares to deliver some spite-checks,
beginning with 35. Re4+. But you see, the
spite will amount to nothing, while the loss
of time (one tempo) gives me the initiative!
What ought to have happened was a pawn
race in which GetClub eats my pawns (on g7
and h7), then runs its h-pawn up the board,
while I eat the White pawns (on a4 and c4),
and run my a-pawn up the board. I have no
idea who wins, because it's too close to call.
Whichever side makes an unneccessary
move, losing a tempo, may well end up
losing the game (or throwing away a draw).
Anyway, if both sides play actively, it's
very, very close; what the annotators like to
call "equal chances".
After what happened in the game: 34. Kf3
...Rb4, Black is on top because the a-pawn
falls and that gives me an outside passed
pawn (way, way outside because the White
King is now on the Kingside).
The only way I know of to correct this kind
of problem is to *remove* the bonus your
program obviously gives for spite-checks;
spite-checks are irrelevant.
Rybka plays this ending very, very
differently, keeping the threats coming at
me like bees near a disturbed hive!
As of 37. c5, the game is already decided.
Any exchange of Rooks along the rank will
result in a won King and pawn ending. Even
so, after my 38. ...e5, the move 39. Rf5 was
much better than going down without a fight.
-- help bot
|
