Asri-unix.217
net.chess
utzoo!decvax!ucbvax!menlo70!sri-unix!mclure
Mon Dec 14 06:16:44 1981
Elite vs. MCT match summary
In the Fidelity Elite vs. Applied Concepts Master Chess Trio
match, the Elite won 5 out of 5. Some comments about the games
are included here.
In game 1, MCT managed to gain a decisive advantage of bishop and
knight vs. rook going into the endgame, but through careless play
allowed allowed the exchange of all but one of its pawns. This
left Elite with a couple more pawns, all connected, which proved
crushing. Elite's only blunder was 20. Bf7, exchanging its
bishop and knight for a rook and pawn, a common computer error
for programs which consider bishop and knight equal to 6 pawns
and rook and pawn equal to 6 pawns. However, Elite played very
surprisingly 63 Rd2, sacrificing the exchange in the endgame in
order to eliminate any of MCT's remote counterplay which would
have lengthened the game by a few moves. Apparently the
programmers have told Elite how to recognize in king and pawn
endgames the technique of whether the enemy king is within the
square of the pawn which is to queen. Elite noticed this and gave
up the last pieces on the board to shorten the game.
In game 2, Elite inflicted doubled pawns on MCT's c file with 7.
.. Bc3, gained control of the d file with 10. ... Rd8, and
proceeded to put inexorable pressure on the weak pawns. MCT
entered the endgame in an unfavorable position and a pawn or two
down. Elite quickly forced the win by promoting a pawn and
forcing MCT to sacrifice its rook for the queen.
In game 3, we see the first crush of the match. A 33 mover in
which MCT left the opening with a better position, but quickly
deteriorated (under the inferior play of the Morphy module).
Elite then plays a risky maneuver with 17. Bh6, a position in
which the bishop can often be trapped. However, MCT then trades
its remaining minor piece for three of Elite's pawns, two of them
in front of the Elite castled position. In some circumstances,
this can be favorable, but more force must be available to take
advantage of the open king position. MCT had no such force. Its
rooks were out of play. So in effect, Elite was playing a piece
up, unless MCT could trap Elite's wayward bishop. Unfortunately
for MCT, it couldn't. Elite's rooks were just too strong, and
MCT's rooks never did get into play so Elite was then playing
with two rooks, bishop, and queen against MCT's sole queen. The
outcome was a quick mate.
In game 4, Elite again came out of the opening with an inferior
position. By move 18, MCT had a very superior position, but in a
long combination it failed to see as far as Elite and lost a
piece because on the last ply, the equivalent enemy piece could
not be recaptured because of a double attack by Elite's queen
against MCT's rook and king which could not defend each other.
This combination started with Elite's 18. ... Nf3. Elite easily
won the enusing endgame. Amusingly, it shortened the win again as
in game 1 by sacrificing material in order to get a pawn outside
of the square of the Enemy king, even though it was a full piece
ahead. This produces a very nice "human" effect for the Elite
endgame in these two games. Perhaps even more amusing is that
Elite promoted a second pawn, this one to a knight instead of a
queen in order to shorten the mate still further.
In game 5, the final game, both machines made very strange moves
at 6. Nc6 dc. 6. Nc6 is not good in the Sicilian because it
strengthens Black's center. However, this is a very common
computer error. I believe chess 4.x made it against Levy in their
historic match in Toronto for the bet. I think Levy was counting
on the precise error. Inexplicably, MCT recaptured with the e
pawn, allowing the exchange of queens and preventing itself from
castling! This looks like a much more serious bug than what
allowed Elite to play 6. Nc6. The resulting pressure on the d
file after Elite's queenside castle was simply too great and MCT
lost a bishop. Another serious program bug occurred when MCT
exchanged with 15. ... Bc3 and gave up its only developed piece.
The Elite will not trade when down material unless it produces a
forced draw by lack of sufficient mating material, perpetual
check, gain of material, etc. The Capablanca cartridge doesn't
seem robust in this respect. Also, Capablanca refused to develop
its rooks after the exchange and lost still more material. Elite
obtained a crushing positional advantage with doubled rooks on
the 7th rank and ended the game quickly thereafter.
Summary: MCT has many serious bugs in the Morphy and Capablanca
modules. Morphy seems especially weak. In practically every
game, MCT left the opening with a superior position, but quickly
degraded its advantage through the inferior play of the Morphy
module. So in this respect, the Gruenfeld opening play appears
slightly superior to the Elite opening book. I don't think this
is too surprising since I believe the Gruenfeld book is larger
than the Elite's. Elite demonstrated an admirable ability to
wrest the initiative from MCT and apply continuous pressure until
MCT crumbled. A sixth game was also played, but not recorded.
Elite won from the black side of a Sicilian, after again having a
cramped opening against MCT's white. It is unclear why the MCT
has such a high reputation. It is almost universally touted as
the strongest mass commerical chess device (the Elite is a
limited production model), but I now find this reputation very
hard to believe.
-----------------------------------------------------------------
gopher://quux.org/ conversion by John Goerzen <
[email protected]>
of
http://communication.ucsd.edu/A-News/
This Usenet Oldnews Archive
article may be copied and distributed freely, provided:
1. There is no money collected for the text(s) of the articles.
2. The following notice remains appended to each copy:
The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996
Bruce Jones, Henry Spencer, David Wiseman.
