White: Gandalf 5.0
Black: Kenneth Frey (MEX)
Download game in pgn-format.
16 December 2002 - Draw by mutual agreement.
Comment from Kenneth Frey
A short opinion and my thoughts during the game might complement the
match, without any lengthy analysis.
I felt quite happy after the first moves. White must improve on the
variation if he pretends to keep the advantage. Black is comfortable after
16.Nxe4. I guess choosing an opening is as difficult for Gandalf as for the
rest of us. As children we learned not to bring out the queen early, and this
games just proves the accuracy of the axiom, as her majesty was forced back
with no glory.
Now the first important and difficult decision was reached : to keep the
tension, take on d4 or play 16...c4. Every modern master has warned against
releasing the tension, because with a secure center, white then has a free
hand on the kingside. Certainly white´s prospects there could not be taken
lightly. I think my choice was correct, but only narrowly.
Gandalf had the very difficult task of proving that 16...c4 wasn´t
accurate, but 18.h3 gave me the impression he couldn't solve the strategical
complexities. I couldn't figure out why he thought the move was necessary.
White must play sharper before black's queenside begins to roll. This could be
achieved by bringing the queen or a rook to e4 and push the h-pawn to h4 and
h5. So white's critical point is his 18th move.
Then came the hideous 20...Nxc3. It stopped any attacking chances on the
kingside but killed the game that now turned very dull. The alternative was
the double-edged 20...Qd7 21.Qe4 Rfd1 or 21.Re4 a5 with interesting
possibilities for both sides, but white must find how to increase the pressure
on the black king. Unfortunately I have lost my analysis of the position
because I had little time to write down the variations at that period. In
general, white is not without chances and the spectrum of a mate finally
haunted me. So I failed to test Gandalf properly, but the little I saw is
My thanks to Gordon for his friendship and efficient management of
Gandalf, to Marijus Kulvietis, our captain, always ready to take another
challenge and Rebel, now a familiar company for me, after annotating every
game in the Odyssey 2001 tournament.
Comment from Gordon Rattray
As mainly a spectator myself, I can't comment much on the game directly.
However, I enjoyed playing a part and it was interesting to watch Gandalf
analyse upon receiving a move from Kenneth.
Once out of its opening book, Gandalf's numerical evaluation of the game
positions didn't change very much. I think at the very most it once
estimated about 0.3 of a pawn in its favour. The final position was
evaluated at 0.0. So, in summary, Gandalf estimated that the game
was almost level throughout.
It is of great credit to both players that they didn't concede any
disadvantages during such high level play. Congratulations to Kenneth
for playing so well, and congratulations to Steen for writing Gandalf.
Operator of Gandalf on Board 2:
Country: Scotland, UK. (City: Glasgow)
Occupation: Software engineer
Interests: computer chess and AI in general
Used hardware for this match: 2 separate Athlon 1.1 Ghz PC's using 200 Mb hash tables and all 3, 4 and 5 man endgame tablebases.