Meeting M. Campbell and Joe Hoane
Han Schut
Yesterday there was a press conference in Rotterdam
on Deep Blue and the start of the Dutch play-off match.
Some quotes:
- 'There is no rematch between DB and GK scheduled.
That doesn't mean there will never be one.'
- 'The IBM team hopes next time GK will play his own
style of chess, i.e. no anti-computer chess.'
- 'Murray Campbell and Joe Hoane did not know anything about
a possible match between Anand and Deep Blue as stated
in an Indian newspaper last week.'
- 'Deep Blue is probably still slightly weaker as GK at this moment.'
- 'The move Be4 in game 2 (instead of the expected Qb6) was nothing
special. Deep Blue finds it in 20 sec. and so does Rebel and Fritz.
It is just a question of 'chess knowledge'; the computer does not
like to give chances for counter attack.'
- 'The so-called random move in game 1 was a 'bug''
- Hans Bohm (IM) played a 15 minute game against Deep Blue and
lost an even endgame.
I also played several games against Deep Blue jr.
- 5 minutes per person per game
- my rating: FIDE 2130; IECC: 2344; ICC: 2100-2360
- Deep Blue jr.: 15 million positions per second; rating: approx. 2600
- same chess algorithm as Deep Blue sr.
I was amazed that Deep Blue jr. has such a small opening book.
It plays in my opinion inferior variations. Some examples:
(I played all the games with white.)
1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.Bc4 (DB-jr out of book) Nf6
5.0-0 Nxe4 6.Re1 d5 7.Bxd5 Qxd5 8.Nc3 ...
game 1: 8. ... Qa5 9.Nxe4 Be6 10.Neg5 0-0-0 11.Nxe6 fxe6 12.Rxe6 Bd6
13. Bg5 Rf8 14.Qe2 Qb4?! 15.Rb1 h6 16.Bh4 Qa5 17.a3 Qh5
18.Bg3 Bxg3 19.hxg3 and aborted for a presentation
game 2: 8. ... Qc4?! 9.Nxe4 Be6 10.b3 Qd5 11.c4 Qd7 etc.
game 3: 8. ... Qa5 9.Nxe4 Be7?! 10.Bg5 f6 etc.
Murray Campbell said that the opening book was 'intentionally kept
small in order not to make too many choices for the computer.'
Regards,
Han Schut
hans...@pi.net