Author: Amir Ban
Date: 12:37:22 05/01/01
Go up one level in this thread
On May 01, 2001 at 13:30:48, Shay Bushinsky wrote:
> We were happy to take part in the BGN qualification match held in Cadaques.
>We would like to thank Prof. Enrique Irazouqi for his great effort both in
>organizing and in conducting this exciting event. Although the end result was
>painful, we are proud with the performance of Deep Junior and believe that it
>sparkled the imagination of many computer chess followers.
> We would also like to congratulate the Deep Fritz team,
>Frans Morsch and Mathias Feist, for their win and for their very fine program.
>We wish them success towards the ultimate confrontation with Kramnik.
> Lets hope that beyond the disputes that practically characterize the world of
>{computer} chess, we will keep seeing more of such events and that programs will
>continue to play an important role in the game.
>
>Amir & Shay
I'm signed to the above. I'll add a few notes:
We noticed the low NPS number recorded in the log in some of the moves. I did
not have this version of the interface which records NPS, so I had nothing to
compare it to, but I saw that the low number occurred in one of two cases: 1.
When Junior played immediately out of permanent brain 2. When Junior got a
timeout in considering a move. Since otherwise NPS looked normal, my guess was
that something goes wrong in the NPS calculation here (the engine reports total
nodes, the interface converts it to NPS).
It became obvious during the match that match strategy is needed, and the
programs were not behaving rationally for the position in the match (e.g. see
game 22). This could have been handled effectively by the contempt parameter.
Tragically, this option was not available to Junior. I asked Enrique before the
match to use 0 for contempt, and he, pleading forgetfulness, asked me to send a
version that has this value hard-wired, and I did. So later, when I woke up to
what was going on, I could do nothing about it.
Likewise for two games that were scrapped due to technical glitches and
replayed. Both were stopped in positions that were at least not worse for Junior
(and for a still unknown reason, Fritz noted these games as losses). In
hindsight we should have insisted that these games be continued, but we didn't.
My expectation before the match was that it's too close to call.
I believe Junior's playing style would be more effective vs. Kramnik, but he
will be quite busy vs. Fritz. Good luck!
Amir
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.