Author: Dieter Buerssner
Date: 13:56:06 04/07/02
Go up one level in this thread
On April 07, 2002 at 13:45:40, Eduard Nemeth wrote:
>Kramnik - Leko
>
>[D]2br1r2/1pR1Rp2/1p3k2/1B1p1Pp1/PP1P2P1/K7/8/8 w - - 0 1
>
>1. a5!!
>
>This position is now today hard for all programs, or?
It is also very hard for my program. I first tried to do the correct move on the
board, and saw soon convincing scores and lines. But, for the position directly,
Yace needed lots of time (I only had the patience, because I had seen the scores
after a5):
141314362 14:02.7 1.51 17t 1. Kb2 Kg7 2. Kc3 Kf6 3. Kd2 Kg7 4. Kd3 Kf6 5.
Ke2 Kg7H 6. Re6H {HT} {0}
168206928 17:19.8 1.52 17t+ 1. a5 bxa5 2. bxa5 Kg7 3. Re5 Rh8 4. Kb4 Rh1 5.
Be8 Rb1+ 6. Ka3 Ra1+ 7. Kb2 Rxa5 8. Bxf7 Kf6 9.
Be6 Bd7 10. Bxd7 Rc8 {350}
212698963 23:31.2 1.91 17++ 1. a5 bxa5 2. bxa5 Kg7 3. Kb4 Kf6
237794327 26:48.7 1.92 17t 1. a5 bxa5 2. bxa5 Kg7 3. Kb4
But at least, it showed some slightly convincing score jump, and I am confident
(after seeing the scores when I did the move on the board), that with more time,
the score would rise.
After this behaviour (seeing good score soon, after the move is made, but
needing a long time from the root position), I disabled all pruning. Then the
move was found at depth 11 (also score 1.92) - this time it needed about 8
minutes, because of the much inferior branching factor. So, this position seems
to trigger some unfortunate behaviour due to pruning techniques. Another
followup mentions this also for Fritz. I can have the hope, that Yace with
default setup would perhaps play this move at typical tournament time control on
a modern computer with ponder.
Regards,
Dieter
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.