Author: Dieter Buerssner
Date: 06:34:49 05/29/01
Go up one level in this thread
On May 27, 2001 at 12:22:33, Jon Dart wrote:
>Next round Arasan was up against Watcherz (Gnu Chess 5.0 on
>a Pentium III/800), which was a last-minute entrant. Gnu has
>been a real giant-killer in this tournament, at one point
>being tied with Shredder! Arasan was just fine, until this
>point:
>
>[D] 2r4k/p6p/4Q1p1/nr6/8/2qR4/P4PPP/4R1K1 b - -
>
>when it played Qxd3. The rook cannot be taken: this is fatal.
>Arasan never recovered, and lost. It takes the version that
>was playing 28 seconds on an Athlon 1.3GHz to avoid Qxd3,
>so it had a good chance to avoid it in the game. But it didn't.
>Curiously, Arasan 5.4 avoids it right away, while version 5.2
>fails low at ply 11 but still sticks to Qxd3 past a minute
>of search.
I looked at this position with Yace. I think, the target here should not only be
to avoid Qxd3, but to find Qc6. This is what I get on my AMD K6-2 475 with 30M
hash:
615361 4.030 1.61 7. 1...Qxd3 2.Qxc8+ Kg7 3.Re7+ Kh6 4.Qf8+ Kg5
5.h4+ Kg4 6.Qc8+ Kxh4 7.Rxh7+ Kg5 8.Qc1+ Kg4
9.Rxa7 {180}
1169827 7.440 1.21 8-- 1...Qxd3 2.Qxc8+ Kg7 3.Qc7+ Kh6 4.Qf4+ Kh5
5.g4+ Kh4 6.f3 Rd5 7.Qh6+ Rh5 8.gxh5 gxh5 {-160}
3753698 19.908 0.00 8t 1...Qxd3 2.Qxc8+ Kg7 3.Qc7+ Kf8 4.Qc8+ Kf7
5.Qc7+ Kf8 {260}
3935689 20.654 0.01 8t+ 1...Qc6 2.Rd6 Qc4 3.Qxc4 Nxc4 4.Rc6 Rxc6 5.f3
{760}
4207904 22.028 0.55 8t 1...Qc6 2.Rd6 Qc5 3.Qf6+ Kg8 4.Qe6+ Kg7 5.Rd7+
Kh6 6.Qh3+ Qh5 7.Qe3+ Rg5 8.Qe7 Rgc5 9.Qxh7+H
Kg5H 10.Rxa7H Qxh7H 11.Rxh7H {100}
4575665 24.026 0.55 8. 1...Qc6 2.Rd6 Qc5 3.Qf6+ Kg8 4.Qe6+ Kg7 5.Rd7+
Kh6 6.Qh3+ Qh5 7.Qe3+ Rg5 8.Qe7 Rgc5 9.Qxh7+
Kg5 10.Rxa7 Qxh7 11.Rxh7 {100}
4778300 25.103 0.15 9-- 1...Qc6 2.Rd6 Qc5 3.Qf6+ Kg8 4.Qe6+ Kg7 5.Rd7+
Kh6 6.Qh3+ Qh5 7.Qe3+ Rg5 8.Qe7 Rc1 9.Rxc1 Rg4
{-240}
5722617 29.316 0.00 9t 1...Qc6 2.Rd6 Qc5 3.Qf6+ Kg8 4.Qe6+ Kg7 5.Rd7+
Kh6 6.Qh3+ Qh5 7.Qe3+ Qg5 8.Qh3+ Qh5 {260}
6707654 33.830 0.00 9. 1...Qc6 2.Rd6 Qc5 3.Qf6+ Kg8 4.Qe6+ Kg7 5.Rd7+
Kh6 6.Qh3+ Qh5 7.Qe3+ Qg5 8.Qh3+ Qh5 {260}
9770281 47.580 0.00 10t 1...Qc6 2.Rd6 Qc5 3.Qf6+ Kg8 4.Qe6+ Kg7 5.Rd7+
Kh6 6.Qh3+ Qh5H 7.Qe3+H Qg5H 8.Qh3+H Qh5H
9.Qe3+H Qg5H 10.Qh3+H Qh5HR {260}
12316273 59.624 0.00 10. 1...Qc6 2.Rd6 Qc5 3.Qf6+ Kg8 4.Qe6+ Kg7 5.Rd7+
Kh6 6.Qh3+ Qh5 7.Qe3+ Qg5 8.Qh3+ Qh5 9.Qe3+ Qg5
10.Qh3+ Qh5 {260}
19549296 1:33.2 0.00 11t 1...Qc6 2.Rd6 Qc5 3.Qf6+ Kg8 4.Qe6+ Kh8 {260}
28348265 2:17.0 0.00 11. 1...Qc6 2.Rd6 Qc5 3.Qf6+ Kg8 4.Qe6+ Kh8 {260}
And stays with Qc6 and a score of 0.0 for 20 minutes.
When I don't allow Qc6 (sort of poor man's 2 best move), I get:
539770 3.496 1.61 7. 1...Qxd3 2.Qxc8+ Kg7 3.Re7+ Kh6 4.Qf8+ Kg5
5.h4+ Kg4 6.Qc8+ Kxh4 7.Rxh7+ Kg5 8.Qc1+ Kg4
9.Rxa7 {180}
1025324 6.382 1.21 8-- 1...Qxd3 2.Qxc8+ Kg7 3.Qc7+ Kh6 4.Qf4+ Kh5
5.g4+ Kh4 6.f3 a6 7.Qh6+ Rh5 8.gxh5 gxh5 {-160}
2057058 11.482 0.00 8t 1...Qxd3 2.Qxc8+ Kg7 3.Qc7+ Kf8 4.Qc8+ Kf7
5.Qe6+ Kg7 6.Qe7+ Kg8 7.Qe8+ Kg7 8.Qe7+ {260}
2871048 15.691 0.00 8. 1...Qxd3 2.Qxc8+ Kg7 3.Qc7+ Kf8 4.Qc8+ Kf7
5.Qe6+ Kg7 6.Qe7+ Kg8 7.Qe8+ Kg7 8.Qe7+ {260}
3636502 19.542 -0.40 9-- 1...Qxd3 2.Qxc8+ Kg7 3.Qc7+ Kf8 4.Qe7+ Kg8
5.Qe8+ Kg7 6.Re7+ Kh6 7.Qf8+ Kg5 8.h4+ Kg4
9.Kh2 Rg5 10.hxg5 Kxg5 {-160}
7775810 38.955 -1.77 9t 1...Qxd3 2.Qxc8+ Kg7 3.Qc7+ Kf8 4.Qe7+ Kg8
5.Qe8+ Kg7 6.Re7+ Kh6 7.Qf8+ Kg5 8.h4+ Kg4
9.Kh2 Rf5 10.f3+ Rxf3 11.gxf3+ Qxf3 12.Qxf3+
Kxf3 13.Rxa7 {-160}
12426994 1:00.1 -1.76 9t+ 1...Qc5 2.Qf6+ Kg8 3.Rd7 Qf8 4.Qxf8+ Rxf8 5.Re8
14045977 1:07.9 0.00 9t 1...Qc5 2.Qf6+ Kg8 3.Rd7 Qf8 4.Qe6+ Kh8 5.Rf7
Qd8 6.Rd7 Qf8 {260}
14415137 1:09.6 0.00 9. 1...Qc5 2.Qf6+ Kg8 3.Rd7 Qf8 4.Qe6+ Kh8 5.Rf7
Qd8 6.Rd7 Qf8 {260}
20825307 1:39.0 -0.40 10-- 1...Qc5 2.Rd5 Qxd5 3.Qxc8+ Kg7 4.Re7+ Kh6
5.Qf8+ Kh5 6.Rxh7+ Kg5 7.Qe7+ Kg4 8.f3+ Kf5
9.g4+ Kf4 10.Rf7+ Qxf7 11.Qxf7+ Kg5 {-321}
23022559 1:49.3 -2.38 10t 1...Qc5 2.Rd5 Qxd5 3.Qxc8+ Kg7 4.Re7+ Kh6
5.Qf8+ Kg5 6.f4+ Kh5 7.h3 Rb7 8.g4+ Kh4 9.Qh6+
Kg3 10.Re3+ {HT} {260}
37220088 3:02.1 -2.38 10. 1...Qc5 2.Rd5 Qxd5 3.Qxc8+ Kg7 4.Re7+ Kh6
5.Qf8+ Kg5 6.f4+ Kh5 7.h3 Rb7 8.g4+ Kh4 9.Qh6+
Kg3 10.Re3+ {HT} {260}
39981141 3:15.8 -2.78 11-- 1...Qc5 2.Rd5 Qxd5 3.Qxc8+ Kg7 4.Re7+ Kh6
5.Qf8+ Kg5 6.f4+ Kh5 7.h3 Rb7 8.Kh2 Qd1 9.g4+
Qxg4 10.hxg4+ Kxg4 {-581}
74296718 5:51.8 -3.00 11t 1...Qc5 2.Rd5 Qc6 3.Qe5+ Kg8 4.Rxb5 Nc4 5.Qd5+
Qxd5 6.Rxd5 Rc7 7.f4 a5 8.g4 {-240}
137147350 11:25.0 -3.00 11. 1...Qc5 2.Rd5 Qc6 3.Qe5+ Kg8 4.Rxb5 Nc4 5.Qd5+
So, to Yace all moves look losing besides Qc6. It may also be interesting to
note, that at depth 9 both moves, Qc5 and Qc6 show a score of 0.0. This may also
show the great amount of luck, that can be involved in engine games. Yace would
luckily have played Qc6 at rather short time control, even if it gets the same
score as Qc5 first.
I think, this would be a nice addition, to your Arasan test suite.
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.