Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Hiarcs vs Rybka from Leiden
- - By Vasik Rajlich (Silver) Date 2008-05-31 10:14
I took a quick look at this very interesting game. The crucial position was this one:

k6r/pb1nqpb1/2p1p3/1p2P1p1/2pPN3/6B1/PPQ1BPP1/1K1R4 b - - 0 19


Rybka expected to be able to play the natural 19. .. c5 here with a tiny black plus, which seems reasonable. However, white can then force a very interesting draw: 20. Nd6! cxd4 (20. .. Bd5 21. Nxb5 and 20. .. a6 21. Nxb7 Kxb7 22 Bf3+ favor white) 21. Nxb7 d3 (21. .. Kxb7 22. Qe4+ Kb8 [22. .. Ka6 23. a4!=] 23. Rxd4 is also drawn) 22. Bxd3+ Kxb7 23. Be4+ Ka6 24. a4! (24. Qc3 Nb6 and white still has work to do) Nxe5 25. axb5+ Kxb5 26. Bxe5 Bxe5 and now white uncorks 27. Bc6+!! with a perpetual.

The position before 27. Bc6 deserves its own diagram:

7r/p3qp2/4p3/1k2b1p1/2p1B3/8/1PQ2PP1/1K1R4 w - - 0 27


Rybka actually saw this all the way to 27. Bc6 (as otherwise 19. .. c5 is obvious and good) on Lukas' overclocked Skulltrail on move 19, but by then it was too late. She instead tried 19. .. Bf8, but then black isn't really any better and the rest of the game was just balanced.

Vas
Parent - - By Roland Rösler (****) Date 2008-05-31 10:51
Not bad! My Rybka 232 w32 sp needs depth 15 to see the perpetual at move 27. So Rybka 3 has to go up to depth 30 (31) to see it in less than 10 minutes on a skulltrail.
Parent - By Vasik Rajlich (Silver) Date 2008-06-01 07:59
Yeah, she's a lot better now in positions like this. This one was just too deep to get earlier.

Vas
Parent - - By turbojuice1122 (Gold) Date 2008-05-31 13:32
Here's one where perhaps some contempt due to knowledge of having a weaker opponent might have been good--contempt set in a way that assumes the opponent won't find a long, lengthy drawing chance. :-)  (I know, this is probably not a good idea to employ in practice--and if one can find a way to do this in practice, it would probably be pretty difficult to implement.)
Parent - By Vasik Rajlich (Silver) Date 2008-06-01 08:02
Anything is possible, but usually contempt would mean avoiding the forced draw (no matter how deep) to play the slower position.

19. .. Bf8 is a perfectly reasonable way to play for a win. Hiarcs simply played well after that.

Vas
Parent - By Ernst (***) Date 2008-06-01 21:52
Yes nice position and thanks for the explanation as Bf8 seemed strange at first glance. I put the position in Shredder XP and let it run for 15 minutes. Nine minutes on a Q6600@3GHz to switch to Bf8.

Deep Shredder XP by Stefan Meyer-Kahlen, www.shredderchess.com
CPU0: GenuineIntel EM64T Family 6 Model 15 Stepping 7 3000 MHz
CPU1: GenuineIntel EM64T Family 6 Model 15 Stepping 7 3000 MHz
CPU2: GenuineIntel EM64T Family 6 Model 15 Stepping 7 3000 MHz
CPU3: GenuineIntel EM64T Family 6 Model 15 Stepping 7 3000 MHz

12/35   0:01   -0.29   19...c5 20.Nd6 cxd4 21.Nxb7 Kxb7 22.Qe4+ Ka6 23.Qxd4 Nb6 24.Qd2 f5 25.exf6 Qxf6 26.Bg4 Rd8 27.Qb4 Nd5 28.Qa3+ Kb6 29.Bf3 (3.400.568) 2654
13/33   0:02   -0.28   19...c5 20.Nd6 cxd4 21.Nxb7 Kxb7 22.Qe4+ Ka6 23.Qxd4 Nb6 24.Qd2 f5 25.exf6 Qxf6 26.Qb4 Qf8 27.Qd2 Qf5+ 28.Ka1 g4 29.Qd6 Qe4 30.Bf1 (5.712.351) 2748
14/39   0:04   -0.52   19...c5 20.Nd6 cxd4 21.Nxb7 Kxb7 22.Qe4+ Ka6 23.Qxd4 Nb6 24.Bf3 (11.623.685) 2786
15/41   0:06   -0.32   19...c5 20.Nd6 cxd4 21.Nxb7 Kxb7 22.Qe4+ Ka6 23.Qxd4 Nb6 24.Bf3 Qb4 25.Be4 Rc8 26.f4 Qc5 27.Qc3 gxf4 28.Bxf4 Nd5 29.Bxd5 exd5 30.Be3 b4 31.Bxc5 bxc3 32.Rxd5 cxb2 (19.896.276) 2848
16/43   0:10   -0.37   19...c5 20.Nd6 cxd4 21.Nxb7 Kxb7 22.Qe4+ Ka6 23.Qxd4 Nb6 24.Bf3 Qb4 25.Rc1 Rc8 26.Qe4 Nd5 27.a3 Qc5 28.Qg4 Bh6 29.Qh5 g4 30.Qxh6 gxf3 31.gxf3 (28.990.512) 2876
17/49   0:21   -0.49   19...c5 20.Nd6 cxd4 21.Nxb7 Kxb7 22.Qe4+ Ka6 23.Qxd4 Nb6 24.Bf3 Qb4 25.Rc1 Rc8 26.Qe4 Qe7 27.Qc2 Qd7 28.Rd1 Nd5 29.Qh7 Bf8 30.Bh5 (63.780.104) 2905
18/47   0:30   -0.45   19...c5 20.Nd6 cxd4 21.Nxb7 Kxb7 22.Qe4+ Ka6 23.Qxd4 Nb6 24.Bf3 Qb4 25.Rc1 Rc8 26.Qe4 Qe7 27.Qc2 Qd7 28.Rd1 Nd5 29.Qh7 Bf8 30.Qc2 Bb4 31.a3 Be7 32.Qe4 (90.671.315) 2933
19/49   0:47   -0.44   19...c5 20.Nd6 cxd4 21.Nxb7 Kxb7 22.Qe4+ Ka6 23.Qxd4 Nb6 24.Bf3 Qb4 25.Rc1 Rc8 26.Qe4 Qe7 27.Rd1 (141.262.658) 2947
20/56   3:58    0.00   19...c5 20.Nd6 cxd4 21.Nxb7 d3 22.Bxd3 Kxb7 23.Be4+ Ka6 24.Qc3 Nb6 25.Qf3 (710.289.771) 2980
20/56   9:00   -0.15   19...Bf8 20.Qd2 Rg8 21.Ka1 Qd8 22.Bh5 Rg7 23.f4 Rh7 24.Bf3 gxf4 25.Bxf4 Nb6 26.Qa5 (1.591.976.002) 2943
21/49  11:07   -0.15   19...Bf8 20.Qd2 Rg8 21.Qe3 Qd8 22.f4 Be7 23.fxg5 Bxg5 (1.971.033.864) 2951
22/50  12:18   -0.10   19...Bf8 20.Qd2 Rg8 21.Qe3 Qd8 22.f4 Be7 23.fxg5 Bxg5 24.Qf2 f5 25.exf6 Bxf6 26.Bf3 Be7 27.Ka1 Bb4 28.a3 Ba5 29.Bh5 Qe7 30.Bd6 (2.180.083.870) 2951
best move: Bg7-f8 time: 15:00.828 min  n/s: 2.958.571  CPU 399.4%   n/s(1CPU): 740.753  nodes: 2.665.072.099
Up Topic Rybka Support & Discussion / Rybka Discussion / Hiarcs vs Rybka from Leiden

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill