I had a couple of interesting endgame positions occur in some offline test games.
This is one:
[d] 4r3/pkq5/4R2p/p1p1Qp1P/2P2P2/6P1/KP6/8 b - -
Arasan played Qxe5 here and lost quickly. Qe7 is better, but after long analysis the best defense is Rxe6. Rybka plays this rapidly. But my engine and a few others I have tried find this difficult.
Here is my output at 45minutes/move on a quad:
bm Rxe6
result: Rxe6 score: -0.56 ++ solved in 603.39 sec. (1635.63M nodes
)
Rxe6 Qxe6 Qc6 Qf7+ Kb6 Qg6 Kb7 Ka3 Qb6 Qf7+ Ka8 Qe8+ Kb7 Qd7+ Kb8 Ka2 Qb4 Qd8+ Kb7 Qd3 a4 Qxf5 Qxc4+ Ka1 Kb6 Qf6+ Ka5 Qe5 Qc1+ Ka2 Qc4+ Kb1 Qf1+ Kc2 Qf2+ Kc1
result(2): Qd7 score: -1.36 ** not solved in 2500.05 secs. (
6575.28M nodes)
Qd7 Qd5+ Qxd5 cxd5 Rxe6 dxe6 Kc6 Kb3 Kd6 Kc4 a4 Kb5 Kxe6 Kxc5 Ke7 Kd5 Kf6 Kd6 a6
Kd5 a5 Kd6 Kf7 Ke5 a3 bxa3
result(3): Qxe5 score: -1.85 ** not solved in 2500.10 secs. (
6748.49M nodes)
Qxe5 Rxe5 Rxe5 fxe5 Kc6 Kb3 Kd7 Ka4 Ke6 Kb5 Kxe5 Kxc5 a4 Kb4 Kd6 Kb5
The 2nd one is not quite so hard but still interesting:
[d] 8/r7/3p2P1/3P1p1k/5P2/p1PK4/6R1/8 b - -
Arasan played Kh6 and lost. Black is probably losing anyway but should play a2, the least bad move:
bm a2
result: a2 score: -1.32 ++ solved in 138.67 sec. (369.65M nodes)
a2 g7 Rxg7 Rxa2 Kh4 Ra1 Rg3+ Kc4 Rf3 Ra6 Rxf4+ Kb5 Rf3 c4 Rb3+ a4 Rb8 Rxd6 f4 f6 Kg3 c5 f3 c6 f2 c7 Ra8+ Kb5 Kg2 Rg6+ Kf1
result(2): Rg7 score: -3.55 ** not solved in 2500.03 secs.
2454.87M nodes)
Rg7 Kc2
result(3): Kh6 score: -3.01 ** not solved in 2500.05 secs.
6818.63M nodes)
Kh6 g7 Rxg7 Rxg7 Kxg7 Kc2 a2 Kb2 a1=Q+ Kxa1 Kf8 Ka2 Ke7 Ka3 Kd7 c4 Kc7 Kb4 Kc8
--Jon
endgame positions - find the best defense
Moderator: Ras
-
- Posts: 4406
- Joined: Fri Mar 10, 2006 5:23 am
- Location: http://www.arasanchess.org
-
- Posts: 169
- Joined: Sun May 11, 2008 10:31 pm
- Location: Switzerland
Re: endgame positions - find the best defense
[d] 4r3/pkq5/4R2p/p1p1Qp1P/2P2P2/6P1/KP6/8 b - -
No problem for Deep Sjeng WC2008. The program's evaluation is much worse after 1...Qxe5??
4r3/pk6/4R2p/p1p1qp1P/2P2P2/6P1/KP6/8 w - -
No problem for Deep Sjeng WC2008. The program's evaluation is much worse after 1...Qxe5??
Code: Select all
Engine: Deep Sjeng WC2008 (256 MB)
von Gian-Carlo Pascutto
15/34 0:00 +0.79 1...Dd7 2.Dd5+ Dxd5 3.cxd5 Txe6
4.dxe6 Kc6 5.Kb3 Kd6 6.Kc4 a4 7.e7 Kxe7
8.Kxc5 Ke6 9.Kb5 Kd5 10.Ka6 Ke4
11.Kxa7 Kf3 (2.386.958)
15/38 0:01 +0.78 1...Txe6 2.Dxe6 Dc6 3.Df7+ Kb6 4.Dg6 Kc7
5.Ka3 Db6 6.Dh7+ Kd8 7.Dg8+ Kc7
8.Df7+ Kb8 9.Dg8+ Kb7 10.Dd5+ Kc7
11.De5+ Kb7 12.Dd5+ (5.323.481) 3577
16/48 0:01 +0.78 1...Txe6 2.Dxe6 Dc6 3.Df7+ Kb6 4.Dg6 Kc7
5.Ka3 Db6 6.Dh7+ Kd8 7.Dg8+ Kc7
8.Dh7+ (7.250.190) 3720
17/50 0:07 +0.78 1...Txe6 2.Dxe6 Dc6 3.Df7+ Kb8
4.Df8+ Kc7 5.Ka3 Db6 6.Dg7+ Kb8
7.De5+ Kb7 8.Dd5+ Kc7 9.De5+ Kb7 (27.543.953) 3850 TB:3
18/52 0:09 +0.78 1...Txe6 2.Dxe6 Dc6 3.Df7+ Kb8
4.Df8+ Kc7 5.Ka3 Db6 6.Dg7+ Kb8
7.De5+ Kb7 8.Dd5+ Kc7 9.De5+ Kb7 (37.619.647) 3856 TB:27
19/58 0:16 +0.78 1...Txe6 2.Dxe6 Dc6 3.Df7+ Kb8
4.Df8+ Kc7 5.Ka3 Db6 6.Dg7+ Kb8
7.De5+ Kb7 8.Dd5+ Kc7 9.Ka2 Db4
10.Df7+ Kb8 11.De8+ Kc7 12.De7+ Kb8
13.De2 (64.008.169) 3856 TB:48
20/60 0:32 +0.78 1...Txe6 2.Dxe6 Dc6 3.Df7+ Kb8
4.Df8+ Kc7 5.Ka3 Db6 6.Dg7+ Kb8
7.De5+ Kb7 8.Dd5+ Kc7 9.Ka2 Db4
10.Df7+ Kb8 11.De8+ Kc7 12.De7+ Kb6
13.De6+ Kc7 14.Kb1 Dd2 (124.385.817) 3844 TB:142
21/61 1:14 +0.78 1...Txe6 2.Dxe6 Dc6 3.Df7+ Kb8
4.Df8+ Kc7 5.Ka3 Db6 6.Dg7+ Kb8
7.De5+ Kb7 8.Dd5+ Kc7 9.Ka2 Db4
10.De5+ Kb7 11.Kb1 Dxc4 12.Dxf5 Df1+
13.Ka2 Dc4+ 14.b3 De2+ (286.439.062) 3839 TB:890
22/61 3:52 +0.78 1...Txe6 2.Dxe6 Dc6 3.Df7+ Kb8
4.Df8+ Kc7 5.Ka3 Db6 6.Dg7+ Kb8
7.De5+ Kb7 8.Dd5+ Kc7 9.Ka2 Db4
10.De5+ Kb7 11.Kb1 Dxc4 12.Dxf5 Df1+
13.Ka2 Dc4+ 14.b3 De2+ (880.806.781) 3789
TB:1.504
4r3/pk6/4R2p/p1p1qp1P/2P2P2/6P1/KP6/8 w - -
Code: Select all
Engine: Deep Sjeng WC2008 (256 MB)
von Gian-Carlo Pascutto
14/32 0:01 +1.22 2.Txe5 Txe5 3.fxe5 Kc6 4.Kb3 Kd7
5.Ka4 Ke6 6.Kb5 Kxe5 7.Kxc5 f4
8.gxf4+ Kxf4 9.Kd6 (5.417.442)
15/32 0:03 +1.22 2.Txe5 Txe5 3.fxe5 Kc6 4.Kb3 Kd7
5.Ka4 Ke6 6.Kb5 Kxe5 7.Kxc5 f4
8.gxf4+ Kxf4 9.Kd6 (13.058.892) 4138
16/35 0:05 +1.43 2.Txe5 Txe5 3.fxe5 Kc6 4.Kb3 Kd7
5.Ka4 Ke6 6.Kb5 Kxe5 7.Kxc5 f4
8.gxf4+ Kxf4 9.Kd6 (26.007.010) 4401
17/53 0:15 +1.73++ 2.Txe5 (74.526.231) 4655
17/53 0:26 +1.83 2.Txe5 Txe5 3.fxe5 Kc6 4.Kb3 Kd7
5.Ka4 Ke6 6.Kb5 Kxe5 7.Kxc5 f4
8.gxf4+ Kxf4 9.Kd6 (125.826.604) 4690
18/53 0:48 +1.90 2.Txe5 Txe5 3.fxe5 Kc6 4.Kb3 Kd7
5.Ka4 Ke6 6.Kb5 Kxe5 7.Kxc5 f4
8.gxf4+ Kxf4 9.Kd6 (229.750.376) 4695
-
- Posts: 169
- Joined: Sun May 11, 2008 10:31 pm
- Location: Switzerland
Re: endgame positions - find the best defense
Below the output of Naum 4 in the MV-mode with 3 lines. Also this engines seems to understand this position.
Kurt
[d] 4r3/pkq5/4R2p/p1p1Qp1P/2P2P2/6P1/KP6/8 b - -
Kurt
[d] 4r3/pkq5/4R2p/p1p1Qp1P/2P2P2/6P1/KP6/8 b - -
Code: Select all
Engine: Naum 4 (256 MB)
von Aleksandar Naumov
19 0:37 +0.33 1...Txe6 2.Dxe6 Dc6 3.Df7+ Kb6 4.Dg6 Kb7
5.Ka3 Db6 6.Dh7+ Kb8 7.Dg8+ Kc7
8.Df7+ Kd6 9.Dg6+ Kc7 10.Dg7+ Kc8 (119.450.257) 3226
TB:2.366
Code: Select all
19 0:37 +1.68 1...Dd7 2.Dd5+ Dxd5 3.cxd5 Td8 4.Kb1 Txd5
5.Kc2 Td4 6.Txh6 Tc4+ 7.Kd3 Tb4
8.Th7+ Kc6 9.Kc2 Tc4+ 10.Kb1 Td4 (119.450.260) 3226
TB:2.366
Code: Select all
19 0:39 +2.17 1...Dxe5 2.Txe5 Tg8 3.Txc5 Txg3
4.Txa5 Kc6 5.Txf5 Th3 6.b4 Tc3
7.Tf6+ Kc7 8.Txh6 Txc4 9.Th7+ Kc6
10.Kb3 Txf4 (126.388.647) 3240
TB:7.914