Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Rybka's difficulties with pawn endgames.
- - By ernest (****) Date 2007-07-25 19:55
Two examples of positions, where Rybka 2.3.2a is clueless.

First example, coming from http://64.68.157.89/forum/viewtopic.php?t=15262&start=0
and http://www.rochadekuppenheim.de/test/schach/smith.htm
(analysis of a variation, from Anand,V - Chess Genius, London, 1994)

3k4/8/1Pp5/3P2p1/4K1Pp/8/8/8 w - -

White to play:
only d6! wins
dxc6? leads to a draw

Engine: Rybka 2.3.2a 32-bit (256 MB) by Vasik Rajlich
................
43.01  21:35   +3.61   1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Ke3 Kb8 5.c7+ Kb7 6.Ke4 Kc8 7.Kf3 Kd7 (228.187.767) 180  TB:1.051.020
44.01  28:12   +3.61   1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Ke3 Kb8 5.c7+ Kb7 6.Ke4 Kc8 7.Kf3 Kd7 (296.434.951) 179  TB:1.307.127
45.01  46:08   +3.61   1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Ke3 Kb8 5.c7+ Kb7 6.Ke4 Kc8 7.Kf3 Kd7 (479.194.228) 177  TB:1.812.015
best move: d5xc6 time: 56:40.719 min  n/s: 178.175  nodes: 591.705.819 TB: 1.998.792

Curiously, Rybka 1.0 Beta immediately finds the correct move...

Second example, sent to me by Pierre Nolot: a study by Nikolai Grigoriev (1920, 1930...?)

5k2/5p2/8/ppp1P2P/8/1K6/6P1/8 w - -


White to move. The goal: to find the solution g4! and to show that it wins

Rybka feels, on the contrary, that the position is completely lost for White...

Engine: Rybka 2.3.2a 32-bit (256 MB) by Vasik Rajlich
16.01   0:02   -5.61   1.g4 c4+ 2.Kc2 b4 3.g5 Kg7 4.Kb2 a4 5.Kb1 a3 6.Ka2 Kg8 7.h6 Kh7 (459.546) 189  TB:11
16.07   0:10   -5.08   1.Kb2 b4 2.Kb3 Kg7 3.g3 Kh7 4.g4 Kh6 5.Kc4 a4 6.Kd3 a3 7.Kc2 c4 (2.053.443) 201  TB:12
17.01   0:10   -5.09   1.Kb2 b4 2.Kb3 Kg7 3.g3 Kh7 4.g4 Kh6 5.Kc4 a4 6.Kd3 a3 7.Kc2 c4 (2.159.004) 201  TB:20
18.01   0:23   -6.81   1.Kb2 Kg7 2.g3 a4 3.Ka3 Kh6 4.g4 c4 5.Kb4 Kg5 6.Ka3 c3 7.Ka2 b4 (4.580.214) 202  TB:79
19.01   1:10   -7.13   1.Kb2 Kg7 2.g3 a4 3.Ka3 Kh6 4.g4 c4 5.Kb4 Kg5 6.Ka3 c3 7.Ka2 b4 (13.829.468) 200  TB:357
20.01   2:40   -9.60   1.Kb2 Kg7 2.g3 a4 3.Ka3 Kh6 4.g4 c4 5.Kb4 Kg5 6.Ka3 c3 7.Ka2 b4 (32.556.669) 208  TB:993
21.01  18:40   -11.41  1.Kb2 Kg7 2.g3 a4 3.Ka3 Kh6 4.g4 c4 5.Kb4 Kg5 6.Ka3 c3 7.Ka2 b4 (213.515.916) 195  TB:5.229
21.02  32:06   -11.04  1.g4 a4+ 2.Kc2 b4 3.g5 Kg7 4.Kb2 a3+ 5.Kb3 c4+ 6.Ka2 c3 7.Kb3 Kg8 (351.237.647) 186  TB:11.383
best move: g2-g4 time: 85:35.156 min  n/s: 196.889  nodes: 987.341.970 TB: 16.739

Many other engines don't think the position is lost, but give an evaluation score which is far from a win.

An exception is (as often in pawn endings), Yace 0.99.87 !!!

Engine: Yace 0.99.87 (256 MB) by Dieter Buerssner
9/28   0:01   -4.20   1.h6 Kg8 2.g4 b4 3.g5 Kh7 4.Kc2 c4 5.Kb2 Kg6 6.Kc2 b3+ 7.Kc3 Kh7 8.e6 fxe6 9.Kb2 e5 10.Kb1 Kg6 11.Ka1 Kxg5 (3.062.835) 2620
9/28   0:01   -4.19++  1.g4 Kg7 2.g5 b4 3.g6 c4+ 4.Kxc4 a4 5.e6 f5 6.h6+ Kh8 7.h7 b3 8.g7+ Kxh7 9.g8Q+ Kxg8 (4.061.326) 2654
9/28   0:02   -3.35   1.g4 Kg7 2.g5 b4 3.Kc2 a4 4.g6 fxg6 5.h6+ Kxh6 6.Kb2 g5 7.e6 a3+ 8.Ka2 Kg7 9.Kb3 a2 10.e7 a1Q 11.e8Q (5.442.267) 2648
9/30   0:02   -3.35   1.g4 Kg7 2.g5 b4 3.Kc2 a4 4.g6 fxg6 5.h6+ Kxh6 6.Kb2 g5 7.e6 a3+ 8.Ka2 Kg7 9.Kb3 a2 10.e7 a1Q 11.e8Q (6.440.632) 2653
10/32   0:03   -3.75--  1.g4 a4+ 2.Ka2 c4 3.g5 Kg7 4.Kb1 a3 5.Kc1 c3 6.Kb1 b4 7.Ka1 Kh7 8.e6 fxe6 9.g6+ Kg8 10.Kb1 b3 11.g7 Kxg7 12.Ka1 a2 13.h6+ Kf7 14.h7 (9.585.368) 2709
10/32   0:04   -3.76   1.g4 c4+ 2.Kc3 Kg8 3.g5 f5 4.exf6 b4+ 5.Kc2 Kf7 6.Kb2 Ke6 7.h6 c3+ 8.Kc2 b3+ 9.Kxb3 a4+ 10.Kxa4 c2 11.g6 c1Q (12.657.208) 2739
10/32   0:05   -3.76   1.g4 c4+ 2.Kc3 Kg8 3.g5 f5 4.exf6 b4+ 5.Kc2 Kf7 6.Kb2 Ke6 7.h6 c3+ 8.Kc2 b3+ 9.Kxb3 a4+ 10.Kxa4 c2 11.g6 c1Q (13.971.802) 2757
11/32   0:08   -3.50   1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 Kg8 8.g6 fxg6 9.h6 a2 10.Kxa2 g5 11.Kb3 g4 12.e6 g3 13.Kxb4 (24.121.767) 2814
11/32   0:08   -3.50   1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 Kg8 8.g6 fxg6 9.h6 a2 10.Kxa2 g5 11.Kb3 g4 12.e6 g3 13.Kxb4 (25.406.567) 2823
12/35   0:12   -3.10++  1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 Kg8 8.g6 fxg6 9.h6 Kf7 10.e6+ Kf8 11.e7+ Kxe7 12.h7 c2 13.h8Q g5 (36.359.563) 2832  TB:18
12/35   0:14   -2.10++  1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 a2 8.Kxa2 Kg8 9.e6 fxe6 10.Kb3 Kg7 11.Kc2 e5 12.Kb1 b3 13.Kc1 Kg8 14.Kd1 (40.816.473) 2823  TB:18
12/35   0:16   -1.36   1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 c2 8.Kxc2 a2 9.Kb2 b3 10.g6 fxg6 11.hxg6 Kxg6 12.e6 Kf6 13.Ka1 Kxe6 14.Kb2 (46.721.209) 2824  TB:18
12/35   0:17   -1.36   1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 c2 8.Kxc2 a2 9.Kb2 b3 10.g6 fxg6 11.hxg6 Kxg6 12.e6 Kf6 13.Ka1 Kxe6 14.Kb2 (48.717.487) 2832  TB:18
13/38   0:21   -1.05   1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 c2 8.Kxc2 a2 9.Kb2 b3 10.Ka1 Kg8 11.Kb2 Kh7 12.Ka1 Kg7 13.Kb2 Kh8 14.h6 (61.193.615) 2834  TB:66
13/38   0:22   -1.05   1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 c2 8.Kxc2 a2 9.Kb2 b3 10.Ka1 Kg8 11.Kb2 Kh7 12.Ka1 Kg7 13.Kb2 Kh8 14.h6 (64.002.807) 2838  TB:67
14/39   0:29   -0.65++  1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 c2 8.Kxc2 a2 9.Kb2 b3 10.Ka1 Kg8 11.Kb2 Kh7 12.g6+ fxg6 13.e6 a1Q+ 14.Kxa1 (82.744.763) 2827  TB:294
14/39   0:43   +0.35++  1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 c2 8.Kxc2 a2 9.Kb2 Kf8 10.h6 Kg8 11.e6 Kf8 12.h7 Ke7 13.h8Q a1R (120.851.499) 2793  TB:390
14/39   0:48   +0.80   1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 c2 8.Kxc2 a2 9.Kb2 b3 10.Ka1 Kg8 11.g6 fxg6 12.h6 b2+ 13.Kxa2 g5 14.e6 (134.909.950) 2778  TB:398
14/39   0:50   +0.80   1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 c2 8.Kxc2 a2 9.Kb2 b3 10.Ka1 Kg8 11.g6 fxg6 12.h6 b2+ 13.Kxa2 g5 14.e6 (139.502.718) 2782  TB:415
15/41   1:00   +1.20++  1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 c2 8.Kxc2 a2 9.Kb2 b3 10.Ka1 Kg8 11.g6 b2+ 12.Kxa2 Kg7 13.e6 Kh6 14.Kxb2 (166.098.789) 2754  TB:731
15/41   1:10   +2.20++  1.g4 c4+ 2.Kc2 a4 3.g5 Kg7 4.Kb2 b4 5.Kb1 a3 6.Ka2 c3 7.Kb3 c2 8.Kxc2 a2 9.Kb2 Kg8 10.h6 b3 11.g6 fxg6 12.e6 Kf8 13.e7+ Ke8 14.h7 (192.402.314) 2731  TB:756
15/41   1:20   +3.18   1.g4 c4+ 2.Kc2 a4 3.g5 a3 4.h6 b4 5.g6 fxg6 6.h7 b3+ 7.Kb1 Kg7 8.e6 c3 9.h8Q+ Kxh8 10.e7 c2+ 11.Kc1 Kg7 12.e8Q a2 13.Qe7+ Kh6 14.Qf6 (217.673.402) 2719  TB:805
15/41   1:24   +3.18   1.g4 c4+ 2.Kc2 a4 3.g5 a3 4.h6 b4 5.g6 fxg6 6.h7 b3+ 7.Kb1 Kg7 8.e6 c3 9.h8Q+ Kxh8 10.e7 c2+ 11.Kc1 Kg7 12.e8Q a2 13.Qe7+ Kh6 14.Qf6 (231.234.073) 2729  TB:906
16/41   1:43   +3.58++  1.g4 c4+ 2.Kc2 a4 3.g5 a3 4.h6 b4 5.g6 fxg6 6.h7 b3+ 7.Kb1 Kg7 8.e6 c3 9.h8Q+ Kxh8 10.e7 b2 11.e8Q+ Kh7 (278.255.057) 2695  TB:1.619
16/43   1:56   +4.09   1.g4 c4+ 2.Kc2 a4 3.g5 a3 4.h6 b4 5.g6 fxg6 6.h7 b3+ 7.Kb1 Kg7 8.e6 c3 9.h8Q+ Kxh8 10.e7 c2+ 11.Kc1 a2 12.e8Q+ Kg7 13.Qe7+ Kh6 14.Qa3 (313.483.642) 2694  TB:1.767
16/43   2:03   +4.09   1.g4 c4+ 2.Kc2 a4 3.g5 a3 4.h6 b4 5.g6 fxg6 6.h7 b3+ 7.Kb1 Kg7 8.e6 c3 9.h8Q+ Kxh8 10.e7 c2+ 11.Kc1 a2 12.e8Q+ Kg7 13.Qe7+ Kh6 14.Qa3 (334.239.063) 2703  TB:2.132
17/44   2:37   +4.17   1.g4 c4+ 2.Kc2 a4 3.g5 a3 4.h6 b4 5.g6 fxg6 6.h7 b3+ 7.Kb1 Kg7 8.e6 c3 9.h8Q+ Kxh8 10.e7 c2+ 11.Kc1 a2 12.e8Q+ Kh7 13.Qe5 g5 14.Qf6 (421.231.783) 2667  TB:3.666
17/44   2:45   +4.17   1.g4 c4+ 2.Kc2 a4 3.g5 a3 4.h6 b4 5.g6 fxg6 6.h7 b3+ 7.Kb1 Kg7 8.e6 c3 9.h8Q+ Kxh8 10.e7 c2+ 11.Kc1 a2 12.e8Q+ Kh7 13.Qe5 g5 14.Qf6 (442.298.084) 2673  TB:4.061
18/46   4:02   +4.18   1.g4 c4+ 2.Kc2 a4 3.g5 a3 4.h6 b4 5.g6 fxg6 6.h7 b3+ 7.Kb1 Kg7 8.e6 c3 9.h8Q+ Kxh8 10.e7 c2+ 11.Kc1 a2 12.e8Q+ Kh7 13.Qa8 g5 14.Kb2 (633.955.102) 2614  TB:7.667
18/46   4:17   +4.18   1.g4 c4+ 2.Kc2 a4 3.g5 a3 4.h6 b4 5.g6 fxg6 6.h7 b3+ 7.Kb1 Kg7 8.e6 c3 9.h8Q+ Kxh8 10.e7 c2+ 11.Kc1 a2 12.e8Q+ Kh7 13.Qa8 g5 14.Kb2 (676.740.086) 2625  TB:8.552
18/48   6:37   +4.18   1.g4 c4+ 2.Kc2 a4 3.g5 a3 4.h6 b4 5.g6 fxg6 6.h7 b3+ 7.Kb1 Kg7 8.e6 c3 9.h8Q+ Kxh8 10.e7 c2+ 11.Kc1 a2 12.e8Q+ Kh7 13.Qa8 g5 14.Kb2 (1.016.287.766) 2557  TB:14.764
interior recognizer probes 113283873, found 4938869 egtb probes 14764, found 14764 max_depth 48
best move: g2-g4 time: 6:37.422 min  n/s: 2.557.059  CPU 99.2%   n/s(1CPU): -1751923  nodes: 1.016.287.766 TB: 14.764
Parent - - By Vasik Rajlich (Silver) Date 2007-07-26 08:00
Thanks. The one is a matter of comparing the queen endgame after 1. d6 with the k+p endgame where progress is not possible after 1. dxc6.

The second one is just very deep. For a human, it's simple: the black king can only be on g7, while the queenside pawns are in zugzwang against the white king.

Vas
Parent - - By Gaмßito (****) Date 2007-07-27 07:51
Hi Vas,

It is a pity that programs like Deep Fritz, Rybka or any other engine can just display a small number of moves of the exact depth.

Too much time to see a line of just a few moves.. I do not like that.  In this case just 6 moves. (?)

It should be exactly to the number of their depth moves, that it has exactly arrived.  Do not you believe would not be much better that Rybka can display their exact long line to the real depth that it is displaying?  Does some problem exist with this detail?

Regards,
Gambito.

5k2/5p2/8/ppp1P2P/8/1K6/6P1/8 w - - 0 1


Analysis by Deep Fritz 10:

1.g4
  -+  (-4.95)   Depth: 8/26   00:00:00  67kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7 4.Ka2 b4 5.Kb1 c4 6.Ka1 Kf8 7.Ka2 Ke7 8.e6
  -+  (-4.95)   Depth: 9/25   00:00:00  131kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7 4.Ka2 b4 5.Kb1 c4 6.Ka1 Kf8 7.Ka2 Ke7 8.e6
  -+  (-4.95)   Depth: 10/26   00:00:00  161kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7 4.Ka2 b4 5.Kb1 c4 6.Ka1 Kf8 7.Ka2 Ke7 8.e6
  -+  (-4.95)   Depth: 11/26   00:00:00  242kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7 4.Ka2 b4
  -+  (-4.95)   Depth: 12/26   00:00:01  551kN
1.g4 b4
  -+  (-5.23)   Depth: 13/30   00:00:02  949kN
1.g4 b4 2.g5
  -+  (-5.23)   Depth: 13/38   00:00:03  1737kN
1.g4 b4 2.g5 Kg7 3.Kc2 a4 4.Kb2
  -+  (-5.23)   Depth: 14/35   00:00:06  3054kN
1.g4 b4
  -+  (-5.51)   Depth: 15/31   00:00:07  3386kN
1.g4 b4 2.g5 Kg7 3.Kc4 a4 4.g6 a3 5.gxf7 a2 6.Kxc5 a1Q 7.f8Q+ Kxf8
  -+  (-6.74)   Depth: 15/50   00:00:12  7530kN
1.g4 a4+
  -+  (-7.02)   Depth: 16/44   00:00:20  9200kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7 4.Kb2 b4 5.Kc2
  -+  (-7.56)   Depth: 16/46   00:00:35  15265kN
1.g4 a4+
  -+  (-7.84)   Depth: 17/37   00:00:40  17066kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7 4.Kb2 b4 5.Kc2 c4 6.Kc1
  -+  (-7.99)   Depth: 17/52   00:01:02  33577kN
1.g4
  -+  (-7.71)   Depth: 18/45   00:01:25  50777kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7 4.Kb2 c4 5.Kc2 a3 6.Kc3
  -+  (-7.96)   Depth: 19/50   00:01:48  67619kN
1.g4
  -+  (-7.68)   Depth: 20/54   00:03:25  147mN
1.g4
  -+  (-7.40)   Depth: 20/54   00:04:13  189mN
1.g4
  -+  (-6.84)   Depth: 20/54   00:04:27  202mN
1.g4 a4+ 2.Ka2 Kg7 3.g5 b4 4.Kb2 c4 5.Kb1 c3 6.Kc2
  -+  (-3.14)   Depth: 20/59   00:05:24  251mN
1.g4 a4+
  -+  (-3.43)   Depth: 21/41   00:05:32  253mN
1.g4 a4+ 2.Ka2 Kg7 3.g5 b4 4.Kb2 c4 5.Kb1 b3 6.Kb2
  -+  (-3.90)   Depth: 21/65   00:07:09  340mN
1.g4
  -+  (-3.62)   Depth: 22/69   00:09:17  444mN
1.g4
  -+  (-3.34)   Depth: 22/69   00:09:42  465mN
1.g4
  -+  (-2.77)   Depth: 22/69   00:10:14  494mN
1.g4 a4+ 2.Ka2 Kg7 3.g5 b4 4.Kb2 c4 5.Kb1 a3 6.Ka2
  ³  (-0.48)   Depth: 22/69   00:13:56  689mN
1.g4 a4+ 2.Ka2 Kg7 3.g5 b4 4.Kb2 c4 5.Kb1 a3 6.Ka2
  ³  (-0.48)   Depth: 23/69   00:23:20  1245mN
1.g4 a4+ 2.Ka2 Kg7 3.g5 b4 4.Kb2 c4 5.Kb1 a3 6.Ka2
  ³  (-0.48)   Depth: 24/67   00:26:29  1435mN
1.g4 a4+ 2.Ka2 Kg7 3.g5 b4 4.Kb2 c4 5.Kb1 a3 6.Ka2
  ³  (-0.48)   Depth: 25/77   00:38:13  2179mN
1.g4
  =  (-0.20)   Depth: 26/96   01:23:42  5096mN
1.g4 a4+ 2.Ka2 Kg7 3.g5 b4 4.Kb2 c4 5.Kb1 a3 6.Ka2
  =  (0.00)   Depth: 26/96   01:49:53  6677mN

(, AMD 27.07.2007)

5k2/5p2/8/ppp1P2P/8/1K6/6P1/8 w - - 0 1


Analysis by Rybka 2.3.2a 32-bit :

1.Kc3 Kg7
  -+  (-4.11)   Depth: 5   00:00:00
1.h6 Kg8
  -+  (-3.98)   Depth: 5   00:00:00
1.g4 a4+ 2.Kc2
  -+  (-3.84)   Depth: 5   00:00:00
1.g4 a4+ 2.Kc2 b4
  -+  (-4.03)   Depth: 6   00:00:00  1kN
1.g4 a4+ 2.Kc2 b4 3.h6
  -+  (-4.12)   Depth: 7   00:00:00  2kN
1.g4 a4+ 2.Kc2 b4 3.g5 a3
  -+  (-4.45)   Depth: 8   00:00:00  3kN
1.Kc3 a4 2.Kd3 Kg7 3.g3
  -+  (-4.21)   Depth: 8   00:00:00  5kN
1.Kc3 a4 2.g4 b4+ 3.Kb2 c4 4.h6
  -+  (-4.46)   Depth: 9   00:00:00  7kN
1.g4 a4+ 2.Ka3 Ke7 3.g5 Kf8 4.Kb2
  -+  (-4.45)   Depth: 9   00:00:00  7kN
1.g4 a4+ 2.Ka3 Ke7 3.g5 Kf8 4.Kb2 b4
  -+  (-4.45)   Depth: 10   00:00:01  9kN
1.g4 a4+ 2.Ka3 Ke7 3.g5 Kf8 4.Kb2 b4 5.Kc2 a3 6.Kb1 c4 7.h6
  -+  (-4.45)   Depth: 11   00:00:01  11kN
1.g4 Kg8 2.g5 b4 3.Kb2 c4 4.Kc2 a4 5.Kb1 a3 6.Ka2 Kg7
  -+  (-4.48)   Depth: 12   00:00:01  23kN
1.g4 Kg8 2.g5 b4 3.Kb2 c4 4.Kc2 a4 5.Kb1 a3 6.Ka2 Kg7
  -+  (-4.48)   Depth: 13   00:00:01  30kN
1.g4 Kg8 2.g5 b4 3.Kb2 a4 4.Kc2 a3 5.Kc1 c4 6.Kb1 c3
  -+  (-4.85)   Depth: 14   00:00:01  39kN
1.g4 Kg8 2.g5 a4+ 3.Ka3 c4 4.Kb4 Kg7 5.Ka3 c3 6.Ka2 b4 7.Ka1 b3
  -+  (-5.07)   Depth: 15   00:00:01  73kN
1.g4 c4+ 2.Kc2 b4 3.g5 Kg7 4.Kb2 a4 5.Kb1 a3 6.Ka2 Kg8 7.h6 Kh7
  -+  (-5.61)   Depth: 16   00:00:08  463kN
1.Kb2 b4 2.Kb3 Ke7 3.g3 Ke8 4.g4 Kf8 5.g5 Kg7 6.Kb2 a4 7.Ka2 c4
  -+  (-5.08)   Depth: 16   00:00:35  2056kN
1.Kb2 b4 2.Kb3 Kg7 3.g3 Kh7 4.g4 Kh6 5.Kc4 a4 6.Kd3 a3 7.Kc2 c4
  -+  (-5.09)   Depth: 17   00:00:35  2102kN
1.Kb2 Kg7 2.g3 a4 3.Ka3 Kh6 4.g4 c4 5.Kb4 Kg5 6.Ka3 c3 7.Ka2 b4
  -+  (-6.81)   Depth: 18   00:01:04  3889kN
1.Kb2 Kg7 2.g3 a4 3.Ka3 Kh6 4.g4 c4 5.Kb4 Kg5 6.Ka3 c3 7.Ka2 b4
  -+  (-7.13)   Depth: 19   00:03:40  12656kN
1.Kb2 Kg7 2.g3 a4 3.Ka3 c4 4.Kb4 Kh6 5.g4 Kg5 6.Kc3 a3 7.Kc2 b4
  -+  (-9.60)   Depth: 20   00:10:10  37141kN

(, AMD 27.07.2007)
Parent - - By Vasik Rajlich (Silver) Date 2007-07-28 08:11
You can get a longer variation by selecting "Display PV Tips".

For Rybka 3, I may add another option to display the "minimax" variation. These do a better job explaining the search scores.

Vas
Parent - By Gaмßito (****) Date 2007-07-30 08:07
Thanks Vas, 
 
Sure, a new option to display the ''minimax'' variation would be great in the new Rybka. I think that everything new to display more ''thinking information'' is very welcome. I think you should make it.

Regards,
Gambito.
Parent - - By sebakaw (*) Date 2007-07-26 08:14 Edited 2007-07-26 08:19
In the first position my Rybka with 5-men TB plays dxc6 like Your but in the second position is good:

Analysis by Rybka 2.3.2a mp 32-bit :
Intel Core2Duo 2.4GHz 2GB RAM

1.Kc3 Kg7
  -+  (-4.11)   Depth: 5   00:00:00
1.h6 Kg8
  -+  (-3.98)   Depth: 5   00:00:00
1.g4 a4+ 2.Kb2
  -+  (-3.84)   Depth: 5   00:00:00
1.g4 a4+ 2.Kb2 b4
  -+  (-4.03)   Depth: 6   00:00:00  2kN
1.g4 a4+ 2.Ka3 Kg7 3.g5
  -+  (-4.03)   Depth: 7   00:00:00  3kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7
  -+  (-4.15)   Depth: 8   00:00:00  4kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7 4.Kb2
  -+  (-4.14)   Depth: 9   00:00:00  6kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7 4.Kb2 b4 5.Kc2 c4
  -+  (-4.15)   Depth: 10   00:00:00  9kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 c4 4.Kb4 Kg7 5.Ka3
  -+  (-4.52)   Depth: 11   00:00:00  14kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kh7 4.Kb2 b4 5.Kc1 Kh8
  -+  (-4.52)   Depth: 12   00:00:00  20kN
1.g4 a4+ 2.Ka3 Kg8 3.g5 Kg7 4.Kb2 b4 5.Kc2 Kg8 6.Kd3 a3
  -+  (-4.52)   Depth: 13   00:00:00  31kN
1.g4 Kg7 2.g5 b4 3.Kc2 a4 4.Kb2 a3+ 5.Kb3 c4+ 6.Ka2 Kh7 7.Ka1
  -+  (-5.02)   Depth: 14   00:00:00  59kN
1.g4 Kg7 2.g5 b4 3.Kc2 c4 4.Kb2 c3+ 5.Kb3 a4+ 6.Kc2 Kg8 7.h6 Kh7
  -+  (-5.07)   Depth: 15   00:00:01  113kN
1.g4 Kg7 2.g5 b4 3.Kc2 c4 4.Kb1 c3 5.Kc1 b3
  -+  (-5.12)   Depth: 16   00:00:02  174kN
Parent - By ernest (****) Date 2007-07-26 11:17

> In the first position my Rybka with 5-men TB plays dxc6 like Your but in the second position is good:
>


Yes it plays g4, but look at the evaluation score...
Parent - By Vinvin (***) Date 2007-07-26 23:57
The yace engine is certainly the reference in pawn endgames. All programmers sould buy the endgame code from Dieter Bürssners :)
Parent - By Gaмßito (****) Date 2007-07-27 02:56
Deep Fritz 10 can solve the first problem easily. 

Look:

3k4/8/1Pp5/3P2p1/4K1Pp/8/8/8 w - - 0 1


Analysis by Deep Fritz 10:

1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kg2 Kb8 5.Kh3 Kc8
  +-  (4.89)   Depth: 9/15   00:00:00  13kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kg2 Kb8 5.Kh2 Kc8
  +-  (4.95)   Depth: 10/17   00:00:00  25kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kf1 Kb8 5.Ke2 Kc8
  +-  (4.89)   Depth: 11/18   00:00:00  36kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kf1 Kb8 5.Ke2 Kc8
  +-  (4.89)   Depth: 12/26   00:00:00  71kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kf1 Kb8 5.Ke2 Kc8
  +-  (4.89)   Depth: 13/23   00:00:00  138kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kf1 Kd8
  +-  (4.89)   Depth: 14/27   00:00:00  236kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kf1 Kd8 5.Kg1 Kc8
  +-  (4.89)   Depth: 15/29   00:00:00  413kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kf1 Kd8 5.Kg1 Kc8
  +-  (4.89)   Depth: 16/29   00:00:01  679kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kf1 Kd8 5.Kg1 Kc8 6.Kh2 Kd8
  +-  (4.88)   Depth: 17/32   00:00:06  1259kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kf1 Kd8 5.Kg1 Kc8 6.Kh2 Kd8
  +-  (4.88)   Depth: 18/43   00:00:08  2555kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kf1 Kd8 5.Kg1 Kc8 6.Kh2 Kb8
  +-  (4.88)   Depth: 19/42   00:00:11  5430kN
1.dxc6 Kc8
  +-  (4.52)   Depth: 20/51   00:00:15  9963kN
1.dxc6 Kc8 2.Kf3 Kb8 3.Kf2 Kc8 4.Kf1 Kd8 5.Kg1 Kc8 6.Kh2 Kb8 7.Kg2 Kc8
  ±  (1.21)   Depth: 20/58   00:00:17  12447kN
1.d6
  ±  (1.21)   Depth: 20/58   00:00:24  18421kN
1.d6
  +-  (1.87)   Depth: 20/58   00:00:24  18983kN
1.d6
  +-  (2.18)   Depth: 20/58   00:00:24  19170kN
1.d6
  +-  (2.80)   Depth: 20/58   00:00:24  19285kN
1.d6 Kc8 2.Kf5 h3 3.Ke6 h2 4.d7+ Kb7 5.d8Q h1Q 6.Qc7+ Ka6 7.Qa7+ Kb5 8.b7 Qe4+ 9.Kf7 Qd5+ 10.Kg7 Qe5+ 11.Kg6 Qe6+ 12.Kxg5 Qe5+ 13.Kg6
  +-  (4.01)   Depth: 20/59   00:00:30  24727kN
1.d6 Kc8 2.Kf5 h3 3.Ke6 h2 4.d7+ Kb7 5.d8Q h1Q 6.Qc7+ Ka6 7.Qa7+ Kb5 8.b7 Qe4+ 9.Kf7 Qd5+ 10.Kg7 Qe5+
  +-  (4.01)   Depth: 21/55   00:00:47  42805kN
1.d6 Kc8
  +-  (3.73)   Depth: 22/59   00:01:27  83173kN
1.d6 Kc8 2.Kf5
  +-  (3.73)   Depth: 22/70   00:01:53  110mN
1.d6 Kc8
  +-  (3.45)   Depth: 23/64   00:03:18  198mN
1.d6 Kc8 2.Kf5 h3 3.Ke6 h2 4.d7+ Kb7 5.d8Q h1Q 6.Qc7+ Ka6 7.b7 Qh6+ 8.Kf5
  +-  (3.41)   Depth: 23/64   00:04:00  244mN
1.d6
  +-  (3.69)   Depth: 24/96   00:04:40  294mN

(, AMD 26.07.2007)

In the second position, the program sees g4 as the best move for white but with a negative score. It cannot see the win for white. It needs to calculate deeper to see that white has a win here. I will give it enough time and see what happens.

Regards,
Gambito.
Up Topic Rybka Support & Discussion / Rybka Discussion / Rybka's difficulties with pawn endgames.

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill