Testposition Endgame

General discussion about computer chess...
Post Reply
Eduard Nemeth
Posts: 632
Joined: Thu Jun 10, 2010 12:50 am

Re: Testposition Endgame

Post by Eduard Nemeth » Mon Jul 19, 2010 1:09 am

Hello Robert!

Perhaps you are right, perhaps this position is draw. But I am not sure!?
Reason: This position have many possible "zugzwang" motives with pawns!

I does an practical experiment. I have let outplay this position between
"Crab 1.0 w32 beta and Houdini 1.03a w32 with Ponder ON.
Both programs used 1 core.

CPU=Core Duo 2,66 GHz
GUI=Fritz 10
PB=ON
Hash=256 MB, 256 MB
Timecontrol=10 Minutes (for this endgame)

Houdini 1.03a lost it.

8/p4p2/P1k1pn1P/3p4/8/1K2B3/8/8 w - - 0 1

PGN:

[Event "Core 2 Duo 2.66 GHz, Blitz:10'"]
[Site "Privat"]
[Date "2010.07.19"]
[Round "1"]
[White "Crab 1.0 Beta"]
[Black "Houdini 1.03a w32 1_CPU"]
[Result "1-0"]
[Annotator "2.26;0.66"]
[SetUp "1"]
[FEN "8/p4p2/P1k1pn1P/3p4/8/1K2B3/8/8 w - - 0 1"]
[PlyCount "79"]
[TimeControl "600"]

{Intel(R) Core(TM)2 Duo CPU E6750 @ 2.66GHz 2660 MHz W=26.8 ply; 1.429kN/
s B=22.6 ply; 1.952kN/s} 1. Bxa7 {Beide letzter Buchzug 2.26/23 17} Nh7 {
0.66/22 14} 2. Bd4 {1.85/25 10} f6 {0.72/23 15} 3. Kc3 {(Lc5) 1.97/25 0} e5 {
0.64/21 14} 4. Bc5 {1.97/24 1} Ng5 {(d4+) 0.63/21 11} 5. Kc2 {(Le3) 2.06/24 43}
d4 {0.63/20 11} 6. Kb3 {(Kd2) 1.83/22 2} f5 {0.63/20 11} 7. Kc2 {1.19/22 13} f4
{0.46/20 13} 8. Kd3 {(Kd1) 1.52/23 18} Nh7 {(f3) 0.57/18 17} 9. Bd6 {
(Kd2) 1.63/23 29} Nf6 {(Sg5) 0.35/19 15} 10. Ke2 {(Le7) 3.37/23 18} e4 {
0.33/19 16} 11. Bxf4 {5.29/25 0} d3+ {2.88/20 10} 12. Kd2 {6.32/28 2} Kb6 {
5.08/22 73} 13. Be5 {9.61/29 0} Nh7 {6.00/22 42} 14. Ke3 {9.69/32 0} Ka7 {
(Kxa6) 6.05/22 28} 15. Bc3 {11.07/28 19} Kxa6 {6.16/24 0} 16. Kxe4 {10.74/23 15
} Kb5 {6.27/25 0} 17. Bd2 {10.90/22 16} Kc4 {6.41/24 2} 18. Kf5 {10.74/24 9}
Nf8 {6.41/25 0} 19. Kf6 {88.94/28 11} Nd7+ {6.46/23 0} 20. Ke6 {
(Kg7) 89.14/28 10} Nc5+ {6.74/22 9} 21. Kf5 {113.69/28 6} Nb3 {9.25/23 43} 22.
Be3 {(h7) 89.50/31 0} d2 {9.17/19 6} 23. Bxd2 {134.46/28 9} Nxd2 {#18/1 0} 24.
h7 {#17/1 0} Kd5 {#17/1 0} 25. h8=Q {#16/1 0} Nc4 {#16/1 0} 26. Qd8+ {#15/1 0}
Nd6+ {#15/1 0} 27. Kf6 {#14/1 0} Kc5 {#14/1 0} 28. Qe7 {#13/1 0} Kd5 {#13/1 0}
29. Qe5+ {#12/1 0} Kc6 {#12/1 0} 30. Qd4 {#11/1 0} Nb5 {#10/1 0} 31. Qc4+ {
#10/1 0} Kb6 {#10/1 0} 32. Ke6 {#9/1 0} Nc7+ {#8/1 0} 33. Kd7 {#8/1 0} Na6 {
#8/1 0} 34. Kd6 {#7/1 0} Ka5 {#7/1 0} 35. Qb3 {#6/1 0} Nb4 {#6/1 0} 36. Kc5 {
#5/1 0} Na6+ {#5/1 0} 37. Kc4 {#4/1 0} Nb4 {#4/1 0} 38. Qxb4+ {#3/1 0} Ka6 {
#3/1 0} 39. Qb8 {#2/1 0} Ka5 {#2/1 0} 40. Qa7# {#1/1 0} 1-0

ED.
Peace be with you!
Top
Howard E
Posts: 46
Joined: Fri Jun 11, 2010 3:57 am
Real Name: Howard Exner

Re: Testposition Endgame

Post by Howard E » Mon Jul 19, 2010 1:15 am

1. Nxg5 hxg5 2. Bxg5 Nxb3 3. h4 Bc5 4. Be3 white pose problems for black
Top
Eduard Nemeth
Posts: 632
Joined: Thu Jun 10, 2010 12:50 am

Re: Testposition Endgame

Post by Eduard Nemeth » Mon Jul 19, 2010 3:53 am

Howard E wrote:1. Nxg5 hxg5 2. Bxg5 Nxb3 3. h4 Bc5 4. Be3 white pose problems for black
Perhaps only?!

But I was blind, sorry. :shock:

Easy wins this another line after 3...Bc5?!

1. Nxg5 hxg5 2. Bxg5 Nxb3 3. h4 Bc5 4. dxc5 Nxc5+ 5. Kc2 b3+ and now 6. Kb1!
After 6. Kb1 black can not given 6. Nd3+ and not used the pawn on e5! 8-)

6. Kb1! ++-

Here an game between Crab 1.0 beta and Houdini 1.03a, with the same conditions how in experimental first game,
but now after 6. Kb1!

8/p2k1p2/P3pP2/2npP1B1/6PP/1p6/8/1K6 b - - 0 6

[Event "Core 2 Duo 2.66 GHz, Blitz:30'"]
[Site "Privat"]
[Date "2010.07.19"]
[Round "1"]
[White "Crab 1.0 beta"]
[Black "Houdini 1.03a w32 1_CPU"]
[Result "1-0"]
[Annotator "2.96;1.17"]
[SetUp "1"]
[FEN "8/p2k1p2/P3pP2/2npP1B1/6PP/1p6/8/1K6 b - - 0 6"]
[PlyCount "24"]
[TimeControl "1800"]

{Intel(R) Core(TM)2 Duo CPU E6750 @ 2.66GHz 2660 MHz W=31.8 ply; 1.247kN/
s B=24.4 ply; 1.751kN/s} 6... Ne4 {Beide letzter Buchzug 1.17/22 153} 7. Bf4 {
2.96/28 64} d4 {1.06/23 15} 8. Kb2 {3.19/28 33} Nc5 {1.06/24 22} 9. Bg3 {
(Lg5) 3.53/28 33} Ke8 {1.30/20 32} 10. Bh2 {(h5) 3.83/29 37} Nxa6 {
(Sd3+) 2.19/23 165} 11. h5 {3.93/27 65} Nb4 {(d3) 2.19/23 0} 12. Kxb3 {
(h6) 3.93/26 74} d3 {2.18/20 11} 13. Kc3 {(Lf4) 3.93/31 29} Kf8 {2.46/22 35}
14. Bg1 {5.49/31 4} Kg8 {2.87/25 79} 15. Bc5 {7.67/33 0} Nc6 {2.87/25 28} 16.
Bd6 {11.87/31 33} Nd4 {(d2) 3.67/24 0} 17. Kxd3 {13.89/35 60} Nf3 {5.71/28 150}
18. Ke3 {adjud. 21.37/37 0} 1-0

Timecontrol was here now 30 Minutes, but white won very easy and fast.

ED.
Peace be with you!
Top
User avatar
Robert Houdart
Posts: 180
Joined: Thu Jun 10, 2010 4:55 pm
Contact:
Contact Robert Houdart

Re: Testposition Endgame

Post by Robert Houdart » Mon Jul 19, 2010 12:14 pm

Ed, in my opinion these short blitz games are not very useful. Black cannot afford a single mistake to reach a draw, so the only possible outcome of these fast games is that White wins.

Your 6.Kb1 is a very good attempt, but Black can then play d4 to prevent Be3. Again neither Houdini nor Stockfish 1.8 can find a win, even at very large depths.

8/p2k1p2/P3pP2/2npP1B1/6PP/1p6/8/1K6 b - - 0 6
40-minute analysis from Houdini (4 cores), no change in evaluation between depth 24 and 34:

Code: Select all

Houdini_64 DEV:
 24/80	01:11,964	 553.800.269	7.715.000	-1,59	6. ... d4 7.Kb2 Ke8 8.h5 Kf8 9.Bd2 Nxa6 10.Kxb3 Nb8 11.Bb4+ Kg8 12.Kc4 Nc6 13.Bd6 a5 14.g5 a4 15.Kd3 a3 16.Bxa3 Nxe5+ 17.Ke4 Nc4 18.Bc1 d3 19.Kxd3 Ne5+ 20.Ke4 Nd7 21.Be3 Kh7 22.Kf4 Kg8 23.Bf2 Kh7 24.Bd4 Nf8 25.Ke4 Nd7 26.Be3 Kg8 27.Kd3 Kh7 28.Ke4 Kg8
 25/80	01:26,211	 663.771.721	7.716.000	-1,59	6. ... d4 7.Kb2 Ke8 8.h5 Kf8 9.Bd2 Nxa6 10.Kxb3 Nb8 11.Bb4+ Kg8 12.Kc4 Nc6 13.Bd6 a5 14.g5 a4 15.Kd3 a3 16.Bxa3 Nxe5+ 17.Ke4 Nc4 18.Bc1 d3 19.Kxd3 Ne5+ 20.Ke4 Nd7 21.Be3 Kh7 22.Kf4 Kg8 23.Bf2 Kh7 24.Bd4 Nb8 25.Bc3 Nd7 26.Bb4 Nb6 27.Ke4 Nd7 28.Bc3 Nb6 29.Be1 Nd7 30.Bf2 Kg8 31.Bd4 Kh7 32.Bf2 Kg8
 26/80	01:39,690	 767.341.860	7.711.000	-1,59	6. ... d4 7.Kb2 Ke8 8.h5 Kf8 9.Bd2 Nxa6 10.Kxb3 Nb8 11.Bb4+ Kg8 12.Kc4 Nc6 13.Bd6 a5 14.g5 a4 15.Kd3 a3 16.Bxa3 Nxe5+ 17.Ke4 Nc4 18.Bc1 d3 19.Kxd3 Ne5+ 20.Ke4 Nd7 21.Be3 Kh7 22.Kf4 Kg8 23.Bf2 Kh7 24.Bd4 Nb8 25.Bc3 Nd7 26.Ke4 Nb6 27.Be1 Nd7 28.Bf2 Kg8 29.Bd4 Kh7 30.Bf2 Kg8
 27/80	03:41,394	1.749.111.279	7.907.000	-1,62	6. ... d4 7.Kb2 Kc6 8.h5 Kb5 9.h6 Nd3+ 10.Kxb3 Nxe5 11.h7 Ng6 12.Bh6 Kxa6 13.Kc4 e5 14.Kc5 Kb7 15.Bd2 Kc7 16.Be1 Kd8 17.Kd5 Kd7 18.Bg3 d3 19.Be1 Kc7 20.Kc5 Kd7 21.Kd5 Kc7
 28/80	04:00,062	1.900.577.671	7.923.000	-1,62	6. ... d4 7.Kb2 Kc6 8.h5 Kb5 9.h6 Nd3+ 10.Kxb3 Nxe5 11.h7 Ng6 12.Bh6 Kxa6 13.Kc4 e5 14.Kc5 Kb7 15.Bd2 Kc7 16.Be1 Kd8 17.Kd5 Kd7 18.Bg3 d3 19.Be1 Kc7 20.Ba5+ Kb7 21.Kc5 Ka6 22.Bc3 Kb7 23.Bb4 Kc7 24.g5 Kd7 25.Kd5 Kc7 26.Bc3 Kd7 27.Ba5 Kc8 28.Kc5 Kb7 29.Be1 Nh8 30.Kd5 Ng6 31.Ke4 Kc6 32.Kxd3 Kd5 33.Bc3 e4+ 34.Ke3 a6 35.Be1 Ke5 36.Bd2 Kf5
 29/80	05:07,984	2.453.224.574	7.970.000	-1,62	6. ... d4 7.Kb2 Kc6 8.h5 Kb5 9.h6 Nd3+ 10.Kxb3 Nxe5 11.h7 Ng6 12.Bh6 Kxa6 13.Kc4 e5 14.Kc5 Kb7 15.Bd2 Kc7 16.Be1 Kd8 17.Kd5 Kd7 18.Bg3 d3 19.Be1 Kc7 20.Ba5+ Kb7 21.Kc5 Kc8 22.Bc3 Kd8 23.Kd5 Kc8 24.Bd2 Kd7 25.Ba5 Kc8 26.Bc3 Kb7 27.Kc5 Kc8 28.Be1 Kd7 29.Kd5 Kc7
 30/80	06:54,998	3.349.464.304	8.074.000	-1,62	6. ... d4 7.Kb2 Kc6 8.h5 Kb5 9.h6 Nd3+ 10.Kxb3 Nxe5 11.h7 Ng6 12.Bh6 Kxa6 13.Kc4 e5 14.Kc5 Kb7 15.Bd2 Kc7 16.Be1 Kd8 17.Kd5 Kd7 18.Bg3 d3 19.Be1 Kc7 20.Ba5+ Kb7 21.Kc5 Kc8 22.Bc3 Kd8 23.Kd5 Kc8 24.Bb4 Kd7 25.Bd2 Kc7 26.Be1 Kb6 27.Ke4 Kc5 28.Bc3 Kc6 29.Bb4 Kd7 30.Kd5 Kc8 31.Kd6 Kd8 32.Ba5+ Kc8 33.Kc5 Kb7 34.Be1 Kc7 35.Bc3 Kb7 36.Bb4 Kc7 37.Be1 Kd7 38.Kd5 Kc7
 31/80	09:48,189	4.811.865.115	8.183.000	-1,62	6. ... d4 7.Kb2 Kc6 8.h5 Kb5 9.h6 Nd3+ 10.Kxb3 Nxe5 11.h7 Ng6 12.Bh6 Kxa6 13.Kc4 e5 14.Kc5 Kb7 15.Bd2 Kc7 16.Be1 Kd8 17.Kd5 Kd7 18.Bg3 d3 19.Be1 Kc7 20.Ba5+ Kb7 21.Kc5 Kc8 22.Bc3 Kd8 23.Kd5 Kc8 24.Bd2 Kd7 25.Ba5 Kc8 26.Bc3 Kb7 27.Kc5 Kc7 28.g5 Kb7 29.Ba5 Nh8 30.Be1 Ng6 31.Bc3 Kc7 32.Bd2 Kd7 33.Kd5 Kc7 34.Ba5+ Kb7 35.Kc5 Nh8
 32/80	14:21,121	7.155.453.563	8.311.000	-1,62	6. ... d4 7.Kb2 Kc6 8.h5 Kb5 9.h6 Nd3+ 10.Kxb3 Nxe5 11.h7 Ng6 12.Bh6 Kxa6 13.Kc4 e5 14.Kc5 Kb7 15.Bd2 Kc7 16.Be1 Kd8 17.Kd5 Kd7 18.Bg3 d3 19.Be1 Kc7 20.Ba5+ Kb7 21.Kc5 Kc8 22.Bc3 Kb7 23.Bb4 Kc8 24.Ba5 Kd7 25.Kd5 Kc8 26.Bc3 Kb7 27.Kc5 Kc7 28.Bd2 Kd7 29.Kd5 Kc7 30.Be1 Kb6 31.Ke4 Kc5 32.Bc3 Kc6 33.Bb4 Kd7 34.Kd5 Kc8 35.Kd6 Kd8 36.Ba5+ Kc8 37.Kc5 Kd7
 33/80	21:16,460	10.717.474.914	8.397.000	-1,62	6. ... d4 7.Kb2 Kc6 8.h5 Kb5 9.h6 Nd3+ 10.Kxb3 Nxe5 11.h7 Ng6 12.Bh6 Kxa6 13.Kc4 e5 14.Kc5 Kb7 15.Bd2 Kc7 16.Be1 Kd8 17.Kd5 Kd7 18.Bg3 d3 19.Be1 Kc7 20.Ba5+ Kb7 21.Kc5 Kc8 22.Bc3 Kb7 23.Bb4 Kc8 24.g5 Kd7 25.Kd5 Kc7 26.Ba5+ Kb7 27.Kc5 Nh8 28.Kd5 Ng6 29.Kc5
 34/83	38:01,550	19.379.442.532	8.494.000	-1,63	6. ... d4 7.Kb2 Kc6 8.h5 Kb5 9.h6 Nd3+ 10.Kxb3 Nxe5 11.h7 Ng6 12.Bh6 Kxa6 13.Kc4 e5 14.Kc5 Kb7 15.Bd2 Kc7 16.Be1 Kd8 17.Kd5 Ke8 18.Bg3 d3 19.Be1 Kd8 20.Bc3 Kc8 21.Bd2 Kb7 22.Kc5 Kc8 23.g5 Kb7 24.Bc3 Nh8 25.Ba5 Ka6 26.Bb4 Kb7 27.Kd5 Ng6 28.Bd2 Kb6 29.Kc4 a5 30.Kxd3 Kb5 31.Ke4 a4 32.Bc1 Kb4 33.Kd5 a3 34.Bd2+ Kb3
20-minute analysis from SF 1.8, no score change from depth 33 to 42:

Code: Select all

Stockfish-18-64-ja:
  33	01:17,206	 561.533.075	7.274.781	-3,03	6. ... d4 7.Kb2 Nd3+ 8.Kxb3 Nxe5 9.h5 Kc6 10.h6 Ng6 11.h7 Kb5 12.Bh6 e5 13.Bg7 e4 14.Bf8 e3 15.Bd6 Kxa6 16.Kc4 e2 17.Bg3 Kb6 18.Kxd4 Kc6 19.Kd3 a5 20.Kxe2 a4 21.Kd3 Kd5 22.Kc3 Ke6 23.Bh4 Ke5 24.Kb4 Kf4 25.Kxa4 Kxg4 26.Bf2 Kf5 27.Kb4
  34	01:37,548	 706.599.692	7.244.872	-3,15	6. ... d4 7.Kb2 Nd3+ 8.Kxb3 Nxe5 9.h5 Kc6 10.h6 Ng6 11.h7 Kb5 12.Bh6 Kxa6 13.Kc4 Kb7 14.Kxd4 Kc6 15.Bd2 Nh8 16.g5 Ng6 17.Bb4 a6 18.Bf8 a5 19.Be7 Kd7 20.Ke4 Nh8 21.Bc5 Kc6 22.Ba3 Ng6 23.Be7
  35	01:49,356	 793.792.924	7.259.858	-3,15	6. ... d4 7.Kb2 Nd3+ 8.Kxb3 Nxe5 9.h5 Kc6 10.h6 Ng6 11.h7 Kb5 12.Bh6 Kxa6 13.Kc4 Kb7 14.Kxd4 Kc6 15.Bd2 Kb5 16.Bf4
  36	02:29,901	1.090.928.292	7.278.435	-3,11	6. ... d4 7.Kb2 Nd3+ 8.Kxb3 Nxe5 9.h5 Kc6 10.h6 Ng6 11.h7 Kb5 12.Bh6 Kxa6 13.Kc4 Kb7 14.Kxd4 a5 15.Kc5 a4 16.Bc1 Kc7 17.Ba3 Kd7 18.Kc4 Nh8 19.Be7 Kc6 20.Bc5 Kd7 21.Ba3
  37	03:05,048	1.332.484.363	7.201.372	-3,03	6. ... d4 7.Kb2 Nd3+ 8.Kxb3 Nxe5 9.h5 Kc6 10.h6 Ng6 11.h7 Kb5 12.Bh6 Kxa6 13.Kc4 Kb7 14.Kxd4 Kc6 15.Ke4 a5 16.Bf8 Kb6 17.g5 Kc6 18.Be7 a4 19.Kd4 Kd7 20.Bb4 Nh8 21.Bf8 Ng6 22.Kc4 Kc6
  38	04:25,826	1.914.581.536	7.202.846	-3,03	6. ... d4 7.Kb2 Nd3+ 8.Kxb3 Nxe5 9.h5 Kc6 10.h6 Ng6 11.h7 Kb5 12.Bh6 Kxa6 13.Kc4 Kb7 14.Kxd4 Kc6 15.Ke4 a5 16.Bf8 Kb6 17.g5 Kc6 18.Be7 a4 19.Kd4 Kd7 20.Bb4 Nh8 21.Bf8 Ng6
  39	06:28,441	2.805.702.722	7.223.295	-3,07	6. ... d4 7.Kb2 Nd3+ 8.Kxb3 Nxe5 9.h5 Kc6
  40	06:51,763	2.967.774.784	7.207.770	-2,86	6. ... d4 7.Kb2 Nd3+ 8.Kxb3 Nxe5 9.h5 Kc6 10.h6 Ng6 11.h7 Kb5 12.Bh6 Kxa6 13.Kc4 Kb7 14.Kxd4 Kc6 15.Bd2 Kb5 16.Bf4 a5 17.Bd6 Kc6 18.Bc5 Kb5 19.g5 Kc6 20.Kc4 Kd7 21.Bf8 a4 22.Kd4
  41	08:17,313	3.574.370.470	7.187.615	-2,86	6. ... d4 7.Kb2 Nd3+ 8.Kxb3 Nxe5 9.h5 Kc6 10.h6 Ng6 11.h7 Kb5 12.Bh6 Kxa6 13.Kc4 Kb7 14.Kxd4 Kc6 15.Bd2 Kb5 16.Bf4 a5 17.Bd6 Kc6 18.Bc5 Kb5 19.g5 Kc6 20.Kc4 Kd7 21.Bf8 a4 22.Kd4
  42	16:04,425	6.915.158.218	7.170.351	-2,82	6. ... d4 7.Kb2 Nd3+ 8.Kxb3 Nxe5 9.h5 Kc6 10.h6 Ng6 11.h7 Kb5 12.Bh6 Kxa6 13.Kc4 Kb6 14.Kxd4 Kc6 15.Bc1 a5 16.g5 a4 17.Ba3 Kd7 18.Ke4 Kc6 19.Be7 Kd7 20.Kd4 Nh8 21.Bb4 Kc6 22.Bf8 Kd7
White has a very clear advantage, but I still haven't seen a winning line.
Top
User avatar
Robert Houdart
Posts: 180
Joined: Thu Jun 10, 2010 4:55 pm
Contact:
Contact Robert Houdart

Re: Testposition Endgame

Post by Robert Houdart » Mon Jul 19, 2010 12:35 pm

Howard,
Howard E wrote:1. Nxg5 hxg5 2. Bxg5 Nxb3 3. h4 Bc5 4. Be3 white pose problems for black
After 4...Bxd4 5.Bxd4 Nxd4 6.Kxd4 Ke8 we're heading towards a queen ending.

4k3/p4p2/P3pP2/3pP3/1p1K2PP/8/8/8 w - - 1 7

Here's the final line of a 5-minute Houdini analysis:

Code: Select all

 38/82	04:16,999	2.146.951.137	8.360.000	+1,18	7.h5 Kf8 8.g5 Kg8 9.g6 fxg6 10.hxg6 Kf8 11.Kd3 Kg8 12.g7 Kf7 13.Kd4 Kg8 14.Kc5 b3 15.Kd6 b2 16.Ke7 b1Q 17.f7+ Kxg7 18.f8Q+ Kg6 19.Kxe6 Qe4 20.Qf6+ Kh7 21.Qe7+ Kh6 22.Qb7 Qg6+ 23.Kd7 Qg7+ 24.Kd6 Qf8+ 25.Kxd5 Kg5 26.Qc6 Qa3 27.Qf6+ Kg4 28.Qg6+ Kf4 29.Qe4+ Kg5 30.Qg2+ Kf4 31.Qd2+ Kg4 32.e6 Qb3+ 33.Ke5 Qb8+ 34.Ke4 Qb1+ 35.Qd3 Qh1+ 36.Kd4 Qh8+ 37.Kc4 Qc8+
This might be a draw or a win, it should be analyzed with an engine that has the KQPKQ tablebase loaded.
Top
Eduard Nemeth
Posts: 632
Joined: Thu Jun 10, 2010 12:50 am

Re: Testposition Endgame

Post by Eduard Nemeth » Mon Jul 19, 2010 3:18 pm

Robert Houdart wrote:Ed, in my opinion these short blitz games are not very useful. Black cannot afford a single mistake to reach a draw, so the only possible outcome of these fast games is that White wins.
Dear Robert!

My last experimental game was not blitz, it was (ONLY FOR THIS ENDING!) 30 Minutes with Ponder ON! In a real game you have in such last endings on timecontrol 120 Minutes too, not more time. But OK, say me, how many time I should set as Timecontrol, that Houdini can hold here an draw?
Robert Houdart wrote:Your 6.Kb1 is a very good attempt, but Black can then play d4 to prevent Be3. Again neither Houdini nor Stockfish 1.8 can find a win, even at very large depths.

[---]
White has a very clear advantage, but I still haven't seen a winning line.
The Problem of this position is eg. that we have here to fighting with any "zugzwang" motives with pawns, and for all Engines is not easy also to hold a draw (for a Human GM the case is the same).

Of course Houdini loses the next experimental-game (after d4) again.

CPU=Core 2 Duo 2.66 GHz
GUI=Fritz 10
PB=ON
Timecontrol=30 Minutes --> for this ending only!
Both Engines used 1 core with 256 MB Hash

8/p2k1p2/P3pP2/2n1P1B1/3p2PP/1p6/8/1K6 w - - 0 6

PGN:

[Event "Core 2 Duo 2.66 GHz, Rapid 30'"]
[Site "Privat"]
[Date "2010.07.19"]
[Round "1"]
[White "Crab 1.0 beta"]
[Black "Houdini 1.03a w32 1_CPU"]
[Result "1-0"]
[Annotator "3.11;1.31"]
[SetUp "1"]
[FEN "8/p2k1p2/P3pP2/2n1P1B1/3p2PP/1p6/8/1K6 w - - 0 6"]
[PlyCount "49"]
[TimeControl "1800"]

{Intel(R) Core(TM)2 Duo CPU E6750 @ 2.66GHz 2660 MHz W=30.6 ply; 1.255kN/
s B=26.0 ply; 1.836kN/s} 6. Kb2 {Beide letzter Buchzug 3.11/27 48} Kc6 {
1.31/21 100} 7. h5 {3.07/28 50} Nd3+ {1.28/23 4} 8. Kxb3 {2.92/31 40} Nxe5 {
1.14/24 0} 9. h6 {3.01/31 54} Kb5 {(Sg6) 1.08/24 0} 10. h7 {3.07/30 105} Ng6 {
1.24/25 44} 11. Bh6 {3.03/32 25} Kxa6 {1.60/23 59} 12. Kc4 {3.03/31 0} e5 {
1.60/24 60} 13. Kc5 {2.70/29 0} Kb7 {1.60/25 70} 14. Bd2 {2.72/31 115} Kc7 {
1.60/26 0} 15. Ba5+ {(Le1) 3.05/30 69} Kc8 {1.61/25 45} 16. Kc6 {
(Le1) 2.88/30 10} d3 {(Sh8) 1.61/24 38} 17. Bd2 {(Kd5) 3.11/29 42} Kd8 {
1.61/25 37} 18. Bc3 {(Kd5) 2.74/29 45} Ke8 {(Kc8) 1.61/25 44} 19. Kd5 {
3.55/28 50} Kd8 {2.28/25 94} 20. g5 {3.91/29 0} Kd7 {2.31/25 54} 21. Ba5 {
4.24/29 20} a6 {(e4) 2.31/25 35} 22. Bc3 {4.44/29 47} Kc7 {(e4) 2.31/25 7} 23.
Bxe5+ {4.52/29 74} Kb6 {(Kd7) 2.33/26 0} 24. Bc3 {4.58/28 26} a5 {2.52/25 53}
25. Ke4 {5.67/30 0} a4 {(Kb5) 2.54/24 21} 26. Kxd3 {6.68/29 40} a3 {
(Kc5) 5.06/25 132} 27. Kc2 {(Kc4) 20.36/26 39} Kc5 {(Kc6) 5.11/26 21} 28. Kb3 {
26.74/25 25} a2 {5.21/27 0} 29. Kxa2 {28.72/26 36} Kc4 {(Kd5) 7.05/28 111} 30.
Bb2 {(Le1) adjud. 93.41/34 35} 1-0

ED.
Peace be with you!
Top
Eduard Nemeth
Posts: 632
Joined: Thu Jun 10, 2010 12:50 am

Re: Testposition Endgame

Post by Eduard Nemeth » Mon Jul 19, 2010 7:24 pm

Don't worry Robert (and others).

1. Perhaps this position is at last won for white?
2. Crab 1.0 Beta is perhaps in such positions an "Gigant-Engine"!?
(I need say YES)

Why?

Well, the best free Rybka 2.3.2a with 5-men EGTB's and usage=Normal, lost too!

Here is this fantastic experimental-game.

CPU=Core 2 Duo 2.66 GHz
GUI=Fritz 10
PB=ON
EGTB=ON (Rybka only)
Timecontrol=30 Minutes
Both Engines used 1 core with 256 MB Hash

[Event "Core 2 Duo 2.66 GHz, Rapid 30'"]
[Site "Privat"]
[Date "2010.07.19"]
[Round "1"]
[White "Crab 1.0 beta"]
[Black "Rybka 2.3.2a 32-bit"]
[Result "1-0"]
[Annotator "3.11;1.46"]
[SetUp "1"]
[FEN "8/p2k1p2/P3pP2/2n1P1B1/3p2PP/1p6/8/1K6 w - - 0 6"]
[PlyCount "91"]
[TimeControl "1800"]

{Intel(R) Core(TM)2 Duo CPU E6750 @ 2.66GHz 2660 MHz W=27.7 ply; 1.263kN/
s B=25.0 ply; 72kN/s; 146.847 TBAs} 6. Kb2 {Beide letzter Buchzug 3.11/27 48}
Nd3+ {1.46/25 69} 7. Kxb3 {2.94/29 60} Nxe5 {1.49/28 34} 8. h5 {3.15/31 36} Kc6
{(Sxg4) 1.63/27 73} 9. h6 {2.78/26 50} Ng6 {1.63/27 177} 10. h7 {2.90/32 23}
Kb5 {1.63/27 25} 11. Bh6 {3.03/32 40} Kxa6 {1.63/27 73} 12. Kc4 {3.11/32 31} e5
{(Kb6) 1.63/27 17} 13. Kc5 {3.21/29 97} Kb7 {1.63/27 15} 14. Bd2 {3.15/30 27}
Kc7 {(Kc8) 1.63/26 9} 15. Ba5+ {(Lb4) 3.15/28 44} Kd7 {(Kc8) 1.63/27 90} 16.
Kd5 {2.82/30 50} Kc8 {1.63/27 9} 17. Kc6 {(Le1) 2.84/31 29} Nh8 {1.63/27 66}
18. Be1 {(Kd6) 3.01/31 0} Kd8 {(Sg6) 1.63/28 39} 19. Kd5 {2.72/30 37} Ng6 {
1.63/29 34} 20. Kd6 {(Lg3) 2.88/30 0} d3 {1.63/28 127} 21. Ba5+ {
(Kd5) 3.17/33 0} Kc8 {1.63/27 35} 22. Bd2 {(Kd5) 3.35/29 2} Kd8 {1.63/26 43}
23. g5 {(Kd5) 2.72/29 10} Kc8 {1.63/26 59} 24. Kc6 {(Kd5) 2.64/32 43} Kd8 {
1.63/27 109} 25. Ba5+ {(Kd5) 2.64/31 37} Kc8 {(Ke8) 1.63/26 54} 26. Kd5 {
2.46/28 31} Kb7 {1.63/26 13} 27. Kc5 {2.48/28 17} Ka6 {(Sh8) 1.63/25 8} 28. Bc3
{(Ld2) 2.48/29 45} Kb7 {1.63/24 23} 29. Bd2 {(Lb4) 3.11/28 20} Kc7 {1.63/25 68}
30. Ba5+ {(Lb4) 3.07/29 0} Kb7 {1.63/25 127} 31. Kb5 {(Lc3) 3.25/31 0} Nh8 {
1.63/21 20} 32. Kc4 {(Lc3) 3.35/29 7} e4 {1.32/23 21} 33. Kd4 {4.48/28 14} Kc6
{(Ka6) 1.63/26 15} 34. Kxe4 {(Lc3) 4.64/29 42} Kb5 {1.22/22 20} 35. Bd2 {
(Lc3) 5.19/28 99} Kc4 {(Kc6) 1.42/25 12} 36. Kf4 {(Kf5) 5.51/24 25} Kb5 {
2.32/25 58} 37. Kg4 {5.31/25 0} a5 {2.41/24 6} 38. Kh5 {5.95/22 17} a4 {
2.88/23 7} 39. Kh6 {5.95/23 27} a3 {2.98/23 0} 40. Kg7 {6.04/21 23} a2 {
3.99/22 30} 41. Kxh8 {5.55/21 66} a1=Q {4.22/23 0} 42. Kg7 {6.28/22 29} Qd1 {
(Db2) 4.49/21 17} 43. h8=Q {6.60/19 22} Qxd2 {4.49/20 19} 44. Qe8+ {6.60/20 0}
Kb4 {(Kc5) 4.49/20 36} 45. g6 {6.76/19 19} Qf4 {(Dg5) 4.49/18 9} 46. Qe7+ {
6.76/19 27} Kc3 {(Kb3) 4.49/17 0} 47. Qc5+ {8.26/16 20} Kb2 {4.49/16 5} 48.
gxf7 {8.02/17 17} Qg4+ {6.99/15 8} 49. Kh6 {12.48/19 22} Qh4+ {6.99/14 0} 50.
Kg6 {8.26/21 22} Qe4+ {(Dg4+) 6.99/13 0} 51. Qf5 {13.05/20 28} 1-0

Position after move 41:

7K/5p1P/5P2/1k4P1/8/3p4/3B4/q7 w - - 0 1

Analysis by Crab 1.0 beta:

42.Kg7 Dd1 43.h8D Dxd2 44.De8+ Kb4 45.g6 Dg5 46.Dxf7 De5 47.De7+ Dxe7+
+- (5.87) Tiefe: 20 00:00:31 80610kN
42.Kg7 Dd1 43.h8D Dxd2 44.De8+ Kb4 45.g6 Df4 46.De7+ Kc3 47.Da3+ Kd2 48.Da5+ Kc2 49.Dc5+ Kb3 50.Kxf7 d2 51.g7 d1D 52.g8D Ddg4 53.Dxg4 Dxg4 54.Kf8 Kb2 55.f7 Kb3 56.Ke7 De4+ 57.Kd7 Dg4+
+- (5.87) Tiefe: 20 00:00:34 87896kN

(Nemeth, Privat 19.07.2010)

Rybka have here a queen more (!) but Rybka is anyway lost! WOW, fantastic game by Crab 1.0 beta! 8-)

ED.
Peace be with you!
Top
Howard E
Posts: 46
Joined: Fri Jun 11, 2010 3:57 am
Real Name: Howard Exner

Re: Testposition Endgame

Post by Howard E » Mon Jul 19, 2010 9:20 pm

5Q2/p7/P3K1k1/3pP3/4q3/8/8/8 w - - 1 19 bm Qe8;

Hi Robert this is your analysis up to move 19.
I got the same line when I posted initially that Be3 for white will pose problems for black.
But now 20. Qe8+ Kh6 21. Qc6 Qc4 22. Qxd5 Qxa6 is a 6 man tablebase mate in 31.
Of couurse there may be better moves. Houdini is a strong analyis partner and found some good moves.

8/p7/q3K2k/3QP3/8/8/8/8 w - - 0 22 bm ;
Top
User avatar
Robert Houdart
Posts: 180
Joined: Thu Jun 10, 2010 4:55 pm
Contact:
Contact Robert Houdart

Re: Testposition Endgame

Post by Robert Houdart » Mon Jul 19, 2010 10:26 pm

Interesting, ultimately this game seems to be decided by KQPKQP table base.
This kind of endgame surely encourages me to add some EGTB support.

Robert
Top
BB+
Posts: 1484
Joined: Thu Jun 10, 2010 4:26 am

Re: Testposition Endgame

Post by BB+ » Tue Jul 20, 2010 5:16 am

After 4...Bxd4 5.Bxd4 Nxd4 6.Kxd4 Ke8 we're heading towards a queen ending.
Indeed, the next 12 moves or so (until white queens) are essentially forced. I analysed the positions after that, and convinced myself that White wins. Having a QPP vs Q and QP vs QP tablebase is useful.

Code: Select all

7.h5 Kf8 8.g5 Kg8 9.g6 fxg6 10.hxg6 Kf8 11.Kd3 Kg8 12.g7 Kf7 13.Kd4 Kg8 14.Kc5 b3 15.Kd6 b2 16.Ke7 b1Q 17.f7+ Kxg7 18.f8Q+ Kg6
5Q2/p3K3/P3p1k1/3pP3/8/8/8/1q6 w - - 1 19
I don't think Kxe6 directly is best. I like 19. Qf6+ Kh5 20. Kxe6 Kg4 21. Qg7+ Kf4 22. Qf7+ Kg4 23. Qf2! and the scores are getting into the +4 and beyond (QPP vs Q usually) after Black's Qb8 or Qd3. Black can try other things, like d4 instead of Kg4, but that doesn't seem to work either. For instance 23. Qf2 Qb8 24. Qd4+ Kg5 25. Kd7 Qb5+ 26. Ke7

There may still be a draw here, but I don't see it.
Top
Post Reply

Return to “General Topics”