Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Sune Endgame test suite (analysis summary)

Author: Sune Larsson

Date: 02:30:12 12/28/05

Go up one level in this thread


On December 28, 2005 at 03:37:40, Dann Corbit wrote:

>On December 28, 2005 at 03:27:37, Sune Larsson wrote:
>
>>On December 27, 2005 at 23:02:42, Dann Corbit wrote:
>>
>>>sune-eg.epd:
>>>8/1ppr1kp1/p1p4p/8/8/5P2/PPP1RKPP/8 w - - 0 1
>>>8/1pk3pp/p7/3p1p2/3K4/6P1/PP2PP1P/8 b - - 0 1
>>>8/4k3/r4pp1/pR6/P4KP1/5P2/8/8 w - - 0 1
>>>8/p1p1k3/1p1p2p1/3Pq2p/2P1P3/1P3QPK/P7/8 w - - 0 1
>>>8/p3ppk1/1p4pp/3nN3/6PP/1P3P2/P3PK2/8 w - - 0 1
>>>7k/pp4pp/2n5/8/8/P7/1P4PP/2K1B3 b - - 0 1
>>>8/8/p4Bp1/1pPb2P1/1P2kp2/P7/5K2/8 w - - 0 1
>>>8/4k3/p1B4p/2K5/1P4bP/8/8/8 b - - 0 1
>>>3R4/1p6/2b5/2P1k2p/p3p2P/P6r/1P2KB2/8 b - - 0 1
>>>8/5b2/1k2p1p1/2NpP1P1/1K1P4/8/8/8 w - - 0 1
>>>8/5p1k/r5pp/P7/3R3P/6P1/5PK1/8 w - - 0 1
>>>4b3/5k2/4p1p1/3pP2p/2pP1P1P/2P5/6N1/2K5 w - - 0 1
>>>8/1p1b2k1/p1p1p1p1/2P1P2p/1P3P1P/P2B4/5K2/8 w - - 0 1
>>>6k1/2p3np/1p1p2p1/3P4/1PPK1R2/6PB/7P/4r3 w - - 0 1
>>>1n6/4k2p/p3ppp1/1pPp4/3P1PP1/3NP3/P3K2P/8 w - - 0 1
>>>3r4/7p/Rp4k1/5p2/4p3/2P5/PP3P1P/5K2 b - - 0 1
>>>8/4k3/1p2b1pp/p1p1p3/P1P1P3/1P3P1P/5KB1/8 b - - 0 37
>>>2r5/1p1k1p2/p2p2p1/P1PPnp1p/1P1K1N1P/5P2/4R1P1/8 w - - 0 1
>>>8/6b1/4k2p/p1p1p1p1/2P1K1P1/PP3N1P/8/8 w - - 0 1
>>>2r3k1/4pp1p/3p2p1/p2R4/1r6/1P4P1/P3PP1P/R5K1 b - - 0 1
>>>
>>>On 2.2 GHz AMD Athlon, at 3 minutes per position, here are the results:
>>>
>>> Fruit2.2.1   Shredder9UCI Rybka-v1.B32 Ktulu_75    YacePadrbrn  Bestia-090
>>> 1)   Kf2-e3  1)   f3-f4   1)   f3-f4   1)   f3-f4  1)    f3-f4  1)    f3-f4
>>> 2).. Kc7-d6  2).. Kc7-d6  2).. Kc7-d6  2).. Kc7-d6 2) .. Kc7-d6 2) .. Kc7-d6
>>> 3)   g4-g5   3)   Rb5-d5  3)   g4-g5   3)   Rb5-d5 3)    g4-g5  3)    Rb5-d5
>>> 4)   Qf3-f4  4)   Qf3-e2  4)   g3-g4   4)   Qf3-f4 4)    Qf3-e2 4)    Qf3-e3
>>> 5)   Ne5-c6  5)   e2-e3   5)   Ne5-c6  5)   Ne5-c6 5)    Ne5-c6 5)    Ne5-c6
>>> 6).. Kh8-g8  6).. Kh8-g8  6).. Kh8-g8  6).. Kh8-g8 6) .. g7-g6  6) .. Kh8-g8
>>> 7)   Bf6-c3  7)   Bf6-e7  7)   Bf6-c3  7)   Kf2-e2 7)    Bf6-c3 7)    Bf6-c3
>>> 8).. Bg4-c8  8).. Bg4-e2  8).. Ke7-d8  8).. Bg4-c8 8) .. Ke7-f6 8) .. Ke7-f6
>>> 9).. Ke5-f5  9).. Bc6-d5  9).. Bc6-d5  9).. Bc6-d5 9) .. Bc6-d5 9) .. Bc6-b5
>>>10)   Nc5-d3 10)   Nc5-a4 10)   Kb4-b3 10)   Kb4-c3 10)   Nc5-a4 10)   Nc5-a4
>>>11)   Rd4-d5 11)   Rd4-d5 11)   Rd4-d5 11)   Rd4-d5 11)   Rd4-d5 11)   Rd4-d5
>>>12)   Ng2-e1 12)   Ng2-e1 12)   Ng2-e1 12)   Kc1-d2 12)   Kc1-d2 12)   Kc1-d2
>>>13)   Kf2-e2 13)   Kf2-e3 13)   Kf2-e3 13)   Kf2-e3 13)   Kf2-e3 13)   Kf2-e3
>>>14)   Rf4-f2 14)   Kd4-d3 14)   Rf4-f6 14)   Rf4-f2 14)   Rf4-f6 14)   Bh3-d7
>>>15)   f4-f5  15)   f4-f5  15)   f4-f5  15)   f4-f5  15)   f4-f5  15)   f4-f5
>>>16).. Rd8-b8 16).. Rd8-b8 16).. Rd8-b8 16).. Rd8-b8 16).. Rd8-b8 16).. Rd8-b8
>>>17).. b6-b5  17).. b6-b5  17).. b6-b5  17).. b6-b5  17).. b6-b5  17).. Ke7-f6
>>>18)   Re2xe5 18)   Nf4-h3 18)   Re2-e3 18)   Re2-e1 18)   Nf4-h3 18)   Re2-e3
>>>19)   Nf3-e1 19)   Nf3-h2 19)   Nf3-e1 19)   Nf3-e1 19)   Nf3-e1 19)   Nf3-e1
>>>20).. a5-a4  20).. a5-a4  20).. a5-a4  20).. a5-a4  20).. a5-a4  20).. a5-a4
>>>12/27/2005 2:37:12 PM, Total time: 1:02:06 AM Rated time: 1:00:00 = 3600 Seconds
>>>--------------------------------------------------------------------------------
>>
>> Nice! Hopefully there will pop up some testers that will find games in 40/40 -
>> from these positions - interesting. I'm at work right now - but from memory
>>
>> 11)  Rd4-d5? is bad - the rook belongs behind the pawn.
>> 16)..Rd8-b8? is too passive. This is about activity in rook ending.
>> 18)..Re2xe5! (Fruit!) is an excellent move.
>>
>> Furthermore - if 15)  f4-f5 is the old game with Kn+pawns vs Kn+pawns - then
>> this position just might be too easy. I will have a look when I have a chess
>> board in front of me...
>>
>> /S
>
>There looks to be a couple of "solution" moves in the set, so some of the
>positions may not be equal.  Now, I recognize those draws where the engine says
>+177, +177, +177 ply after ply.  But where there is a sudden change for the
>better and it keeps improving, makes it seem that the position is not balanced.
>
>For instance, these two:
>15) 0 1
>    Avoid move:
>    Best move (Rybka v1.0 Beta.w32): f4-f5
>    Not found in: 30:00
>      3	00:01	         242	6.044	+0.30	g4g5
>      4	00:01	         535	10.741	+0.24	g4g5
>      4	00:01	       1.608	23.191	+0.27	f4f5
>      5	00:01	       2.148	21.777	+0.21	f4f5 g6g5
>      6	00:01	       3.365	31.042	+0.18	f4f5 g6g5 f5xe6
>      7	00:01	       6.635	48.186	+0.12	f4f5 g6g5 f5xe6 Nb8c6
>      7	00:01	       8.690	52.038	+0.14	Nd3b4 a6a5 Nb4d3 Nb8c6
>      8	00:01	      13.100	60.698	+0.11	Nd3b4 a6a5 Nb4d3 Nb8c6 Ke2f3
>      9	00:01	      26.426	79.355	+0.06	Nd3b4 a6a5 Nb4d3 Nb8c6 g4g5 b5b4
>      9	00:01	      36.119	89.989	+0.18	g4g5 Nb8c6 a2a3 f6f5 Ke2d2 Ke7d7
>     10	00:01	      48.868	99.881	+0.15	g4g5 Nb8c6 a2a3 f6f5 Nd3b4 Nc6xb4 a3xb4
>Ke7d7
>     11	00:02	      93.817	115.467	+0.10	g4g5 Nb8c6 a2a3 f6f5 Nd3b4 Nc6xb4 a3xb4
>Ke7d7 h2h3
>     12	00:02	     173.470	123.959	+0.13	g4g5 Nb8c6 a2a3 f6f5 Nd3b4 Nc6xb4 a3xb4
>Ke7d7 Ke2d3 Kd7c7
>     13	00:03	     286.259	137.361	+0.12	g4g5 Nb8c6 a2a3 f6f5 Nd3b4 Nc6xb4 a3xb4
>Ke7d7 Ke2d3 Kd7c7
>     14	00:04	     519.703	148.818	+0.12	g4g5 Nb8c6 a2a3 f6f5 Nd3b4 Nc6xb4 a3xb4
>Ke7d7 Ke2d3 Kd7c7
>     15	00:08	   1.089.764	162.907	+0.10	g4g5 Nb8c6 a2a3 f6f5 Nd3b4 Nc6xb4 a3xb4
>Ke7d7 Ke2d3 Kd7c7
>     15	00:14	   2.034.141	169.511	+0.11	a2a3 a6a5 g4g5 Nb8c6 Ke2d2 f6f5 Kd2c3
>Ke7d8 Nd3e1 Kd8d7
>     16	00:21	   3.308.253	174.118	+0.14	a2a3 f6f5 g4xf5 e6xf5 Nd3e5 Ke7d8 h2h4
>Kd8c7 Ne5f3 Nb8c6
>     17	00:30	   4.967.461	177.409	+0.10	a2a3 a6a5 g4g5 Nb8c6 Ke2d2 f6f5 Kd2c3
>Ke7d8 Nd3e1 Kd8d7
>     17	00:40	   6.507.436	171.248	+0.16	f4f5 g6g5 Nd3b4 a6a5 Nb4c2 e6xf5 g4xf5
>Ke7d7 e3e4 d5xe4
>     18	00:51	   8.160.114	166.532	+0.20	f4f5 g6g5 Nd3b4 a6a5 c5c6 Ke7d6 f5xe6
>Nb8xc6 Nb4xc6 Kd6xc6
>     19	01:14	  11.616.726	161.343	+0.60	f4f5 g6g5 Nd3b4 a6a5 c5c6 Ke7d6 f5xe6
>Nb8xc6 Nb4xc6 Kd6xc6
>     20	01:40	  15.427.791	160.706	+0.62	f4f5 g6g5 Nd3b4 a6a5 c5c6 Ke7d6 f5xe6
>Nb8xc6 Nb4xc6 Kd6xc6
>     21	02:11	  20.251.876	160.729	+0.60	f4f5 g6g5 Nd3b4 a6a5 c5c6 Ke7d6 f5xe6
>Nb8xc6 Nb4xc6 Kd6xc6
>     22	03:21	  30.421.534	156.007	+0.66	f4f5 g6g5 Nd3b4 a6a5 c5c6 Ke7d6 f5xe6
>Nb8xc6 Nb4xc6 Kd6xc6
>     23	04:47	  43.473.317	155.818	+0.64	f4f5 g6g5 Nd3b4 a6a5 c5c6 Ke7d6 f5xe6
>Nb8xc6 Nb4xc6 Kd6xc6
>     24	09:00	  77.976.755	148.244	+0.85	f4f5 g6g5 Nd3b4 a6a5 c5c6 Ke7d6 f5xe6
>Nb8xc6 Nb4xc6 Kd6xc6
>     25	27:21	 229.653.704	143.443	+1.36	f4f5 g6g5 Nd3b4 a6a5 c5c6 Ke7d6 f5xe6
>Nb8xc6 Nb4xc6 Kd6xc6
>   12/27/2005 8:25:08 PM, Time for this analysis: 00:30:01, Rated time: 7:30:00
>
>17) 0 37
>    Avoid move:
>    Best move (Rybka v1.0 Beta.w32): b6-b5
>    Not found in: 30:00
>      3	00:01	         128	131.072	+0.50	Ke7d6
>      4	00:01	         235	240.640	+0.56	Ke7d6
>      5	00:01	         684	700.416	+0.48	Ke7d6 Kf2e3
>      6	00:01	       2.048	99.864	+0.32	Ke7d6 h3h4 Be6d7
>      6	00:01	       3.002	146.383	+0.47	Ke7f6 Kf2e3 h6h5
>      6	00:01	       3.735	123.375	+0.54	h6h5 Kf2e3 g6g5
>      7	00:01	       4.766	119.033	+0.45	h6h5 Kf2e3 g6g5 Bg2f1
>      7	00:01	       6.451	129.525	+0.50	g6g5 Kf2e3 Be6d7 Bg2f1
>      8	00:01	      11.557	130.048	+0.51	g6g5 Kf2e2 Ke7d6 Ke2e3 h6h5
>      9	00:01	      14.678	135.407	+0.50	g6g5 Kf2e2 Ke7d6 Ke2e3 Be6c8 Bg2f1
>     10	00:01	      21.625	146.649	+0.51	g6g5 Kf2e2 Ke7d6 Ke2e3 Be6c8 Ke3d2
>Bc8d7 Kd2d3
>     11	00:01	      62.308	154.862	+0.32	g6g5 Kf2g3 Be6d7 h3h4 Ke7f6 Bg2h3
>g5xh4+ Kg3xh4 Bd7xh3 Kh4xh3
>     12	00:02	     152.662	160.829	+0.34	g6g5 Kf2g3 Be6d7 h3h4 Ke7f6 Bg2f1
>g5xh4+ Kg3xh4 Kf6g6 Bf1h3
>     13	00:02	     215.901	164.618	+0.46	g6g5 Kf2g3 Ke7f7 h3h4 Kf7f6 Bg2f1 Kf6g6
>Bf1g2 Kg6h5 h4xg5
>     14	00:03	     274.206	167.734	+0.44	g6g5 Kf2g3 Ke7f7 h3h4 Kf7f6 Bg2f1 Be6d7
>Bf1g2 Bd7c8 Bg2f1
>     15	00:04	     479.190	180.069	+0.48	g6g5 Kf2g3 Ke7f7 Bg2f1 Kf7g6 Bf1g2
>Be6d7 Bg2f1 Bd7c6 h3h4
>     16	00:04	     596.673	183.702	+0.47	g6g5 Kf2g3 Ke7f7 Bg2f1 Kf7g6 Bf1g2
>Be6d7 Bg2f1 Bd7c6 Bf1g2
>     17	00:06	     888.364	182.009	+0.46	g6g5 Kf2g3 Ke7f7 Bg2f1 Kf7g6 Bf1g2
>Be6d7 Bg2f1 Bd7e8 Bf1g2
>     18	00:09	   1.283.614	173.819	+0.46	g6g5 Kf2g3 Ke7f7 Bg2f1 Kf7g6 Bf1g2
>Be6d7 Bg2f1 Bd7e8 Bf1g2
>     19	00:12	   1.993.658	199.365	+0.46	g6g5 Kf2g3 Ke7f7 Bg2f1 Kf7g6 Bf1g2
>Be6d7 Bg2f1 Bd7e8 Bf1g2
>     20	00:17	   2.841.630	189.442	+0.46	g6g5 Kf2g3 Ke7f7 Bg2f1 Kf7g6 Bf1g2
>Be6d7 Bg2f1 Bd7e8 Bf1g2
>     21	00:30	   5.260.979	187.892	+0.46	g6g5 Kf2g3 Ke7f7 Bg2f1 Kf7g6 Bf1g2
>Be6d7 Bg2f1 Bd7e8 Bf1g2
>     21	05:09	  41.213.342	137.377	+2.28	b6b5 c4xb5 Ke7d6 Kf2e3 Be6xb3 f3f4
>e5xf4+ Ke3xf4 Bb3xa4 b5b6
>     22	06:01	  49.114.240	139.926	+2.41	b6b5 c4xb5 Ke7d6 Kf2e3 Be6xb3 f3f4
>e5xf4+ Ke3xf4 Bb3xa4 b5b6
>     23	07:14	  60.607.006	143.278	+2.42	b6b5 c4xb5 Ke7d6 b5b6 Kd6c6 Bg2f1
>Be6xb3 Bf1b5+ Kc6xb6 h3h4
>     24	10:46	  86.054.351	136.811	+2.57	b6b5 c4xb5 Ke7d6 b5b6 Kd6c6 Bg2f1
>Be6xb3 Bf1b5+ Kc6xb6 h3h4
>     25	21:30	 153.031.859	121.646	+2.84	b6b5 c4xb5 Ke7d6 b5b6 Kd6c6 Bg2f1
>Be6xb3 Bf1b5+ Kc6xb6 Kf2e3
>   12/27/2005 9:25:12 PM, Time for this analysis: 00:30:01, Rated time: 8:30:00
>
>
>It would seem odd to have a tournament based on positions that are won for one
>side or the other.
>
>Some of the others are apparent draws that give engines a wood-counting
>headache.


 The first 9 are rather balanced positions (Nunn) - the added 11 are either
 a test in bringing home the full point (like the 3 positions with "good knight
 vs "bad bishop", won rook ending, won bishop ending)or saving half a point in
 an otherwise lost ending - like the RookE where several engines chose Rb8? -
 that position is about the need for activity in rook ending. The point is
 not finding a solution move - but showing in games that the points can be
 brought home. From memory I don't recall more than one position that is a
 draw with best play - out of the last 11 that is.

 /S



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.