Surprise ending (from Winge - Toro ICCF)

[Jouni]

[D]r7/8/2K4p/8/1B1k4/P6P/8/1Rb5 b - - 0 93

-Bxa3! = forced draw

[Ajedrecista]

Hello Jouni:

[D]r7/8/2K4p/8/1B1k4/P6P/8/1Rb5 b - - 0 93

-Bxa3! = forced draw

Spark does not have problems in this position.

Analysis by Spark 1.0 w32 (using two cores) in an Intel Pentium D930 (3 GHz), with 512 MB of hash, running under Arena 2.0.1 GUI:

FEN: r7/8/2K4p/8/1B1k4/P6P/8/1Rb5 b - - 0 93

Spark-1.0-win32-mp:

1 00:00 32 0 -4,00 h6h5 Bb4c5+ Kd4e4 Rb1xc1
1 00:00 52 0 -3,97 Bc1b2 Rb1xb2
1 00:00 105 0 -2,70 Bc1d2 Bb4xd2 Ra8a6+ Kc6d7 Ra6xa3
1 00:00 117 0 -0,12 Bc1xa3

2/1 00:00 337 0 -0,14 Bc1xa3 Rb1d1+ Kd4e4 Rd1e1+ Ke4d4

3/3 00:00 883 0 -0,10 Bc1xa3 Rb1a1 Ra8c8+ Kc6d7

4/4- 00:00 1.345 0 -0,35 Bc1xa3 Rb1a1 Ra8c8+ Kc6d7
4/6- 00:00 2.759 0 -0,69 Bc1xa3 Rb1a1 Ra8c8+ Kc6d7
4/6 00:00 4.906 0 -0,57 Kd4d3 Kc6d5 Kd3c2 Rb1a1
4/6 00:00 5.831 0 -0,48 Kd4e4 Rb1a1 Bc1b2 Ra1e1+ Ke4d3

5/6- 00:00 7.420 0 -0,73 Kd4e4 Rb1a1 Bc1b2 Ra1e1+ Ke4d3
5/7 00:00 14.897 0 -0,69 Bc1f4 Bb4c5+ Kd4d3 Rb1b4 Ra8c8+ Kc6d5 Bf4e5 Rb4b3+ Kd3c2
5/8 00:00 23.821 0 -0,63 Bc1g5 Bb4c5+ Kd4d3 Bc5d6 Ra8a6+ Kc6d5 Ra6a5+ Kd5e6

6/8 00:00 30.299 0 -0,68 Bc1g5 Bb4d6 Ra8a6+ Rb1b6 Ra6xb6+ Kc6xb6 Bg5d8+ Kb6c6 Bd8h4 Bd6c5+ Kd4c4

7/12 00:00 79.054 718.672 -0,65 Bc1g5 Bb4c5+ Kd4d3 Bc5d6 Ra8a6+ Rb1b6 Ra6a7 Rb6b3+ Kd3c2 Rb3b5

8/13 00:00 166.084 818.147 -0,64 Bc1g5 Kc6b7 Ra8a4 Bb4f8 Kd4c3 Rb1b5 Bg5e3 Bf8g7+ Kc3c2

9/16 00:00 402.496 756.571 -0,67 Bc1g5 Bb4d6 Kd4d3 Rb1b3+ Kd3c2 Rb3b5 Bg5e3 Kc6d5 Ra8a4 Bd6f8

10/16 00:00 547.916 833.966 -0,68 Bc1g5 Bb4d6 Kd4d3 Rb1b3+ Kd3c2 Rb3b5 Bg5f6 Kc6d5 Ra8a4 Bd6c5 h6h5

11/17 00:01 886.225 886.225 -0,70 Bc1g5 Bb4d6 Kd4d3 Rb1b3+ Kd3c2 Rb3f3 Ra8a6+ Kc6d5 h6h5 Bd6b4 Bg5h4 Rf3c3+ Kc2b2 Rc3c4

12/18 00:01 1.410.405 893.792 -0,67 Bc1g5 Kc6b7 Ra8a4 Rb1d1+ Kd4e4 Kb7c6 Ra4a8 Rd1e1+ Ke4d3 Kc6d5 Bg5h4 Re1d1+ Kd3c2 Rd1d4 Bh4f6 Rd4d2+ Kc2b3

13/20 00:02 2.634.040 926.174 -0,67 Bc1g5 Kc6b7 Ra8a4 Rb1d1+ Kd4c4 Rd1d6 Bg5f4 Rd6f6 Bf4g5 Rf6e6 Kc4d5 Re6d6+ Kd5c4 Rd6c6+ Kc4d4 Kb7b6
13/22+ 00:03 3.243.885 947.950 -0,42 Bc1xa3 Rb1a1
13/22 00:03 3.327.863 950.818 -0,40 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 Rh8h7+ Kd7e6 Rh7c7

14/22 00:03 3.567.182 975.439 -0,37 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 Rh8h7+ Kd7e6 Kb2c3 Ke6d5 Rh7d7+ Kd5e5

15/23 00:03 4.049.952 1.024.526 -0,35 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 Rh8h7+ Kd7e6 Kb2c3 Ra6a3+ Kc3d4 Ra3a4+ Kd4e3

16/25 00:04 4.882.182 1.034.579 -0,28 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 h6h5 Ra6a5 Kb2c3 Kd7e7 Kc3d4

17/26 00:05 5.971.510 1.070.546 -0,29 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 Ra6c6 Rh8h7+ Kd7e8 h6h5 h3h4 Rh7a7 Rc6c5 Ra7a8+ Ke8f7

18/26 00:06 7.263.590 1.114.731 -0,35 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 Ra6c6 Rh8h7+ Kd7e8 h6h5 h3h4 Rh7a7 Rc6c5 Ra7a8+ Ke8f7 Ra8a7+ Kf7g6

19/29 00:09 10.988.648 1.160.486 -0,29 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 Ra6c6 Rh8h7+ Kd7e8 h6h5 h3h4 Kb2b3 Rc6c5 Kb3b4 Rc5d5 Kb4c4

20/29 00:17 20.122.076 1.129.628 -0,22 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 h6h5 Ra6a4 Kb2b3 Ra4a5 Kb3c4 Kd7e7 Kc4d4 Ke7f7 Rh8h7+ Kf7f6 Rh7h6+ Kf6g5

21/30 00:27 30.333.071 1.099.901 -0,22 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 h6h5 Ra6a4 Kb2b3 Ra4a5 Kb3c3 Kd7d6 Rh8h6+ Kd6e7 Kc3d4 Ke7f7 Rh6h7+ Kf7g6

22/34 00:47 50.933.745 1.070.171 -0,22 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 h6h5 Ra6a4 Kb2b3 Ra4a5 Kb3c3 Kd7d6 Rh8h6+ Kd6e7 Kc3d4 Ra5b5 Rh6h7+ Ke7f6 Rh7h6+ Kf6g5
22/35 01:00 67.395.316 1.120.044 -0,22 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 h6h5 Ra6a4 Kb2b3 Ra4a5 Kb3c3 Kd7d6 Rh8h6+ Kd6e7 Kc3d4 Ra5b5 Rh6h7+ Ke7f6 Rh7h6+ Kf6g5

I stopped the analysis after a minute. Spark found ...,Bxa3 with less than 3.3e+6 searched nodes!

Regards from Spain.

Ajedrecista.

[syzygy]

I stopped the analysis after a minute. Spark found ...,Bxa3 with less than 3.3e+6 searched nodes!

Using tablebases?

[Ajedrecista]

Hello Ronald:

I stopped the analysis after a minute. Spark found ...,Bxa3 with less than 3.3e+6 searched nodes!

Using tablebases?

AFAIK, Spark can not use EGTB. I just tried a K + R vs. K position with Spark and it runs into depth instead giving the exact DTM value; furthermore, Arena showed '0 TBHits'.

I have all 3, 4 and 5-man EGTB of both Nalimov and Gaviota formats but Spark can not use them. I also looked into the configuration window of Spark and I did not find a place to type the EGTB path.

Regards from Spain.

Ajedrecista.

[syzygy]

AFAIK, Spark can not use EGTB. I just tried a K + R vs. K position with Spark and it runs into depth instead giving the exact DTM value; furthermore, Arena showed '0 TBHits'.

I have all 3, 4 and 5-man EGTB of both Nalimov and Gaviota formats but Spark can not use them. I also looked into the configuration window of Spark and I did not find a place to type the EGTB path.

Interesting. Using 6-piece tablebases my engine finds Bxa3 immediately and gives it a draw value after 147839 nodes, but without tablebases it prefers other moves. After Bxa3 Ra1 Bxb4 Rxa8 the position is a tablebase draw, but without tablebases most engines will probably not understand this.

Now that I have a better look at Spark's solution: it does not work . After Bxa3 Ra1 Kc4 the position is won for white:
[D]r7/8/2K4p/8/1Bk5/b6P/8/R7 w - - 2 3
My engine gives Bxa3 Ra6+ Kc7 Ra5 Bb2 Rh5 Ra3 Rd5 Bc1 h5 Ra6 Rf5 h4 Rf7+ Kd8 Rf8+ Ke7 Rc8 Ra5 Rc7+ Kf6 Rc6+ Kg5 Kb4 Bd2+ Rc3 Bxc3+ after which it is a tablebase win.

[michiguel]

Hello Jouni:

[D]r7/8/2K4p/8/1B1k4/P6P/8/1Rb5 b - - 0 93

-Bxa3! = forced draw

Spark does not have problems in this position.

Analysis by Spark 1.0 w32 (using two cores) in an Intel Pentium D930 (3 GHz), with 512 MB of hash, running under Arena 2.0.1 GUI:

FEN: r7/8/2K4p/8/1B1k4/P6P/8/1Rb5 b - - 0 93

Spark-1.0-win32-mp:

1 00:00 32 0 -4,00 h6h5 Bb4c5+ Kd4e4 Rb1xc1
1 00:00 52 0 -3,97 Bc1b2 Rb1xb2
1 00:00 105 0 -2,70 Bc1d2 Bb4xd2 Ra8a6+ Kc6d7 Ra6xa3
1 00:00 117 0 -0,12 Bc1xa3

2/1 00:00 337 0 -0,14 Bc1xa3 Rb1d1+ Kd4e4 Rd1e1+ Ke4d4

3/3 00:00 883 0 -0,10 Bc1xa3 Rb1a1 Ra8c8+ Kc6d7

4/4- 00:00 1.345 0 -0,35 Bc1xa3 Rb1a1 Ra8c8+ Kc6d7
4/6- 00:00 2.759 0 -0,69 Bc1xa3 Rb1a1 Ra8c8+ Kc6d7
4/6 00:00 4.906 0 -0,57 Kd4d3 Kc6d5 Kd3c2 Rb1a1
4/6 00:00 5.831 0 -0,48 Kd4e4 Rb1a1 Bc1b2 Ra1e1+ Ke4d3

5/6- 00:00 7.420 0 -0,73 Kd4e4 Rb1a1 Bc1b2 Ra1e1+ Ke4d3
5/7 00:00 14.897 0 -0,69 Bc1f4 Bb4c5+ Kd4d3 Rb1b4 Ra8c8+ Kc6d5 Bf4e5 Rb4b3+ Kd3c2
5/8 00:00 23.821 0 -0,63 Bc1g5 Bb4c5+ Kd4d3 Bc5d6 Ra8a6+ Kc6d5 Ra6a5+ Kd5e6

6/8 00:00 30.299 0 -0,68 Bc1g5 Bb4d6 Ra8a6+ Rb1b6 Ra6xb6+ Kc6xb6 Bg5d8+ Kb6c6 Bd8h4 Bd6c5+ Kd4c4

7/12 00:00 79.054 718.672 -0,65 Bc1g5 Bb4c5+ Kd4d3 Bc5d6 Ra8a6+ Rb1b6 Ra6a7 Rb6b3+ Kd3c2 Rb3b5

8/13 00:00 166.084 818.147 -0,64 Bc1g5 Kc6b7 Ra8a4 Bb4f8 Kd4c3 Rb1b5 Bg5e3 Bf8g7+ Kc3c2

9/16 00:00 402.496 756.571 -0,67 Bc1g5 Bb4d6 Kd4d3 Rb1b3+ Kd3c2 Rb3b5 Bg5e3 Kc6d5 Ra8a4 Bd6f8

10/16 00:00 547.916 833.966 -0,68 Bc1g5 Bb4d6 Kd4d3 Rb1b3+ Kd3c2 Rb3b5 Bg5f6 Kc6d5 Ra8a4 Bd6c5 h6h5

11/17 00:01 886.225 886.225 -0,70 Bc1g5 Bb4d6 Kd4d3 Rb1b3+ Kd3c2 Rb3f3 Ra8a6+ Kc6d5 h6h5 Bd6b4 Bg5h4 Rf3c3+ Kc2b2 Rc3c4

12/18 00:01 1.410.405 893.792 -0,67 Bc1g5 Kc6b7 Ra8a4 Rb1d1+ Kd4e4 Kb7c6 Ra4a8 Rd1e1+ Ke4d3 Kc6d5 Bg5h4 Re1d1+ Kd3c2 Rd1d4 Bh4f6 Rd4d2+ Kc2b3

13/20 00:02 2.634.040 926.174 -0,67 Bc1g5 Kc6b7 Ra8a4 Rb1d1+ Kd4c4 Rd1d6 Bg5f4 Rd6f6 Bf4g5 Rf6e6 Kc4d5 Re6d6+ Kd5c4 Rd6c6+ Kc4d4 Kb7b6
13/22+ 00:03 3.243.885 947.950 -0,42 Bc1xa3 Rb1a1
13/22 00:03 3.327.863 950.818 -0,40 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 Rh8h7+ Kd7e6 Rh7c7

14/22 00:03 3.567.182 975.439 -0,37 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 Rh8h7+ Kd7e6 Kb2c3 Ke6d5 Rh7d7+ Kd5e5

15/23 00:03 4.049.952 1.024.526 -0,35 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 Rh8h7+ Kd7e6 Kb2c3 Ra6a3+ Kc3d4 Ra3a4+ Kd4e3

16/25 00:04 4.882.182 1.034.579 -0,28 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 h6h5 Ra6a5 Kb2c3 Kd7e7 Kc3d4

17/26 00:05 5.971.510 1.070.546 -0,29 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 Ra6c6 Rh8h7+ Kd7e8 h6h5 h3h4 Rh7a7 Rc6c5 Ra7a8+ Ke8f7

18/26 00:06 7.263.590 1.114.731 -0,35 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 Ra6c6 Rh8h7+ Kd7e8 h6h5 h3h4 Rh7a7 Rc6c5 Ra7a8+ Ke8f7 Ra8a7+ Kf7g6

19/29 00:09 10.988.648 1.160.486 -0,29 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 Ra6c6 Rh8h7+ Kd7e8 h6h5 h3h4 Kb2b3 Rc6c5 Kb3b4 Rc5d5 Kb4c4

20/29 00:17 20.122.076 1.129.628 -0,22 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 h6h5 Ra6a4 Kb2b3 Ra4a5 Kb3c4 Kd7e7 Kc4d4 Ke7f7 Rh8h7+ Kf7f6 Rh7h6+ Kf6g5

21/30 00:27 30.333.071 1.099.901 -0,22 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 h6h5 Ra6a4 Kb2b3 Ra4a5 Kb3c3 Kd7d6 Rh8h6+ Kd6e7 Kc3d4 Ke7f7 Rh6h7+ Kf7g6

22/34 00:47 50.933.745 1.070.171 -0,22 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 h6h5 Ra6a4 Kb2b3 Ra4a5 Kb3c3 Kd7d6 Rh8h6+ Kd6e7 Kc3d4 Ra5b5 Rh6h7+ Ke7f6 Rh7h6+ Kf6g5
22/35 01:00 67.395.316 1.120.044 -0,22 Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2 Ra8c8+ Kc6d7 Kb3xb2 Ra1a6 Rc8h8 h3h4 h6h5 Ra6a4 Kb2b3 Ra4a5 Kb3c3 Kd7d6 Rh8h6+ Kd6e7 Kc3d4 Ra5b5 Rh6h7+ Ke7f6 Rh7h6+ Kf6g5

I stopped the analysis after a minute. Spark found ...,Bxa3 with less than 3.3e+6 searched nodes!

Regards from Spain.

Ajedrecista.

Spark might be finding it for the wrong reasons. The PV looks weird, for instance:

Bc1xa3 Rb1a1 Kd4c4 Bb4xa3 Kc4b3 Ba3b2

Rather than Ba3b2, Kd6 wins.

Miguel