SYZYGY Base question

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

IWB
Posts: 1539
Joined: Thu Mar 09, 2006 2:02 pm

SYZYGY Base question

Post by IWB »

Hi all

Maybe this is discussed already but I am to lazy to read all posting about these bases.

On the CSS Forum we stumbled upon this position:

[d]

Nothing special. #51 with a capture in move 35. This is what SYZ-Fish is doing with it against an optimal Shredder with Nalimovs:

[Event "1 Minutes/Game + 1 Seconds/Move"]
[Site "Engine Match"]
[Date "2014.01.19"]
[Round "1"]
[White "Stockfish 180114 SYZ"]
[Black "Deep Shredder 12"]
[Result "1-0"]
[SetUp "1"]
[FEN "8/6n1/6k1/8/3B1K2/1B6/8/8 w - -"]

1. Ke5 {+151.00/34 4s} Nh5 {+M53/1 8s} 2. Bc2+ {+151.00/44
0s} Kf7 {+M52/1 0s} 3. Bd1 {+151.00/43 1s} Ng7 {+M54/1 0s
(Kg6)} 4. Bb3+ {+151.00/44 1s} Kg6 {+M55/1 0s} 5. Bc2+
{+151.00/45 1s} Kf7 {+M55/1 0s} 6. Bd1 {+151.00/45 1s} Kg6
{+M54/1 0s} 7. Be3 {+151.00/46 1s} Kf7 {+M55/1 0s (Ne8)}
8. Bb3+ {+151.00/44 1s} Kg6 {+M54/1 0s} 9. Bd4 {+151.00/100
0s} Nh5 {+M53/1 0s} 10. Kd6 {+151.00/38 1s} Ng7 {+M52/1 0s}
11. Bc2+ {+151.00/31 0s} Kf7 {+M51/1 0s} 12. Kd7
{+151.00/29 0s} Nh5 {+M50/1 0s} 13. Bb3+ {+151.00/29 0s}
Kg6 {+M49/1 0s} 14. Ke7 {+151.00/20 0s} Kf5 {+M48/1 0s}
15. Bc2+ {+151.00/34 2s} Kf4 {+M47/1 0s} 16. Ke6
{+151.00/38 1s} Ng3 {+M47/1 0s} 17. Be5+ {+151.00/35 1s}
Kf3 {+M46/1 0s} 18. Bb8 {+151.00/19 0s} Ne2 {+M45/1 0s}
19. Ke5 {+151.00/21 0s} Ke3 {+M44/1 0s} 20. Be4 {+151.00/21
0s} Nc3 {+M44/1 0s} 21. Bh7 {+151.00/36 1s} Nb5 {+M43/1 0s}
22. Bg6 {+151.00/28 0s} Kd2 {+M42/1 0s} 23. Kd5 {+151.00/30
0s} Nc3+ {+M41/1 0s} 24. Kd4 {+151.00/32 0s} Ne2+ {+M40/1
0s} 25. Ke5 {+151.00/32 0s} Ke3 {+M39/1 0s} 26. Ba7+
{+151.00/24 0s} Kd2 {+M38/1 0s} 27. Bb6 {+151.00/18 0s} Kc3
{+M37/1 0s} 28. Ke4 {+151.00/19 0s} Kb4 {+M36/1 0s} 29. Bd8
{+151.00/16 0s} Kc5 {+M35/1 0s (Nc3+)} 30. Be7+ {+151.00/32
2s} Kc4 {+M34/1 0s} 31. Bf7+ {+151.00/35 2s} Kc3 {+M33/1
0s} 32. Ke3 {+151.00/36 0s} Nc1 {+M32/1 0s} 33. Bh4
{+151.00/27 0s} Nb3 {+M31/1 0s} 34. Be1+ {+151.00/28 0s}
Kb2 {+M30/1 0s} 35. Kd3 {+151.00/36 3s} Nc1+ {+M29/1 0s}
36. Kc4 {+151.00/36 0s} Kc2 {+M28/1 0s} 37. Ba5 {+151.00/25
0s} Kd1 {+M27/1 0s (Ne2)} 38. Be6 {+151.00/1 0s} Kc2
{+M26/1 0s (Ne2)} 39. Bf5+ {+151.00/1 0s} Kb2 {+M25/1 0s}
40. Bd2 {+151.00/33 3s} Na2 {+M24/1 0s (Ne2)} 41. Bh6
{+151.00/37 19s} Ka3 {+M23/1 0s} 42. Bg7 {+151.00/17 0s}
Nc1 {+M22/1 0s} 43. Bg4 {+151.00/15 1s} Na2 {+M21/1 0s}
44. Bd1 {+151.00/24 0s} Nc1 {+M20/1 0s} 45. Kc3 {+151.00/24
0s} Na2+ {+M19/1 0s} 46. Kc2 {+151.00/25 0s} Kb4 {+M18/1
0s} 47. Kb2 {+151.00/23 0s} Kc4 {+M17/1 0s (Kb5)} 48. Kxa2
{+151.00/1 0s} Kd3 {+M16/1 0s (Kd5)} 49. Kb3 {+151.00/26
2s} Kd2 {+M15/1 0s (Ke4)} 50. Bc2 {+M16/33 2s} Ke1 {+M14/1
0s (Ke2)} 51. Kc3 {+M14/36 3s} Ke2 {+M13/1 0s} 52. Bh6
{+M13/36 2s} Ke1 {+M12/1 0s} 53. Kd3 {+M12/37 2s} Kf2
{+M11/1 0s} 54. Ke4 {+M11/37 2s} Kg2 {+M10/1 0s} 55. Ke3
{+M10/38 3s} Kg3 {+M9/1 0s} 56. Bf4+ {+M9/39 2s} Kg2 {+M8/1
0s (Kg4)} 57. Bg6 {+M8/41 2s} Kf1 {+M7/1 0s (Kh3)} 58. Kf3
{+M7/45 1s} Ke1 {+M6/1 0s} 59. Bc2 {+M6/64 1s} Kf1 {+M5/1
0s} 60. Bd2 {+M5/100 1s} Kg1 {+M4/1 0s} 61. Kg3 {+M4/100
0s} Kf1 {+M3/1 0s} 62. Bd3+ {+M3/100 0s} Kg1 {+M2/1 0s}
63. Be3+ {+M2/100 0s} Kh1 {+M1/1 0s} 64. Be4# {+M1/100 0s}
1-0

SYZ-Fish needs 14 moves more than nessesary but this is just to prove that my SYS-Fish is using the SYZYGY Bases.

Now this, more complex:

[d]

Mate in 65 and capturing a piece in move 49. Here is a perfect Nalimove game out of this:

[Event "TEST"]
[Site "HOME"]
[Date "2014.01.19"]
[Round "1"]
[White "NalimovA"]
[Black "NalimovB"]
[Result "1-0"]
[SetUp "1"]
[FEN "B7/8/8/8/5nk1/8/8/4K1B1 w - -"]

1. Bh2 {0s} Kg5 {+M64/1 0s} 2. Kd2 {0s} Kf5 {+M63/1 0s}
3. Ke3 {0s} Nh5 {+M62/1 0s} 4. Bg2 {0s} Kg5 {+M61/1 0s}
5. Kd4 {1s} Kf5 {+M60/1 0s} 6. Bh3+ {0s} Kg5 {+M59/1 0s}
7. Ke5 {0s} Ng7 {+M58/1 0s} 8. Bf1 {0s} Kg6 {+M57/1 0s}
9. Bg1 {0s} Kf7 {+M56/1 0s} 10. Bd4 {0s} Kg6 {+M55/1 0s}
11. Bd3+ {1s} Kf7 {+M54/1 0s} 12. Bc2 {0s} Ne8 {+M53/1 0s}
13. Bb3+ {0s} Kg6 {+M52/1 0s} 14. Kf4 {0s} Ng7 {+M51/1 0s}
15. Kg4 {0s} Ne8 {+M50/1 0s} 16. Bc2+ {0s} Kf7 {+M49/1 0s}
17. Kg5 {0s} Ke6 {+M48/1 0s} 18. Bf2 {1s} Nc7 {+M47/1 0s}
19. Bb3+ {0s} Kd6 {+M46/1 0s} 20. Bg3+ {0s} Kc6 {+M45/1 0s}
21. Ba4+ {0s} Kb6 {+M44/1 0s} 22. Kf5 {0s} Nb5 {+M43/1 0s}
23. Bf2+ {0s} Ka5 {+M42/1 0s} 24. Bb3 {0s} Kb4 {+M41/1 0s}
25. Be6 {0s} Kc3 {+M40/1 0s} 26. Ke4 {0s} Kc2 {+M39/1 0s}
27. Ke3 {0s} Nc3 {+M38/1 0s} 28. Bf5+ {0s} Kb3 {+M37/1 0s}
29. Kd3 {0s} Na4 {+M36/1 0s} 30. Be6+ {0s} Kb4 {+M35/1 0s}
31. Kd4 {1s} Nc5 {+M34/1 0s} 32. Bf5 {0s} Kb5 {+M33/1 0s}
33. Bg4 {0s} Na4 {+M32/1 0s} 34. Bh5 {0s} Nb6 {+M31/1 0s}
35. Be8+ {0s} Kb4 {+M30/1 0s} 36. Be1+ {0s} Kb3 {+M29/1 0s}
37. Kd3 {0s} Ka3 {+M28/1 0s} 38. Bh5 {0s} Ka4 {+M27/1 0s}
39. Kd4 {0s} Kb5 {+M26/1 0s} 40. Be8+ {1s} Ka6 {+M25/1 0s}
41. Bc6 {0s} Ka7 {+M24/1 0s} 42. Bg2 {0s} Ka6 {+M23/1 0s}
43. Kc5 {0s} Nc8 {+M22/1 0s} 44. Bh4 {0s} Na7 {+M21/1 0s}
45. Bd8 {0s} Nc8 {+M20/1 0s} 46. Kc6 {0s} Na7+ {+M19/1 0s}
47. Kc7 {0s} Ka5 {+M18/1 0s} 48. Kb7+ {0s} Kb4 {+M17/1 0s}
49. Kxa7 {0s} Kc5 {+M16/1 0s} 50. Bc7 {1s} Kb5 {+M15/1 0s}
51. Bb6 {0s} Kb4 {+M14/1 0s} 52. Ka6 {0s} Kc4 {+M13/1 0s}
53. Ka5 {0s} Kb3 {+M12/1 0s} 54. Bf1 {0s} Ka3 {+M11/1 0s}
55. Bf2 {0s} Kb3 {+M10/1 0s} 56. Be1 {0s} Ka3 {+M9/1 0s}
57. Bc4 {0s} Kb2 {+M8/1 0s} 58. Ka4 {0s} Kc2 {+M7/1 0s}
59. Bb4 {0s} Kb2 {+M6/1 0s} 60. Bb3 {0s} Ka1 {+M5/1 0s}
61. Ka3 {0s} Kb1 {+M4/1 0s} 62. Bd2 {0s} Ka1 {+M3/1 0s}
63. Bg8 {0s} Kb1 {+M2/1 0s} 64. Bh7+ {0s} Ka1 {+M1/1 0s}
65. Bc3# {0s} 1-0

And now my SYZ-Fish is doing this:

[Event "1 Minutes/Game + 1 Seconds/Move"]
[Site "Engine Match"]
[Date "2014.01.19"]
[Round "1"]
[White "Stockfish 180114 SYZ"]
[Black "DS12"]
[Result "1/2-1/2"]
[SetUp "1"]
[FEN "B7/8/8/8/5nk1/8/8/4K1B1 w - -"]

1. Kd2 {+0.01/30 2s} Kf5 {+M64/1 0s} 2. Bh2 {+0.01/22 0s}
Kg5 {+M63/1 0s (Ng6)} 3. Ke3 {+0.01/35 1s} Nh5 {+M62/1 0s
(Ne6)} 4. Bb7 {+0.48/30 2s} Kf5 {+M61/1 0s (Kf6)} 5. Bc8+
{+0.48/32 1s} Kg5 {+M60/1 0s (Kf6)} 6. Kd4 {+0.48/1 0s} Kg6
{+M59/1 0s (Nf6)} 7. Ke5 {+0.48/1 0s} Ng7 {+M58/1 0s (Kg7)}
8. Bg1 {+0.48/39 1s} Kg5 {+M57/1 0s (Nh5)} 9. Bd4 {+0.48/40
1s} Kg6 {+M56/1 0s} 10. Bd7 {+0.48/43 1s} Kg5 {+M55/1 0s}
11. Ba4 {+0.48/32 0s} Kg6 {+M54/1 0s} 12. Bb3 {+0.48/31 0s}
Nh5 {+M53/1 0s} 13. Bc2+ {+0.48/41 1s} Kf7 {+M52/1 0s}
14. Kd6 {+0.48/28 0s} Ng7 {+M51/1 0s} 15. Kd7 {+0.48/29 0s}
Nh5 {+M50/1 0s} 16. Bb3+ {+0.48/23 0s} Kg6 {+M49/1 0s}
17. Ke7 {+0.48/18 0s} Kf5 {+M48/1 0s} 18. Bc2+ {+0.48/34
2s} Kf4 {+M47/1 0s} 19. Ke6 {+0.48/36 1s} Ng3 {+M47/1 0s}
20. Be5+ {+0.48/41 2s} Kf3 {+M46/1 0s} 21. Bb8 {+0.48/16
0s} Ne2 {+M45/1 0s (Nf1)} 22. Ke5 {+0.48/1 0s} Ke3 {+M44/1
0s} 23. Be4 {+0.48/17 0s} Nc3 {+M44/1 0s (Nc1)} 24. Bf5
{+0.48/32 2s} Nb5 {+M43/1 0s} 25. Bg6 {+0.48/22 0s} Kd2
{+M42/1 0s} 26. Kd5 {+0.48/23 0s} Nc3+ {+M41/1 0s (Na3)}
27. Kd4 {+0.48/1 0s} Ne2+ {+M40/1 0s} 28. Ke5 {+0.48/24 0s}
Ke3 {+M39/1 0s (Kc3)} 29. Ba7+ {+0.48/1 0s} Kd2 {+M38/1 0s}
30. Bb6 {+0.48/21 0s} Kc3 {+M37/1 0s} 31. Ke4 {+0.48/18 0s}
Kb4 {+M36/1 0s} 32. Bd8 {+0.48/16 0s} Kc5 {+M35/1 0s (Kc4)}
33. Be7+ {+0.48/32 5s} Kc4 {+M34/1 0s} 34. Bf7+ {+0.48/32
8s} Kc3 {+M33/1 0s} 35. Ke3 {+0.48/24 0s} Nc1 {+M32/1 0s}
36. Bh4 {+0.48/29 0s} Nb3 {+M31/1 0s} 37. Be1+ {+0.48/30
0s} Kb2 {+M30/1 0s} 38. Kd3 {+0.48/36 2s} Nc1+ {+M29/1 0s}
39. Kc4 {+0.48/35 0s} Kc2 {+M28/1 0s} 40. Ba5 {+0.48/26 0s}
Kd1 {+M27/1 0s (Ne2)} 41. Be6 {+0.48/1 0s} Kc2 {+M26/1 0s
(Ke2)} 42. Bf5+ {+0.48/1 0s} Kb2 {+M25/1 0s} 43. Bc3+
{+0.48/24 6s} Ka3 {+M24/1 0s (Ka2)} 44. Bc2 {+0.48/44 2s}
Ka2 {+M23/1 0s (Ne2)} 45. Be5 {+0.48/1 0s} Ka3 {+M22/1 0s}
46. Bd1 {+0.48/47 0s} Na2 {+M21/1 0s (Ka2)} 47. Bg7
{+0.48/1 0s} Nc1 {+M20/1 0s} 48. Kc3 {+0.48/80 0s} Na2+
{+M19/1 0s} 49. Kc2 {+0.48/100 0s} Kb4 {+M18/1 0s (Nb4+)}
50. Kb2 {+0.50/1 0s} Kc4 {+M17/1 0s} 1/2-1/2

Draw because of 50 move rule (which I don't like) It doesn't find the capturing move at move 49?

Am I doing somethign wrong, are my SYZ Bases (who invented that name?) defective, but why is this working in the first example?

I am a bit puzzled.

Thx for any help
Ingo
syzygy
Posts: 5774
Joined: Tue Feb 28, 2012 11:56 pm

Re: SYZYGY Base question

Post by syzygy »

I've added pgn tags.
Nothing special. #51 with a capture in move 35. This is what SYZ-Fish is doing with it against an optimal Shredder with Nalimovs:

[pgn][Event "1 Minutes/Game + 1 Seconds/Move"]
[Site "Engine Match"]
[Date "2014.01.19"]
[Round "1"]
[White "Stockfish 180114 SYZ"]
[Black "Deep Shredder 12"]
[Result "1-0"]
[SetUp "1"]
[FEN "8/6n1/6k1/8/3B1K2/1B6/8/8 w - - 0 1"]

1. Ke5 {+151.00/34 4s} Nh5 {+M53/1 8s} 2. Bc2+ {+151.00/44
0s} Kf7 {+M52/1 0s} 3. Bd1 {+151.00/43 1s} Ng7 {+M54/1 0s
(Kg6)} 4. Bb3+ {+151.00/44 1s} Kg6 {+M55/1 0s} 5. Bc2+
{+151.00/45 1s} Kf7 {+M55/1 0s} 6. Bd1 {+151.00/45 1s} Kg6
{+M54/1 0s} 7. Be3 {+151.00/46 1s} Kf7 {+M55/1 0s (Ne8)}
8. Bb3+ {+151.00/44 1s} Kg6 {+M54/1 0s} 9. Bd4 {+151.00/100
0s} Nh5 {+M53/1 0s} 10. Kd6 {+151.00/38 1s} Ng7 {+M52/1 0s}
11. Bc2+ {+151.00/31 0s} Kf7 {+M51/1 0s} 12. Kd7
{+151.00/29 0s} Nh5 {+M50/1 0s} 13. Bb3+ {+151.00/29 0s}
Kg6 {+M49/1 0s} 14. Ke7 {+151.00/20 0s} Kf5 {+M48/1 0s}
15. Bc2+ {+151.00/34 2s} Kf4 {+M47/1 0s} 16. Ke6
{+151.00/38 1s} Ng3 {+M47/1 0s} 17. Be5+ {+151.00/35 1s}
Kf3 {+M46/1 0s} 18. Bb8 {+151.00/19 0s} Ne2 {+M45/1 0s}
19. Ke5 {+151.00/21 0s} Ke3 {+M44/1 0s} 20. Be4 {+151.00/21
0s} Nc3 {+M44/1 0s} 21. Bh7 {+151.00/36 1s} Nb5 {+M43/1 0s}
22. Bg6 {+151.00/28 0s} Kd2 {+M42/1 0s} 23. Kd5 {+151.00/30
0s} Nc3+ {+M41/1 0s} 24. Kd4 {+151.00/32 0s} Ne2+ {+M40/1
0s} 25. Ke5 {+151.00/32 0s} Ke3 {+M39/1 0s} 26. Ba7+
{+151.00/24 0s} Kd2 {+M38/1 0s} 27. Bb6 {+151.00/18 0s} Kc3
{+M37/1 0s} 28. Ke4 {+151.00/19 0s} Kb4 {+M36/1 0s} 29. Bd8
{+151.00/16 0s} Kc5 {+M35/1 0s (Nc3+)} 30. Be7+ {+151.00/32
2s} Kc4 {+M34/1 0s} 31. Bf7+ {+151.00/35 2s} Kc3 {+M33/1
0s} 32. Ke3 {+151.00/36 0s} Nc1 {+M32/1 0s} 33. Bh4
{+151.00/27 0s} Nb3 {+M31/1 0s} 34. Be1+ {+151.00/28 0s}
Kb2 {+M30/1 0s} 35. Kd3 {+151.00/36 3s} Nc1+ {+M29/1 0s}
36. Kc4 {+151.00/36 0s} Kc2 {+M28/1 0s} 37. Ba5 {+151.00/25
0s} Kd1 {+M27/1 0s (Ne2)} 38. Be6 {+151.00/1 0s} Kc2
{+M26/1 0s (Ne2)} 39. Bf5+ {+151.00/1 0s} Kb2 {+M25/1 0s}
40. Bd2 {+151.00/33 3s} Na2 {+M24/1 0s (Ne2)} 41. Bh6
{+151.00/37 19s} Ka3 {+M23/1 0s} 42. Bg7 {+151.00/17 0s}
Nc1 {+M22/1 0s} 43. Bg4 {+151.00/15 1s} Na2 {+M21/1 0s}
44. Bd1 {+151.00/24 0s} Nc1 {+M20/1 0s} 45. Kc3 {+151.00/24
0s} Na2+ {+M19/1 0s} 46. Kc2 {+151.00/25 0s} Kb4 {+M18/1
0s} 47. Kb2 {+151.00/23 0s} Kc4 {+M17/1 0s (Kb5)} 48. Kxa2
{+151.00/1 0s} Kd3 {+M16/1 0s (Kd5)} 49. Kb3 {+151.00/26
2s} Kd2 {+M15/1 0s (Ke4)} 50. Bc2 {+M16/33 2s} Ke1 {+M14/1
0s (Ke2)} 51. Kc3 {+M14/36 3s} Ke2 {+M13/1 0s} 52. Bh6
{+M13/36 2s} Ke1 {+M12/1 0s} 53. Kd3 {+M12/37 2s} Kf2
{+M11/1 0s} 54. Ke4 {+M11/37 2s} Kg2 {+M10/1 0s} 55. Ke3
{+M10/38 3s} Kg3 {+M9/1 0s} 56. Bf4+ {+M9/39 2s} Kg2 {+M8/1
0s (Kg4)} 57. Bg6 {+M8/41 2s} Kf1 {+M7/1 0s (Kh3)} 58. Kf3
{+M7/45 1s} Ke1 {+M6/1 0s} 59. Bc2 {+M6/64 1s} Kf1 {+M5/1
0s} 60. Bd2 {+M5/100 1s} Kg1 {+M4/1 0s} 61. Kg3 {+M4/100
0s} Kf1 {+M3/1 0s} 62. Bd3+ {+M3/100 0s} Kg1 {+M2/1 0s}
63. Be3+ {+M2/100 0s} Kh1 {+M1/1 0s} 64. Be4# {+M1/100 0s}
1-0[/pgn]

SYZ-Fish needs 14 moves more than nessesary but this is just to prove that my SYS-Fish is using the SYZYGY Bases.
They are not DTM but DTZ50, which means they don't go for the fastest mate but go for the fastest capture or pawn move (by either side) as these moves reset the 50-move counter.
Mate in 65 and capturing a piece in move 49. Here is a perfect Nalimove game out of this:

[pgn][Event "TEST"]
[Site "HOME"]
[Date "2014.01.19"]
[Round "1"]
[White "NalimovA"]
[Black "NalimovB"]
[Result "1-0"]
[SetUp "1"]
[FEN "B7/8/8/8/5nk1/8/8/4K1B1 w - - 0 1"]

1. Bh2 {0s} Kg5 {+M64/1 0s} 2. Kd2 {0s} Kf5 {+M63/1 0s}
3. Ke3 {0s} Nh5 {+M62/1 0s} 4. Bg2 {0s} Kg5 {+M61/1 0s}
5. Kd4 {1s} Kf5 {+M60/1 0s} 6. Bh3+ {0s} Kg5 {+M59/1 0s}
7. Ke5 {0s} Ng7 {+M58/1 0s} 8. Bf1 {0s} Kg6 {+M57/1 0s}
9. Bg1 {0s} Kf7 {+M56/1 0s} 10. Bd4 {0s} Kg6 {+M55/1 0s}
11. Bd3+ {1s} Kf7 {+M54/1 0s} 12. Bc2 {0s} Ne8 {+M53/1 0s}
13. Bb3+ {0s} Kg6 {+M52/1 0s} 14. Kf4 {0s} Ng7 {+M51/1 0s}
15. Kg4 {0s} Ne8 {+M50/1 0s} 16. Bc2+ {0s} Kf7 {+M49/1 0s}
17. Kg5 {0s} Ke6 {+M48/1 0s} 18. Bf2 {1s} Nc7 {+M47/1 0s}
19. Bb3+ {0s} Kd6 {+M46/1 0s} 20. Bg3+ {0s} Kc6 {+M45/1 0s}
21. Ba4+ {0s} Kb6 {+M44/1 0s} 22. Kf5 {0s} Nb5 {+M43/1 0s}
23. Bf2+ {0s} Ka5 {+M42/1 0s} 24. Bb3 {0s} Kb4 {+M41/1 0s}
25. Be6 {0s} Kc3 {+M40/1 0s} 26. Ke4 {0s} Kc2 {+M39/1 0s}
27. Ke3 {0s} Nc3 {+M38/1 0s} 28. Bf5+ {0s} Kb3 {+M37/1 0s}
29. Kd3 {0s} Na4 {+M36/1 0s} 30. Be6+ {0s} Kb4 {+M35/1 0s}
31. Kd4 {1s} Nc5 {+M34/1 0s} 32. Bf5 {0s} Kb5 {+M33/1 0s}
33. Bg4 {0s} Na4 {+M32/1 0s} 34. Bh5 {0s} Nb6 {+M31/1 0s}
35. Be8+ {0s} Kb4 {+M30/1 0s} 36. Be1+ {0s} Kb3 {+M29/1 0s}
37. Kd3 {0s} Ka3 {+M28/1 0s} 38. Bh5 {0s} Ka4 {+M27/1 0s}
39. Kd4 {0s} Kb5 {+M26/1 0s} 40. Be8+ {1s} Ka6 {+M25/1 0s}
41. Bc6 {0s} Ka7 {+M24/1 0s} 42. Bg2 {0s} Ka6 {+M23/1 0s}
43. Kc5 {0s} Nc8 {+M22/1 0s} 44. Bh4 {0s} Na7 {+M21/1 0s}
45. Bd8 {0s} Nc8 {+M20/1 0s} 46. Kc6 {0s} Na7+ {+M19/1 0s}
47. Kc7 {0s} Ka5 {+M18/1 0s} 48. Kb7+ {0s} Kb4 {+M17/1 0s}
49. Kxa7 {0s} Kc5 {+M16/1 0s} 50. Bc7 {1s} Kb5 {+M15/1 0s}
51. Bb6 {0s} Kb4 {+M14/1 0s} 52. Ka6 {0s} Kc4 {+M13/1 0s}
53. Ka5 {0s} Kb3 {+M12/1 0s} 54. Bf1 {0s} Ka3 {+M11/1 0s}
55. Bf2 {0s} Kb3 {+M10/1 0s} 56. Be1 {0s} Ka3 {+M9/1 0s}
57. Bc4 {0s} Kb2 {+M8/1 0s} 58. Ka4 {0s} Kc2 {+M7/1 0s}
59. Bb4 {0s} Kb2 {+M6/1 0s} 60. Bb3 {0s} Ka1 {+M5/1 0s}
61. Ka3 {0s} Kb1 {+M4/1 0s} 62. Bd2 {0s} Ka1 {+M3/1 0s}
63. Bg8 {0s} Kb1 {+M2/1 0s} 64. Bh7+ {0s} Ka1 {+M1/1 0s}
65. Bc3# {0s} 1-0[/pgn]
This position is in fact a draw under the 50-move rule. NalimovB somewhere screwed up and allowed a capture at move 49. This would not happen if you let black use Syzygybases.
And now my SYZ-Fish is doing this:

[pgn][Event "1 Minutes/Game + 1 Seconds/Move"]
[Site "Engine Match"]
[Date "2014.01.19"]
[Round "1"]
[White "Stockfish 180114 SYZ"]
[Black "DS12"]
[Result "1/2-1/2"]
[SetUp "1"]
[FEN "B7/8/8/8/5nk1/8/8/4K1B1 w - - 0 1"]

1. Kd2 {+0.01/30 2s} Kf5 {+M64/1 0s} 2. Bh2 {+0.01/22 0s}
Kg5 {+M63/1 0s (Ng6)} 3. Ke3 {+0.01/35 1s} Nh5 {+M62/1 0s
(Ne6)} 4. Bb7 {+0.48/30 2s} Kf5 {+M61/1 0s (Kf6)} 5. Bc8+
{+0.48/32 1s} Kg5 {+M60/1 0s (Kf6)} 6. Kd4 {+0.48/1 0s} Kg6
{+M59/1 0s (Nf6)} 7. Ke5 {+0.48/1 0s} Ng7 {+M58/1 0s (Kg7)}
8. Bg1 {+0.48/39 1s} Kg5 {+M57/1 0s (Nh5)} 9. Bd4 {+0.48/40
1s} Kg6 {+M56/1 0s} 10. Bd7 {+0.48/43 1s} Kg5 {+M55/1 0s}
11. Ba4 {+0.48/32 0s} Kg6 {+M54/1 0s} 12. Bb3 {+0.48/31 0s}
Nh5 {+M53/1 0s} 13. Bc2+ {+0.48/41 1s} Kf7 {+M52/1 0s}
14. Kd6 {+0.48/28 0s} Ng7 {+M51/1 0s} 15. Kd7 {+0.48/29 0s}
Nh5 {+M50/1 0s} 16. Bb3+ {+0.48/23 0s} Kg6 {+M49/1 0s}
17. Ke7 {+0.48/18 0s} Kf5 {+M48/1 0s} 18. Bc2+ {+0.48/34
2s} Kf4 {+M47/1 0s} 19. Ke6 {+0.48/36 1s} Ng3 {+M47/1 0s}
20. Be5+ {+0.48/41 2s} Kf3 {+M46/1 0s} 21. Bb8 {+0.48/16
0s} Ne2 {+M45/1 0s (Nf1)} 22. Ke5 {+0.48/1 0s} Ke3 {+M44/1
0s} 23. Be4 {+0.48/17 0s} Nc3 {+M44/1 0s (Nc1)} 24. Bf5
{+0.48/32 2s} Nb5 {+M43/1 0s} 25. Bg6 {+0.48/22 0s} Kd2
{+M42/1 0s} 26. Kd5 {+0.48/23 0s} Nc3+ {+M41/1 0s (Na3)}
27. Kd4 {+0.48/1 0s} Ne2+ {+M40/1 0s} 28. Ke5 {+0.48/24 0s}
Ke3 {+M39/1 0s (Kc3)} 29. Ba7+ {+0.48/1 0s} Kd2 {+M38/1 0s}
30. Bb6 {+0.48/21 0s} Kc3 {+M37/1 0s} 31. Ke4 {+0.48/18 0s}
Kb4 {+M36/1 0s} 32. Bd8 {+0.48/16 0s} Kc5 {+M35/1 0s (Kc4)}
33. Be7+ {+0.48/32 5s} Kc4 {+M34/1 0s} 34. Bf7+ {+0.48/32
8s} Kc3 {+M33/1 0s} 35. Ke3 {+0.48/24 0s} Nc1 {+M32/1 0s}
36. Bh4 {+0.48/29 0s} Nb3 {+M31/1 0s} 37. Be1+ {+0.48/30
0s} Kb2 {+M30/1 0s} 38. Kd3 {+0.48/36 2s} Nc1+ {+M29/1 0s}
39. Kc4 {+0.48/35 0s} Kc2 {+M28/1 0s} 40. Ba5 {+0.48/26 0s}
Kd1 {+M27/1 0s (Ne2)} 41. Be6 {+0.48/1 0s} Kc2 {+M26/1 0s
(Ke2)} 42. Bf5+ {+0.48/1 0s} Kb2 {+M25/1 0s} 43. Bc3+
{+0.48/24 6s} Ka3 {+M24/1 0s (Ka2)} 44. Bc2 {+0.48/44 2s}
Ka2 {+M23/1 0s (Ne2)} 45. Be5 {+0.48/1 0s} Ka3 {+M22/1 0s}
46. Bd1 {+0.48/47 0s} Na2 {+M21/1 0s (Ka2)} 47. Bg7
{+0.48/1 0s} Nc1 {+M20/1 0s} 48. Kc3 {+0.48/80 0s} Na2+
{+M19/1 0s} 49. Kc2 {+0.48/100 0s} Kb4 {+M18/1 0s (Nb4+)}
50. Kb2 {+0.50/1 0s} Kc4 {+M17/1 0s} 1/2-1/2[/pgn]

Draw because of 50 move rule (which I don't like) It doesn't find the capturing move at move 49?
There was no forced capture at move 49 or earlier. In the other game NalimovB screwed up and allowed such a capture, but from the start there was never a guarantee that this would be possible.
IWB
Posts: 1539
Joined: Thu Mar 09, 2006 2:02 pm

Re: SYZYGY Base question

Post by IWB »

I remove the PGN tags as no one can use this without the EPD in a GUI.

[/quote="syzygy"]

[d]B7/8/8/8/5nk1/8/8/4K1B1 w - - 0 1"[d]

This position is in fact a draw under the 50-move rule. NalimovB somewhere screwed up and allowed a capture at move 49. This would not happen if you let black use Syzygybases.
[/quote="syzygy"]

OK, here is Shredder with White and Nalimovs and Stockfish with Black SYS bases:

[Event "1 Minutes/Game + 1 Seconds/Move"]
[Site "Engine Match"]
[Date "2014.01.19"]
[Round "1"]
[White "Stockfish 180114 SYZ 1T"]
[Black "DS12 1T"]
[Result "1/2-1/2"]
[SetUp "1"]
[FEN "B7/8/8/8/5nk1/8/8/4K1B1 w - -"]

1. Kd2 {+0.01/30 2s} Kf5 {+M64/1 0s} 2. Bh2 {+0.01/22 0s}
Kg5 {+M63/1 0s (Ng6)} 3. Ke3 {+0.01/35 1s} Nh5 {+M62/1 0s
(Ne6)} 4. Bb7 {+0.48/30 2s} Kf5 {+M61/1 0s (Kf6)} 5. Bc8+
{+0.48/32 1s} Kg5 {+M60/1 0s (Kf6)} 6. Kd4 {+0.48/1 0s} Kg6
{+M59/1 0s (Nf6)} 7. Ke5 {+0.48/1 0s} Ng7 {+M58/1 0s (Kg7)}
8. Bg1 {+0.48/39 1s} Kg5 {+M57/1 0s (Nh5)} 9. Bd4 {+0.48/40
1s} Kg6 {+M56/1 0s} 10. Bd7 {+0.48/43 1s} Kg5 {+M55/1 0s}
11. Ba4 {+0.48/32 0s} Kg6 {+M54/1 0s} 12. Bb3 {+0.48/31 0s}
Nh5 {+M53/1 0s} 13. Bc2+ {+0.48/41 1s} Kf7 {+M52/1 0s}
14. Kd6 {+0.48/28 0s} Ng7 {+M51/1 0s} 15. Kd7 {+0.48/29 0s}
Nh5 {+M50/1 0s} 16. Bb3+ {+0.48/23 0s} Kg6 {+M49/1 0s}
17. Ke7 {+0.48/18 0s} Kf5 {+M48/1 0s} 18. Bc2+ {+0.48/34
2s} Kf4 {+M47/1 0s} 19. Ke6 {+0.48/36 1s} Ng3 {+M47/1 0s}
20. Be5+ {+0.48/41 2s} Kf3 {+M46/1 0s} 21. Bb8 {+0.48/16
0s} Ne2 {+M45/1 0s (Nf1)} 22. Ke5 {+0.48/1 0s} Ke3 {+M44/1
0s} 23. Be4 {+0.48/17 0s} Nc3 {+M44/1 0s (Nc1)} 24. Bf5
{+0.48/32 2s} Nb5 {+M43/1 0s} 25. Bg6 {+0.48/22 0s} Kd2
{+M42/1 0s} 26. Kd5 {+0.48/23 0s} Nc3+ {+M41/1 0s (Na3)}
27. Kd4 {+0.48/1 0s} Ne2+ {+M40/1 0s} 28. Ke5 {+0.48/24 0s}
Ke3 {+M39/1 0s (Kc3)} 29. Ba7+ {+0.48/1 0s} Kd2 {+M38/1 0s}
30. Bb6 {+0.48/21 0s} Kc3 {+M37/1 0s} 31. Ke4 {+0.48/18 0s}
Kb4 {+M36/1 0s} 32. Bd8 {+0.48/16 0s} Kc5 {+M35/1 0s (Kc4)}
33. Be7+ {+0.48/32 5s} Kc4 {+M34/1 0s} 34. Bf7+ {+0.48/32
8s} Kc3 {+M33/1 0s} 35. Ke3 {+0.48/24 0s} Nc1 {+M32/1 0s}
36. Bh4 {+0.48/29 0s} Nb3 {+M31/1 0s} 37. Be1+ {+0.48/30
0s} Kb2 {+M30/1 0s} 38. Kd3 {+0.48/36 2s} Nc1+ {+M29/1 0s}
39. Kc4 {+0.48/35 0s} Kc2 {+M28/1 0s} 40. Ba5 {+0.48/26 0s}
Kd1 {+M27/1 0s (Ne2)} 41. Be6 {+0.48/1 0s} Kc2 {+M26/1 0s
(Ke2)} 42. Bf5+ {+0.48/1 0s} Kb2 {+M25/1 0s} 43. Bc3+
{+0.48/24 6s} Ka3 {+M24/1 0s (Ka2)} 44. Bc2 {+0.48/44 2s}
Ka2 {+M23/1 0s (Ne2)} 45. Be5 {+0.48/1 0s} Ka3 {+M22/1 0s}
46. Bd1 {+0.48/47 0s} Na2 {+M21/1 0s (Ka2)} 47. Bg7
{+0.48/1 0s} Nc1 {+M20/1 0s} 48. Kc3 {+0.48/80 0s} Na2+
{+M19/1 0s} 49. Kc2 {+0.48/100 0s} Kb4 {+M18/1 0s (Nb4+)}
50. Kb2 {+0.50/1 0s} Kc4 {+M17/1 0s} 1/2-1/2

Draw!

And now I start to see the point. NalimovB Did not make an error (or it did, depends on the view). It is jsut that the Nalimoves always play the shortest to a mate but sometimes there are more moves with the same mate count. Then it simply pickes the first and not the second which might push the capture behind the 50 move rule ...


I will check the first position now.

Thx Ingo






syzygy wrote:
And now my SYZ-Fish is doing this:

[pgn][Event "1 Minutes/Game + 1 Seconds/Move"]
[Site "Engine Match"]
[Date "2014.01.19"]
[Round "1"]
[White "Stockfish 180114 SYZ"]
[Black "DS12"]
[Result "1/2-1/2"]
[SetUp "1"]
[FEN "B7/8/8/8/5nk1/8/8/4K1B1 w - - 0 1"]

1. Kd2 {+0.01/30 2s} Kf5 {+M64/1 0s} 2. Bh2 {+0.01/22 0s}
Kg5 {+M63/1 0s (Ng6)} 3. Ke3 {+0.01/35 1s} Nh5 {+M62/1 0s
(Ne6)} 4. Bb7 {+0.48/30 2s} Kf5 {+M61/1 0s (Kf6)} 5. Bc8+
{+0.48/32 1s} Kg5 {+M60/1 0s (Kf6)} 6. Kd4 {+0.48/1 0s} Kg6
{+M59/1 0s (Nf6)} 7. Ke5 {+0.48/1 0s} Ng7 {+M58/1 0s (Kg7)}
8. Bg1 {+0.48/39 1s} Kg5 {+M57/1 0s (Nh5)} 9. Bd4 {+0.48/40
1s} Kg6 {+M56/1 0s} 10. Bd7 {+0.48/43 1s} Kg5 {+M55/1 0s}
11. Ba4 {+0.48/32 0s} Kg6 {+M54/1 0s} 12. Bb3 {+0.48/31 0s}
Nh5 {+M53/1 0s} 13. Bc2+ {+0.48/41 1s} Kf7 {+M52/1 0s}
14. Kd6 {+0.48/28 0s} Ng7 {+M51/1 0s} 15. Kd7 {+0.48/29 0s}
Nh5 {+M50/1 0s} 16. Bb3+ {+0.48/23 0s} Kg6 {+M49/1 0s}
17. Ke7 {+0.48/18 0s} Kf5 {+M48/1 0s} 18. Bc2+ {+0.48/34
2s} Kf4 {+M47/1 0s} 19. Ke6 {+0.48/36 1s} Ng3 {+M47/1 0s}
20. Be5+ {+0.48/41 2s} Kf3 {+M46/1 0s} 21. Bb8 {+0.48/16
0s} Ne2 {+M45/1 0s (Nf1)} 22. Ke5 {+0.48/1 0s} Ke3 {+M44/1
0s} 23. Be4 {+0.48/17 0s} Nc3 {+M44/1 0s (Nc1)} 24. Bf5
{+0.48/32 2s} Nb5 {+M43/1 0s} 25. Bg6 {+0.48/22 0s} Kd2
{+M42/1 0s} 26. Kd5 {+0.48/23 0s} Nc3+ {+M41/1 0s (Na3)}
27. Kd4 {+0.48/1 0s} Ne2+ {+M40/1 0s} 28. Ke5 {+0.48/24 0s}
Ke3 {+M39/1 0s (Kc3)} 29. Ba7+ {+0.48/1 0s} Kd2 {+M38/1 0s}
30. Bb6 {+0.48/21 0s} Kc3 {+M37/1 0s} 31. Ke4 {+0.48/18 0s}
Kb4 {+M36/1 0s} 32. Bd8 {+0.48/16 0s} Kc5 {+M35/1 0s (Kc4)}
33. Be7+ {+0.48/32 5s} Kc4 {+M34/1 0s} 34. Bf7+ {+0.48/32
8s} Kc3 {+M33/1 0s} 35. Ke3 {+0.48/24 0s} Nc1 {+M32/1 0s}
36. Bh4 {+0.48/29 0s} Nb3 {+M31/1 0s} 37. Be1+ {+0.48/30
0s} Kb2 {+M30/1 0s} 38. Kd3 {+0.48/36 2s} Nc1+ {+M29/1 0s}
39. Kc4 {+0.48/35 0s} Kc2 {+M28/1 0s} 40. Ba5 {+0.48/26 0s}
Kd1 {+M27/1 0s (Ne2)} 41. Be6 {+0.48/1 0s} Kc2 {+M26/1 0s
(Ke2)} 42. Bf5+ {+0.48/1 0s} Kb2 {+M25/1 0s} 43. Bc3+
{+0.48/24 6s} Ka3 {+M24/1 0s (Ka2)} 44. Bc2 {+0.48/44 2s}
Ka2 {+M23/1 0s (Ne2)} 45. Be5 {+0.48/1 0s} Ka3 {+M22/1 0s}
46. Bd1 {+0.48/47 0s} Na2 {+M21/1 0s (Ka2)} 47. Bg7
{+0.48/1 0s} Nc1 {+M20/1 0s} 48. Kc3 {+0.48/80 0s} Na2+
{+M19/1 0s} 49. Kc2 {+0.48/100 0s} Kb4 {+M18/1 0s (Nb4+)}
50. Kb2 {+0.50/1 0s} Kc4 {+M17/1 0s} 1/2-1/2[/pgn]

Draw because of 50 move rule (which I don't like) It doesn't find the capturing move at move 49?
There was no forced capture at move 49 or earlier. In the other game NalimovB screwed up and allowed such a capture, but from the start there was never a guarantee that this would be possible.
syzygy
Posts: 5774
Joined: Tue Feb 28, 2012 11:56 pm

Re: SYZYGY Base question

Post by syzygy »

IWB wrote:I remove the PGN tags as no one can use this without the EPD in a GUI.
Just click on the b8 square.
And now I start to see the point. NalimovB Did not make an error (or it did, depends on the view). It is jsut that the Nalimoves always play the shortest to a mate but sometimes there are more moves with the same mate count. Then it simply pickes the first and not the second which might push the capture behind the 50 move rule ...
Yes, or it might be that in some position any of the "best" moves (in terms of DTM) give up the 50-move draw.

See e.g. here where Nalimov suggests to play 1...Kg6 which loses, whereas 1...Kf6 preserves the draw.
IWB
Posts: 1539
Joined: Thu Mar 09, 2006 2:02 pm

Re: SYZYGY Base question

Post by IWB »

Ok, this needs more attention:

[d]

Accoring to ofline and online Nalimove this is a mate in #51 with a sure capture before the 50 move rule. So the 50 move rule is not important here.

This is

SYS-FISH with Black and Shredder with Nalimovs:

[Event "1 Minutes/Game + 1 Seconds/Move"]
[Site "Engine Match"]
[Date "2014.01.19"]
[Round "1"]
[White "DS12 1T"]
[Black "Stockfish 180114 SYZ 1T"]
[Result "1-0"]
[SetUp "1"]
[FEN "8/6n1/6k1/8/3B1K2/1B6/8/8 w - -"]

1. Kg4 {+M51/1 0s} Ne8 {+151.00/1 0s} 2. Bc2+ {+M50/1 0s}
Kf7 {+151.00/1 0s} 3. Kg5 {+M49/1 0s} Ke6 {+151.00/1 0s}
4. Bb3+ {+M48/1 0s (Bf5+)} Kd6 {+151.00/1 0s} 5. Bf2
{+M47/1 0s} Nc7 {+151.00/1 0s} 6. Bg3+ {+M46/1 0s} Kc6
{+151.00/17 0s} 7. Ba4+ {+M45/1 0s (Kf5)} Kb6 {+151.00/1
0s} 8. Kf5 {+M44/1 0s (Be5)} Nb5 {+151.00/1 0s} 9. Bf2+
{+M43/1 0s} Ka5 {+151.00/19 0s} 10. Bb3 {+M42/1 0s (Bc2)}
Kb4 {+151.00/1 0s} 11. Be6 {+M41/1 0s} Kc3 {+151.00/1 0s}
12. Ke4 {+M40/1 0s} Kb4 {+151.00/1 0s} 13. Bg8 {+M39/1 0s}
Nd6+ {+151.00/1 0s} 14. Kd5 {+M38/1 0s (Ke5)} Nc4
{+151.00/1 0s} 15. Bf7 {+M37/1 0s (Be6)} Kc3 {+151.00/1 0s}
16. Bd4+ {+M36/1 0s} Kb4 {+151.00/21 0s} 17. Bf6 {+M35/1 0s
(Bg6)} Nd2 {+151.00/1 0s} 18. Bg6 {+M34/1 0s} Nf3
{+151.00/1 0s} 19. Bd8 {+M33/1 0s} Nd2 {+151.00/1 0s}
20. Bh4 {+M32/1 0s (Be7+)} Kb3 {+151.00/1 0s} 21. Kc5
{+M31/1 0s} Nf3 {+151.00/1 0s} 22. Bf6 {+M30/1 0s} Ng1
{+151.00/1 0s} 23. Bh5 {+M29/1 0s (Bd4)} Kc2 {+151.00/1 0s}
24. Kd4 {+M28/1 0s} Kd2 {+151.00/1 0s} 25. Bg5+ {+M27/1 0s}
Ke1 {+151.00/22 0s} 26. Ke3 {+M26/1 0s} Kf1 {+151.00/28 0s}
27. Bh4 {+M25/1 0s (Be7)} Kg2 {+151.00/1 0s} 28. Bg4
{+M24/1 0s} Kh2 {+151.00/38 3s} 29. Kf2 {+M20/1 0s (Be7)}
Nh3+ {+151.00/1 0s} 30. Ke2 {+M19/1 0s (Kf3)} Ng1+
{+151.00/1 0s} 31. Kf1 {+M18/1 0s (Ke3)} Nh3 {+151.00/1 0s}
32. Bf5 {+M17/1 0s} Nf4 {+151.00/1 0s} 33. Kf2 {+M16/1 0s}
Nh3+ {+151.00/1 0s} 34. Kf3 {+M15/1 0s} Ng1+ {+151.00/29
0s} 35. Kg4 {+M14/1 0s (Ke3)} Ne2 {+151.00/1 0s} 36. Bd3
{+M13/1 0s} Nd4 {+151.00/1 0s} 37. Bf6 {+M12/1 0s} Nb3
{+151.00/1 0s} 38. Kf3 {+M8/1 0s} Nc5 {+M7/35 3s} 39. Bf5
{+M7/1 0s} Kh1 {+M6/43 3s} 40. Kf2 {+M6/1 0s} Ne4+ {+M5/96
3s} 41. Bxe4+ {+M5/1 0s} Kh2 {+M4/1 0s} 42. Bd5 {+M4/1 0s
(Bd3)} Kh3 {+M3/1 0s} 43. Be6+ {+M3/1 0s} Kh2 {+M2/100 0s}
44. Be5+ {+M2/1 0s} Kh1 {+M1/100 0s} 45. Bd5# {+M1/1 0s} 1-0

Mate in 45- Either Nalimov is wrong with a Mate in 51 or SYS-Fish is not defining optimal. OR I make something wrong - which I will and can't exclude completly!

Bye
Ingo
IWB
Posts: 1539
Joined: Thu Mar 09, 2006 2:02 pm

Re: SYZYGY Base question

Post by IWB »

syzygy wrote:
IWB wrote:I remove the PGN tags as no one can use this without the EPD in a GUI.
Just click on the b8 square.
[/quote="syzygy"]

Thats all but intuitive, but thx.
syzygy wrote:
Yes, or it might be that in some position any of the "best" moves (in terms of DTM) give up the 50-move draw.

See e.g. here where Nalimov suggests to play 1...Kg6 which loses, whereas 1...Kf6 preserves the draw.
The online Nalimovs (and the ones on my harddrive) want to play c3 with a #100 and they sort Kf6 prior to Kg6 as Kf6 is a mate in 74 while Kg6 is a mate in 73. As this is for Black they would use the longest possible mate but would favor Kf6 and not Kg6! So they would do exactly as you would like to have it (at this particular point!).

Bye
Ingo
Sven
Posts: 4052
Joined: Thu May 15, 2008 9:57 pm
Location: Berlin, Germany
Full name: Sven Schüle

Re: SYZYGY Base question

Post by Sven »

IWB wrote:Accoring to ofline and online Nalimove this is a mate in #51 with a sure capture before the 50 move rule. So the 50 move rule is not important here.
Why are you so sure about that? It might be possible (in theory - not in this case) that the defender can postpone the first capture just to move 51, e.g. Bxg7# with appropriate position of the other pieces. Then it would be a 50 moves draw. I just want to point out that in general you can't derive the information "with a sure capture before the 50 moves rule" only from a DTM tablebase.
IWB wrote:Mate in 45- Either Nalimov is wrong with a Mate in 51 or SYS-Fish is not defining optimal.
As Ronald already explained, an engine using Syzygy Bases will not aim for the shortest mate (from winner's viewpoint) resp. the longest mate (from loser's viewpoint) but for the shortest sequence to the next capture (resp. longest from loser's viewpoint). This is not necessarily optimal from DTM perspective but it will correctly find draws according to 50 moves rule even in positions that are "won" solely based on DTM.

Sven
Sven
Posts: 4052
Joined: Thu May 15, 2008 9:57 pm
Location: Berlin, Germany
Full name: Sven Schüle

Re: SYZYGY Base question

Post by Sven »

IWB wrote:OK, here is Shredder with White and Nalimovs and Stockfish with Black SYS bases:

[Event "1 Minutes/Game + 1 Seconds/Move"]
[Site "Engine Match"]
[Date "2014.01.19"]
[Round "1"]
[White "Stockfish 180114 SYZ 1T"]
[Black "DS12 1T"]
[Result "1/2-1/2"]
[SetUp "1"]
[FEN "B7/8/8/8/5nk1/8/8/4K1B1 w - -"]

...
You probably wanted to enter a different PGN, this one is Shredder with Black again.
syzygy
Posts: 5774
Joined: Tue Feb 28, 2012 11:56 pm

Re: SYZYGY Base question

Post by syzygy »

IWB wrote:
syzygy wrote:See e.g. here where Nalimov suggests to play 1...Kg6 which loses, whereas 1...Kf6 preserves the draw.
The online Nalimovs (and the ones on my harddrive) want to play c3 with a #100 and they sort Kf6 prior to Kg6 as Kf6 is a mate in 74 while Kg6 is a mate in 73. As this is for Black they would use the longest possible mate but would favor Kf6 and not Kg6! So they would do exactly as you would like to have it (at this particular point!).
I don't know what position you are looking at. I am looking at the position after 1.Ne3, where c3 is not an option:
[d]8/8/8/6kp/2p5/2N1NK2/8/8 b - - 1 1
In this position, Nalimov gives:

Code: Select all

Kg5-g6   Lose in 70
Kg5-f6   Lose in 68
Kg5-h6   Lose in 68
h5-h4    Lose in 67
Kg5-h4   Lose in 10
So with Nalimov an engine would play 1...Kg6. However, this leads to a lost position whereas the position after 1...Kf6 is a draw under the 50-move rule.
IWB
Posts: 1539
Joined: Thu Mar 09, 2006 2:02 pm

Re: SYZYGY Base question

Post by IWB »

Sven Schüle wrote:
IWB wrote:Accoring to ofline and online Nalimove this is a mate in #51 with a sure capture before the 50 move rule. So the 50 move rule is not important here.
Why are you so sure about that? It might be possible (in theory - not in this case) that the defender can postpone the first capture just to move 51, e.g. Bxg7# with appropriate position of the other pieces. Then it would be a 50 moves draw. I just want to point out that in general you can't derive the information "with a sure capture before the 50 moves rule" only from a DTM tablebase.
[/quote="Sven Schüle"]

I did not get the information from the Nalimov (how) but because I am playing around with this position now for about 24h. There is a capture before the 50th move!
Sven Schüle wrote:
IWB wrote:Mate in 45- Either Nalimov is wrong with a Mate in 51 or SYS-Fish is not defining optimal.
As Ronald already explained, an engine using Syzygy Bases will not aim for the shortest mate (from winner's viewpoint) resp. the longest mate (from loser's viewpoint) but for the shortest sequence to the next capture (resp. longest from loser's viewpoint). This is not necessarily optimal from DTM perspective but it will correctly find draws according to 50 moves rule even in positions that are "won" solely based on DTM.

Sven
I understood that as well, it is not not relevant here. Shredder with Nalimov managed to mate Stockfish with SYZ in 45 moves. The whole 50 move DTM whatever is irrelevant here. Either Nalimov is right with "Mate in 51", then SYZ.Fish defended suboptimal, or Nalimov is wrong with its #51.

At least I don't see why this should be related to DTM or 50 move rule in this case.

BYe
INgo