Wrong chess studies

[lantonov]

Here are some published chess studies analysed with Stockfish + Syzygy and proven wrong:

Gia Nadareishvili, Шахматы / Chess, 1962
[d]B7/4k3/2K5/4P1P1/8/p3P3/8/8 w - - 0 1
Solution: 1.g6 a2 2.g7 Kf7 3.e6+ Kxg7 4.e7 a1=Q 5.e8=N+! Kf7 6.Nc7 Qc3+ 7.Kb7! Qxe3 8.Nd5 Qb3+ 9.Ka7 Positional draw.
Wrong!
Instead of 7 ... Qxe3, Black plays 7 ... Qb3.
SF shows -123.45 on move 1 and it is mate in 29:
1. g6 a2 2. g7 Kf7 3. e6+ Kxg7 4. e7 a1=Q 5. e8=N+ Kf7 6. Nc7 Qc3+ 7. Kb7 Qb3+ 8. Kc8 Qxe3 9. Bg2 Qc3 10. Kd7 Qd2+ 11. Bd5+ Kf6 12. Kc8 Ke7 13. Bc6 Qd8+ 14. Kb7 Kd6 15. Ne8+ Kc5 16. Ba4 Qd4 17. Bc6 Qb4+ 18. Kc7 Qb6+ 19. Kd7 Qxc6+ 20. Ke7 Kd5 21. Nf6+ Ke5 22. Nd7+ Kf5 23. Nb8 Qb6 24. Nd7 Qc7 25. Ke8 Ke6 26. Nf8+ Kd6 27. Nd7 Qxd7+ 28. Kf8 Qh7 29. Ke8 Qe7#

Gia Nadareishvili, Альманах "Чадраки", 1974
[d]2k2nq1/6pp/8/4P1PP/1p3N2/p7/5R2/KB6 w - - 0 1
Solution: 1.Ba2 b3 2.e6 bxa2 3.e7 Kd7 4.Re2 Ke8 5.exf8=Q+ Kxf8 6.Ng6+! hxg6 7.hxg6 Qh8 8.Rh2! Qg8 9.Re2! Positional draw.
Wrong!
The author gives the variant 3 ... Nd7 4.Re2 Qe8 5.h6! gxh6 6.Nh5! ostensibly with a White advantage. But:
1. Ba2 b3 2. e6 bxa2 3. e7 Nd7 4. Re2 Qe8 5. h6 Nc5 6. Kxa2 Qf7+ 7. Ka1 Nb3+ 8. Kb1 a2+ 9. Rxa2 Qxe7 10. Rc2+ Kb8 11. hxg7 Qxg7 12. Rb2 Qf7 13. Nd3 Kc7 14. Rc2+ Kd8 15. Rc3 Nd4 16. Nb2 Qf5+ 17. Rd3 Ke7 18. Ka2 Qe4 19. Ra3 Qd5+ 20. Kb1 Qxg5 21. Ra7+ Ke6 22. Ra6+ Kd7 23. Ra7+ Kc6 24. Ra3 Qf5+ 25. Nd3 Qf1+ 26. Ka2 Nc2 27. Rc3+ Kd5 28. Kb2 Kd4 29. Nc1 Qxc1+ 30. Kxc1 Kxc3 -123.49

Gia Nadareishvili, "Italia scacchistica" competition, 1975
[d]8/8/8/2B4B/K7/8/p7/1k1n4 w - - 0 1
Solution: 1. Ba3 Nb2+ 2. Kb3 a1=Q 3. Bg6+ Kc1 4. Bc2 Kd2 5. Bb4+ Kc1 6. Ba3 Kd2 7. Bb4+ Ke2 8. Bc3 Qa6 9. Kxb2 Positional draw
It is true that the final position is a draw but instead of 1 ... Nb2+ Black plays 1 ... Kc2 and mates in 5 e.g. in the following variant:
1. Ba3 Kc2 2. Bg6+ Kd2 3. Bb4+ Nc3+ 4. Ka3 a1=Q+ 5. Kb3 Qa2#

Gia Nadareishvili, "64", 1974
[d]8/1K6/2P5/k7/P4b2/7b/8/5N2 w - - 0 1
Solution: 1. Nh2 Bxh2 2. c7 Bg2+ 3. Ka7 Bg1+ 4. Kb8 Bh2 5. Ka7 Bh3 6. Kb7 Bf1 7. Kc6 Bh3 8. Kb7 Bg2+ 9. Ka7 draw
Instead of 1 ... Bxh2 SF plays 1 ... Bf5 with mate (-123.40)
Variant: 1. Nh2 Bf5 2. Nf3 Bd3 3. Kc8 Be4 4. Nd4 Be5 5. Nb3+ Kb6 6. c7 Bf5+ 7. Kd8 Bf6+ 8. Ke8 Kxc7 9. Kf7 Bc3 10. Nc1 Kd6 11. Ne2 Be5 12. Kg8 Bg4 13. Nc1 Be6+ 14. Kf8 Bc4 15. a5 Bf4 16. Kg7 Bxc1 17. Kg6 Ke5 18. Kh5 Bd2 19. Kg6 Bxa5 20. Kg5 Be6 21. Kh4 Be1+ 22. Kg5 Bf5 23. Kh6 Kf6 24. Kh5 Bf2 25. Kh6 Bg4 26. Kh7 Kf7 27. Kh8 Be3 28. Kh7 Bf5+ 29. Kh8 Bd4#

T. Laza, 1843
[d]1k6/nn6/1K6/8/Q7/8/8/8 b - - 0 1
Solution: 1 ... Nc8+ 2. Kc6 Ne7+ 3. Kb5 Nd6+ 4. Kb6 Kc8 5. Qa7 Kd8 6. Qc7+ Ke8 7. Qb8+ Kd7 8. Kc5 Nf7 draw
However, 1 ... Nc8+ 2. Kb5 Na7+ 3. Kb4 Nc8 4. Qc6 Ncd6 5. Kc3 Nf7 6. Qd7 Nfd6 7. Kd4 Nc8 8. Qc6 Nd8 9. Qc2 Nb7 10. Qh2+ Ka7 11. Kd5 Ne7+ 12. Ke6 Nc6 13. Kd7 Nb4 14. Kc7 Na6+ 15. Kc6 Nd8+ 16. Kd7 Nb7 17. Qf2+ Ka8 18. Qd4 Nac5+ 19. Kc6 Kb8 20. Qg4 Ka7 21. Qb4 Ka8 22. Kb6 Nd7+ 23. Kc7 Ndc5 24. Qa3+ Na5 25. Qxa5+ Na6+ 26. Qxa6#

A. Troitsky, Шахматный журнал / Chess Journal, 1900
[d]3N4/8/6k1/2K5/8/B4p2/8/8 w - - 0 1
Solution: 1. Nc6 f2 2. Ne5+ Kg7 3. Kd5 f1=Q 4. Be7 draw
But SF says that this position is won for Black. Let's see:
4 ... Qg2+ 5. Ke6 Kh6 6. Nf7+ Kg6 7. Ne5+ Kh5 8. Kf6 Qh3 9. Bc5 Qf1+ 10. Ke6 Qg2 11. Be7 Qg8+ 12. Kf6 Qh8+ 13. Ke6 Qc8+ 14. Kd5 Qb7+ 15. Ke6 Qb3+ 16. Kf6 Qb6+ 17. Kf5 Qa6 18. Bb4 Qf1+ 19. Ke6 Qh3+ 20. Kd5 Kg5 21. Nc4 Qd3+ 22. Kc5 Kf5 23. Ba3 Ke4 24. Nd6+ Kf4 25. Bc1+ Kf3 26. Bb2 Qc2+ 27. Nc4 Ke4 28. Kb4 Kd3 29. Ne5+ Ke3 30. Nc4+ Ke4 31. Bc3 Qf2 32. Nb2 Qb6+ 33. Kc4 Qb7 34. Bb4 Qd5+ 35. Kc3 Qb5 36. Ba3 Qe5+ 37. Kc2 Qc7+ 38. Kb3 Qg3+ 39. Ka2 Qg8+ 40. Kb1 Qb3 41. Be7 Ke3 42. Bh4 Qd5 43. Kc2 Qe4+ 44. Kc3 Qd4+ 45. Kc2 Qxh4 46. Kc3 Qf6+ 47. Kc2 Qc6+ 48. Kb3 Kd4 49. Nd1 Qf3+ 50. Kb4 Qxd1 51. Kb5 Kd5 52. Kb6 Kd6 53. Kb7 Qb3+ 54. Ka6 Kc6 55. Ka5 Qa3#

Yu. Mendheim, 1832
[d]8/3K1kn1/2Q2b2/8/8/8/8/8 w - - 0 1
Solution: 1. Qd5+ Kg6 2. Qg2+ Kf7 3. Qg4 Be5 4. Qc4+ Kg6 5. Qe4+ Kf6 6. Kc6 Nf5 7. Qf3 Ke6 8. Kc5 Bd6+ 9. Kc4 Ne7 draw
SF (Syzygy) finds a White win: 10. Qb7 Bh2 11. Qh1 Be5 12. Qd1 Nf5 13. Qd5+ Kf6 14. Qc6+ Bd6 15. Kd3 Ke6 16. Qe4+ Kf6 17. Qe8 Be7 18. Qc6+ Ke5 19. Qc3+ Kf4 20. Qc1+ Ke5 21. Qe1+ Kf4 22. Qe4+ Kg5 23. Qf3 Bd6 24. Ke4 Ng3+ 25. Kd5 Bc7 26. Ke6 Bf4 27. Qd5+ Kg4 28. Qd1+ Kg5 29. Qf3 Nh5 30. Qd5+ Kg4 31. Qf5+ Kh4 32. Qf7 Kg4 33. Qg6+ Bg5 34. Qf5+ Kh4 35. Qf3 Ng3 36. Kf7 Kh3 37. Kg6 Bc1 38. Qf6 Bd2 39. Qf7 Bc3 40. Qe6+ Kh2 41. Kg5 Bd4 42. Qd5 Bf2 43. Kg4 Nf1 44. Kf3 Kg1 45. Qg8+ Ng3 46. Qg6 Kf1 47. Qb1+ Be1 48. Qc2 Ne2 49. Qxe2+ Kg1 50. Qg2#

G. Rink, Schakkverlden, 1939
[d]6q1/8/5K2/8/8/4B3/8/k2B4 w - - 0 1
Solution: 1. Be2 Qd5 2. Bf4 Qe4 3. Be5+ Ka2 4. Bh5 Qh4+ 5. Kg6 Qe4+ 6. Kf6 draw
But not according to SF: 6 ... Qh4+ 7. Kg6 Kb3 8. Bf6 Qh2 9. Kg5 Qd2+ 10. Kf5 Qd7+ 11. Kg5 Kc2 12. Bg6+ Kd2 13. Be5 Qe6 14. Bf4+ Kc3 15. Bf5 Qg8+ 16. Kh5 Qg7 17. Bg5 Qf7+ 18. Kg4 Kd4 19. Bf4 Qf6 20. Bg3 Ke3 21. Bf4+ Ke2 22. Bg5 Qc6 23. Kf4 Qc1+ 24. Kg4 Qc4+ 25. Bf4 Qg8+ 26. Bg5 Kf2 27. Be4 Qf7 28. Bf5 Qd5 29. Bh4+ Ke2 30. Bf6 Qf3+ 31. Kg5 Qg3+ 32. Kh6 Qg8 33. Bg6 Ke3 34. Kg5 Qd5+ 35. Kh6 Qg8 36. Kg5 Kf3 37. Kf5 Qc8+ 38. Ke5 Kg4 39. Be7 Qc3+ 40. Ke6 Kf3 41. Bf6 Qb3+ 42. Ke7 Qg8 43. Bf7 Qh7 44. Kf8 Ke3 45. Be6 Qg6 46. Ke7 Ke4 47. Bd7 Kd5 48. Be6+ Kc5 49. Be5 Kc6 50. Bf6 Qd3 50-move rule
[d]8/4K3/2k1BB2/8/8/3q4/8/8 w - - 100 51
Position after move 50: 1. Bg7 Qg6 2. Be5 Qg5+ 3. Bf6 Qf4 4. Bg7 Kb6 5. Ba1 Kb5 6. Bf6 Qc7+ 7. Kf8 Qh7 8. Bf7 Kb6 9. Ke8 Kc7 10. Ke7 Kc6 11. Ke6 Qh3+ 12. Ke5 Kd7 13. Bg5 Qc3+ 14. Kf5 Qc5+ 15. Kg6 Qc2+ 16. Kg7 Qg2 17. Kf6 Qf3+ 18. Kg7 Qg4 19. Kf6 Kd6 20. Bh6 Qf3+ 21. Kg6 Qe4+ 22. Kf6 Qe7+ 23. Kg6 Ke5 24. Bg7+ Kf4 25. Bf6 Qe4+ 26. Kh6 Qh1+ 27. Kg7 Qb7 28. Kg6 Qg2+ 29. Kh7 Kf5 30. Bh8 Qg5 31. Ba1 Qh4+ 32. Kg7 Qg3+ 33. Kh7 Qh2+ 34. Kg8 Qg2+ 35. Bg7 Qa8+ 36. Kh7 Kg5 37. Bh8 Qf8 38. Bc4 Qe7+ 39. Kg8 Qd8+ 40. Kg7 Qd4+ 41. Kf7 Qxh8 42. Ke7 Qf6+ 43. Kd7 Qd4+ 44. Kc6 Qxc4+ 45. Kb6 Kf6 46. Ka5 Ke5 47. Kb6 Kd6 48. Ka5 Qb3 49. Ka6 Kc6 50. Ka5 Qa3#

G. Rink, Magyar sakkvilag, 1939
[d]8/1k6/3K4/8/3B4/8/8/3B1q2 w - - 0 1
Solution: 1. Bb3 Qf8+ 2. Ke5 Qh8+ 3. Ke4 Qh7+ 4. Ke5 Qg7+ 5. Ke4 Qg6+ 6. Kf4 Qd6+ 7. Ke4 draw
But SF thinks otherwise: 7 ... Qg6+ 8. Kf4 Kc6 9. Bc4 Qc2 10. Be6 Qh2+ 11. Ke4 Qh1+ 12. Kf4 Qh4+ 13. Bg4 Kd5 14. Bg7 Kc4 15. Be5 Kd3 16. Bg7 Qf2+ 17. Kg5 Qg1 18. Kf4 Qe3+ 19. Kf5 Qg3 20. Bf6 Qd6 21. Bh5 Qd5+ 22. Be5 Qe4+ 23. Ke6 Qc6+ 24. Kf5 Qc8+ 25. Kf4 Qf8+ 26. Kg5 Ke4 27. Bh2 Qf5+ 28. Kh4 Qf1 29. Kg3 Qf4+ 30. Kh3 Qe3+ 31. Kh4 Kf5 32. Bg4+ Kf6 33. Bc7 Qe1+ 34. Kh3 Qc3+ 35. Bg3 Kg5 36. Be2 Qe3 37. Bc4 Qe4 38. Bg8 Qg6 39. Bc4 Qh7+ 40. Kg2 Qe4+ 41. Kf2 Qxc4 42. Ke3 Qc3+ 43. Kf2 Kg4 44. Bh2 Qd2+ 45. Kg1 Kf3 46. Bg3 Qg2#

G. Rink, Amaneker, 1947
[d]3N4/3q4/6N1/3k3K/8/8/8/8 w - - 0 1
1. Ne6 Qh7+ 2. Kg5 Ke4 3. Nef4 Qf7 4. Nh5 Qf5+ 5. Kh6 Qf2 6. Kg5 Qg2+ 7. Kf6 draw
Black win (SF): 7 ... Qb2+ 8. Kg5 Qd2+ 9. Ngf4 Qa5+ 10. Kg6 Qf5+ 11. Kh6 Kf3 12. Ng6 Kg4 13. Nhf4 Qg5+ 14. Kh7 Kf5 15. Kg7 Qf6+ 16. Kh7 Kg5 17. Nh3+ Kg4 18. Nhf4 Kf5 19. Kh6 Qf7 20. Nh8 Qf8+ 21. Kh7 Qe7+ 22. Kg8 Qd7 23. Nh5 Qe8+ 24. Kh7 Qxh5+ 25. Kg7 Qe8 26. Ng6 Qxg6+ 27. Kh8 Qg5 28. Kh7 Kf6 29. Kh8 Qg7#

S. Kaminer, Шахматы / Chess, 1926
[d]r3k3/p5pp/5p2/1K2n3/P4PN1/8/1R4P1/8 w q - 0 1
1. Re2 Rb8+ 2. Ka5 Rb2 3. Re4 Rxg2 4. Ne3 Re2 5. fxe5 f5 6. Nxf5 Rxe4 7. Nd6+ White win
Instead of 5 ... f5 SF prefers 5 ... fxe5 and the game is probable draw:
5 ... fxe5 6. Rxe5+ Kf7 7. Ka6 h5 8. Kxa7 h4 9. Nc4 Rc2 10. Nd6+ Kf6 11. Rh5 g5 12. a5 Kg6 13. Rh8 Kg7 14. Rb8 h3 15. Nf5+ Kg6 16. Ng3 Rg2 17. Rb6+ Kf7 18. Nf5 h2 19. Rh6 Rf2 20. Ng3 Kg7 21. Rh3 Kg6 22. Nh1 Rf7+ 23. Kb6 Rf6+ 24. Kc5 Rf5+ 25. Kb4 Rf4+ 26. Kb3 Rf5 27. Rxh2 Rxa5 28. Ng3 g4 29. Kb4 Re5 30. Kc4 Re3 31. Rg2 Kg5 32. Rg1 Kf4 33. Nf1 Re1 34. Kc3 g3 35. Nxg3 Re3+ 36. Kd4 Rxg3 37. Rxg3 Kxg3 =

A.S. Gurvich, Бакинский рабочий / Baku Worker, 1927
[d]4N3/bB3N2/4k3/3p1n2/8/8/3KP3/8 w - - 0 1
Solution: 1. Nh8 Kd7 2. Nf6+ Kc7 3. Ba6 Bd4 4. Nxd5+ Kc6 5. Ng6 Kxd5 6. e4+ Kxe4 7. Bb7#
SF variation: 1. Nh8 Nh6 2. Bc6 Nf5 3. Bb5 Bb8 4. Bd3 Ng3 5. e3 Ke7 6. Bg6 Ba7 7. Nc7 Nf1+ 8. Kc3 Nxe3 9. Bb1 Bb8 10. Ng6+ Kf6 11. Na6 Be5+ 12. Kd2 Nc4+ 13. Ke2 Bd6 14. Nh4 Nb6 15. Ke3 Nd7 16. Kd4 Nb8 17. Nxb8 Bxb8 18. Kc5 Ba7+ 19. Kd6 Bb8+ 20. Kd7 Ke5 21. Nf3+ Kf4 22. Ne1 Ke3 23. Bh7 Bg3 24. Nd3 Kd4 25. Bf5 Kc4 26. Kd8 Bd6 27. Bg6 Kc3 28. Kd7 Bh2 29. Kc6 Bg3 30. Kb5 Bh4 31. Bf5 Bg3 32. Kc6 Kd4 33. Bh7 Bh2, etc. with draw

About 1 in 20 studies analysed are wrong.

[Arpad Rusz]

About 1 in 20 studies analysed are wrong.

Harold van der Heijden's endgame study database (version IV) contains 76132 studies. 18203 of them are already marked as "cooked" (with some sort of flaws). Since nobody ever checked all the studies with computer, the situation is probably even worse... Sometimes the positions are very complex and even the engines cannot help...

[Dirt]

Here are some published chess studies analysed with Stockfish + Syzygy and proven wrong:

Gia Nadareishvili, Шахматы / Chess, 1962
[d]B7/4k3/2K5/4P1P1/8/p3P3/8/8 w - - 0 1
Solution: 1.g6 a2 2.g7 Kf7 3.e6+ Kxg7 4.e7 a1=Q 5.e8=N+! Kf7 6.Nc7 Qc3+ 7.Kb7! Qxe3 8.Nd5 Qb3+ 9.Ka7 Positional draw.
Wrong!
Instead of 7 ... Qxe3, Black plays 7 ... Qb3.
SF shows -123.45 on move 1 and it is mate in 29:
1. g6 a2 2. g7 Kf7 3. e6+ Kxg7 4. e7 a1=Q 5. e8=N+ Kf7 6. Nc7 Qc3+ 7. Kb7 Qb3+ 8. Kc8 Qxe3 9. Bg2 Qc3 10. Kd7 Qd2+ 11. Bd5+ Kf6 12. Kc8 Ke7 13. Bc6 Qd8+ 14. Kb7 Kd6 15. Ne8+ Kc5 16. Ba4 Qd4 17. Bc6 Qb4+ 18. Kc7 Qb6+ 19. Kd7 Qxc6+ 20. Ke7 Kd5 21. Nf6+ Ke5 22. Nd7+ Kf5 23. Nb8 Qb6 24. Nd7 Qc7 25. Ke8 Ke6 26. Nf8+ Kd6 27. Nd7 Qxd7+ 28. Kf8 Qh7 29. Ke8 Qe7#

A good test of whether my Syzygy bases are installed right. Thank you!

[noctiferus]

Actually, I find, using Houdini4 or Stockfish151214, both with Syzigies, that white can delay mate by one move (mate in 30) playing 24...Kd7 instead of 24...Nd7:

24. Kd7 Qxb8 25. Kc6 Qb4 26. Kc7 Qb5 27. Kd6 Kf6 28. Kc7
Ke6 29. Kc8 Kd6 30. Kd8 Qd7#

instead of
24.Nd7 Qc7 25. Ke8 Ke6 26. Nf8+ Kd6 27. Nd7 Qxd7+ 28. Kf8 Qh7 29. Ke8 Qe7#
or
24. Na6 Qxa6 25. Kd7 Qb5+ 26. Ke7 Qc6 27. Kd8 Qb7 28. Ke8 Ke6 29. Kf8 Qf7#

Do you agree?

[noctiferus]

"mate by one move (mate in 30) playing 24...Kd7 instead of 24...Nd7:"

Stupid error by me: of course I meant "24. Kd7" and "24. Nd7".
Sorry

[lantonov]

About 1 in 20 studies analysed are wrong.

Harold van der Heijden's endgame study database (version IV) contains 76132 studies. 18203 of them are already marked as "cooked" (with some sort of flaws). Since nobody ever checked all the studies with computer, the situation is probably even worse... Sometimes the positions are very complex and even the engines cannot help...

Didn't know about Harold van der Heijden's database. Were the cooked studies checked with tablebases ?

[dadij]

Harold van der Heijden's endgame study database (version IV) contains 76132 studies. 18203 of them are already marked as "cooked" (with some sort of flaws). Since nobody ever checked all the studies with computer, the situation is probably even worse... Sometimes the positions are very complex and even the engines cannot help...

Here are a few additional "cooked" positions which I tested a couple of years ago:

A few studies tested with Aquarium

More studies tested with Aquarium Lomonosov tables

[Dirt]

Actually, I find, using Houdini4 or Stockfish151214, both with Syzigies, that white can delay mate by one move (mate in 30) playing 24...Kd7 instead of 24...Nd7

Actually, I'm getting -123.49 for the score instead of -123.45. Syzygy should work by white trying to delay conversion, which makes distance to mate not so relevant.

[noctiferus]

Try using a backward analysis on the proposed line: it shows the above correction.

[petero2]

Gia Nadareishvili, Шахматы / Chess, 1962
[d]B7/4k3/2K5/4P1P1/8/p3P3/8/8 w - - 0 1
Solution: 1.g6 a2 2.g7 Kf7 3.e6+ Kxg7 4.e7 a1=Q 5.e8=N+! Kf7 6.Nc7 Qc3+ 7.Kb7! Qxe3 8.Nd5 Qb3+ 9.Ka7 Positional draw.
Wrong!
Instead of 7 ... Qxe3, Black plays 7 ... Qb3.
SF shows -123.45 on move 1 and it is mate in 29:
1. g6 a2 2. g7 Kf7 3. e6+ Kxg7 4. e7 a1=Q 5. e8=N+ Kf7 6. Nc7 Qc3+ 7. Kb7 Qb3+ 8. Kc8 Qxe3 9. Bg2 Qc3 10. Kd7 Qd2+ 11. Bd5+ Kf6 12. Kc8 Ke7 13. Bc6 Qd8+ 14. Kb7 Kd6 15. Ne8+ Kc5 16. Ba4 Qd4 17. Bc6 Qb4+ 18. Kc7 Qb6+ 19. Kd7 Qxc6+ 20. Ke7 Kd5 21. Nf6+ Ke5 22. Nd7+ Kf5 23. Nb8 Qb6 24. Nd7 Qc7 25. Ke8 Ke6 26. Nf8+ Kd6 27. Nd7 Qxd7+ 28. Kf8 Qh7 29. Ke8 Qe7#

Actually this is mate in 44 according to lomonosov tablebases, and texel 1.05 using both syzygy and gaviota tablebases agree:

Code: Select all

uci
setoption name GaviotaTbPath value /home/petero/chess/gtb
setoption name GaviotaTbCache value 1024
setoption name MinProbeDepth value 1
setoption name SyzygyPath value /home/petero/chess/rtb/wdl:/home/petero/chess/rtb/dtz:/home/petero/chess/rtb/6wdl:/home/petero/chess/rtb/6dtz
setoption name threads value 16
setoption name hash value 8192
position fen B7/4k3/2K5/4P1P1/8/p3P3/8/8 w - - 0 1
go infinite
...
info depth 25 score mate -44 time 42491 nodes 547673074 nps 12889154 tbhits 29175169 pv g5g6 a3a2 g6g7 e7f7 e5e6 f7g7 e6e7 a2a1q e7e8n g7g6 e8c7 a1c3 c6b7 c3b3 b7c6 b3e3 c7d5 e3e5 a8b7 e5b8 b7a6 g6f5 a6b7 f5e5 d5c7 b8a7 c7b5 a7g1 c6c7 g1c5 b7c6 e5e6 c7b7 e6f6 b7c7 f6e7 c7b7 e7d8 b5a7 c5e7 b7b6 e7c7 b6a6 d8e7 c6b7 c7d6 a7c6 e7e6 a6b5 d6d5 b5b4 e6d6 c6a5 d5d2 b4a4 d2d4 a4b5 d4b2 b5c4 b2b6 a5b3 b6b7 b3d2 b7b2 c4d3 d6d5 d3e3 b2c3 e3e2 d5e5 d2f3 e5f4 f3e1 f4e4 e2f1 c3b2 f1g1 b2c1 g1f1 e4e3 f1g1 c1e1 g1g2 e1h4 g2g1 e3f3 g1f1 h4h1

[pgn][Event "Edited position"]
[Site "localhost"]
[Date "2015.02.22"]
[Round "-"]
[White "-"]
[Black "-"]
[Result "0-1"]
[FEN "B7/4k3/2K5/4P1P1/8/p3P3/8/8 w - - 0 1"]
[SetUp "1"]

1. g6 a2 2. g7 Kf7 3. e6+ Kxg7 4. e7 a1=Q 5. e8=N+ Kg6 6. Nc7 Qc3+ 7. Kb7
Qb3+ 8. Kc6 Qxe3 9. Nd5 Qe5 10. Bb7 Qb8 11. Ba6 Kf5 12. Bb7 Ke5 13. Nc7 Qa7
14. Nb5 Qg1 15. Kc7 Qc5+ 16. Bc6 Ke6 17. Kb7 Kf6 18. Kc7 Ke7 19. Kb7 Kd8
20. Na7 Qe7+ 21. Kb6 Qc7+ 22. Ka6 Ke7 23. Bb7 Qd6+ 24. Nc6+ Ke6 25. Kb5
Qd5+ 26. Kb4 Kd6 27. Na5 Qd2+ 28. Ka4 Qd4+ 29. Kb5 Qb2+ 30. Kc4 Qb6 31. Nb3
Qxb7 32. Nd2 Qb2 33. Kd3 Kd5 34. Ke3 Qc3+ 35. Ke2 Ke5 36. Nf3+ Kf4 37. Ne1
Ke4 38. Kf1 Qb2 39. Kg1 Qc1 40. Kf1 Ke3 41. Kg1 Qxe1+ 42. Kg2 Qh4 43. Kg1
Kf3 44. Kf1 Qh1#
{Black mates} 0-1
[/pgn]