About KNNvKP again (DTM50)

Endgame analysis using tablebases, EGTB generation, exchange, sharing, discussions, etc..

About KNNvKP again (DTM50)

Postby pa1ny » Mon Sep 12, 2016 11:36 am

Hello everybody,

what is the maximum legal(!) Mate in X position for KNNvKP ?

All I found out so far, this should be a valid DTM50 mate in 63 moves:
--> N7/8/3k4/5p2/5K2/6N1/8/8 w - - 0 1

Anyone else?

Cheers, pa1ny
pa1ny
 
Posts: 8
Joined: Tue Sep 06, 2016 3:36 pm

Re: About KNNvKP again (DTM50)

Postby pa1ny » Mon Sep 19, 2016 2:50 pm

I've done some research myself lately..

I think this is a legal DTM50 mate in 90(!) moves for KNNvKP:

8/5k2/8/N3K3/p7/N7/8/8 w - - 0 1
8/5k2/8/N3K3/p7/N7/8/8 w - - 0 1

Current DTM bases (Nalimov/Gaviota) say that 1.Nc6 will win in 83 moves, but DTZ50 bases (Syzygy) can destroy this dream and would draw the "optimal" DTM line with the 50-move-rule.

This is the solution:

[Event "Computer Chess EGTB"]
[Date "2016.09.18"]
[White "Nalimov DTM"]
[Black "Syzygy DTZ50"]
[SetUp "1"]
[FEN "8/5k2/8/N3K3/p7/N7/8/8 w - - 0 1"]
[PlyCount "179"]

1. Nc6 Kg6 2. Nd8 Kg5 3. Nf7+ Kg4 4. Ke4 Kg3 5. Ne5 Kf2 6. Nec4 Kg2 7. Kf4
Kf2 8. Nd6 Ke2 9. Ndb5 (9. Ke4 Kf2 10. Nf5 Ke2 11. Nd4+ Kd2 12. Ndc2 Ke2
13. Nb4 Kd2 14. Nb5 a3) 9. .. Kd3 10. Kf3 Kd2 11. Ke4 Ke2 12. Nc3+ Kd2 13.
Kd4 (13. Na2 Ke2 14. Nc2 a3) 13. .. Kc1 14. Ke3 (14. Ncb5 Kd2 15. Nc4+ Ke1
16. Nc3 a3) 14. .. Kb2 15. Ncb5 Kb3 16. Kd3 Kb4 17. Kd4 Kb3 18. Kc5 Kb2 19.
Kb4 Kc1 20. Kc3 Kd1 21. Kd3 Ke1 22. Ke3 Kf1 23. Nd4 Kg2 24. Nf5 Kf1 25. Nh4
Kg1 26. Ke2 Kh2 27. Kf3 Kg1 28. Nf5 Kf1 29. Ng3+ Kg1 30. Ne2+ Kh2 31. Nf4
Kg1 32. Ke2 Kh2 33. Kf2 Kh1 34. Kg3 Kg1 35. Ng2 Kf1 36. Kf3 Kg1 37. Ne3 Kh2
38. Kg4 Kg1 39. Kg3 Kh1 40. Nd1 Kg1 41. Nc3 Kf1 42. Kf3 Kg1 43. Nc4 Kh2 44.
Ne3 a3 (44. .. Kh3 45. Ng2 Kh2 46. Nf4 Kg1 47. Ke2 Kh2 48. Kf2 Kh1 49. Ne4
Kh2 50. Ng3 a3 51. Nf1+ Kh1 52. Nh5 a2 53. Nhg3#) 45. Na2 Kg1 46. Kg3 Kh1
47. Kf2 Kh2 48. Ng2 Kh1 49. Nf4 Kh2 50. Kf3 Kg1 51. Ke2 Kh1 52. Kf1 Kh2 53.
Kf2 Kh1 54. Kg3 Kg1 55. Ne2+ Kf1 56. Kf3 Ke1 57. Ng3 Kd1 58. Ke4 Kd2 59.
Kd4 Kd1 60. Kd3 Ke1 61. Ke3 Kd1 62. Nf1 Ke1 63. Nd2 Kd1 64. Nf3 Kc2 65. Ke2
Kb2 66. Kd2 Kb3 67. Kd3 Ka4 68. Kc4 Ka5 69. Kc5 Ka6 70. Kc6 Ka7 71. Ne5 Kb8
72. Kd7 Ka7 73. Kc7 Ka6 74. Kc6 Ka7 75. Nd7 Ka6 76. Nb8+ Ka7 77. Kc7 Ka8
78. Na6 Ka7 79. Nc5 Ka8 80. Kb6 Kb8 81. Nb7 Kc8 82. Kc6 Kb8 83. Nd6 Ka7 84.
Kb5 Kb8 85. Kb6 Ka8 86. Kc7 Ka7 87. Nb4 Ka8 88. Nc6 a2 89. Nc8 a1=Q 90. Nb6#

Additional Notes:

I used Stockfish-7 for Syzygy DTM50 bases, because Houdini-4 and syzygy-tables.info (online) unfortunately ignore the move-counter! :/
I used an old (free) Rybka-2.2 for Nalimov DTM bases, because newer versions surprisingly cannot handle it so precise and fast.

KNNvKP - DTM50(!) max mate in 90.jpg
KNNvKP - DTM50(!) max mate in 90.jpg (287.09 KiB) Viewed 335 times


In the first 14 moves some of white's DTM moves need to be replaced by the optimal DTZ50 moves to avoid draw by 50-move-rule, that means more than 100 plys between two pawn moves without captures or mate (see red mark and arrows)
Then, at move 44, a pawn move is forced if black don't wanna die on move 53, zeroing forced exactly on move 50 (blue mark).
So from move 45 on, we go strictly DTM again, as we have only 89 plys (< 100) left.
pa1ny
 
Posts: 8
Joined: Tue Sep 06, 2016 3:36 pm

Re: About KNNvKP again (DTM50)

Postby pa1ny » Thu Sep 22, 2016 9:58 am

Ok, this topic is obsolete.

I found the solution from kronsteen and galen about the max mate DTM50 in 112 and many other interesting things for KNNvKP (see http://galen.metapath.org/egtb50/)

6k1/p7/8/8/7N/7K/2N5/8 w - - 0 1
6k1/p7/8/8/7N/7K/2N5/8 w - - 0 1

This thread can be closed.
pa1ny
 
Posts: 8
Joined: Tue Sep 06, 2016 3:36 pm


Return to Endgame Tablebases

Who is online

Users browsing this forum: No registered users and 2 guests