tricky position

[jdart]

[d]r7/P2R4/4p3/4k3/1K2p3/6R1/7P/r7 w - - 0 1

I believe Kc3 here draws, other moves lose.

Arasan has trouble with this, even with tablebases. It wants to play Rg5+. Stockfish has no trouble here, maybe because of better eval.

[Joost Buijs]

Nightmare on my 18 core 10980XE wants to play Rg5+ too. At depth 29 it switches to Kc3 and at depth 31 it sees that Kc3 is a draw.

Code: Select all

info depth 24 score cp -47 time 7661 nodes 169829108 nps 22168007 tbhits 200980 pv g3g5 e5f4 g5g7 e4e3 d7f7 f4e4 g7g4 e4d5 g4g5 d5d4 g5g4 d4d5
info depth 25 score cp -42 time 11587 nodes 259047399 nps 22356727 tbhits 297608 pv g3g5 e5f4 g5g7 e4e3 d7f7 f4e4 g7g4 e4d5 g4g5 d5d4 g5g4 d4d3
info depth 26 score cp -52 time 16371 nodes 370577304 nps 22636204 tbhits 397357 pv g3g5 e5f4 g5g7 e4e3 d7f7 f4e4 g7g4 e4d5 g4g5 d5d4 g5g4 d4d3
info depth 27 score cp -71 time 27485 nodes 610633125 nps 22216959 tbhits 729658 pv g3g5 e5f4 g5g7 e4e3 d7f7 f4e4 g7g4 e4d3 f7d7 d3e2 g4g7 e6e5
info depth 28 score cp -74 time 35534 nodes 787138229 nps 22151692 tbhits 978159 pv g3g5 e5f4 g5g7 e4e3 d7f7 f4e4 g7g4 e4d5 g4g5 d5d4 g5g4 d4e5
info depth 29 score cp -43 time 82333 nodes 1805802884 nps 21932917 tbhits 2536337 pv b4c3 a8a7 d7a7 a1a7 c3d2 e5f4 d2e2 a7a2 e2f1 f4f5 g3g2 a2a1
info depth 30 score cp -6 time 107567 nodes 2354225492 nps 21886131 tbhits 3675667 pv b4c3 a8a7 d7a7 a1a7 c3d2 e5f4 g3g8 a7a2 d2e1 e6e5 e1f1 a2d2
info depth 31 score cp 0 time 120120 nodes 2629804885 nps 21893147 tbhits 4083134 pv b4c3 a8a7 d7a7 a1a7 c3d2 e5f4 g3g8 a7a2 d2e1 e6e5 e1f1 a2d2
info depth 32 score cp 0 time 136107 nodes 2971708463 nps 21833619 tbhits 4662495 pv b4c3 a8a7 d7a7 a1a7 c3d2 e5f4 g3g8 a7a2 d2e1 e6e5 e1f1 a2d2
info depth 33 score cp 0 time 158189 nodes 3460693690 nps 21876955 tbhits 5344748 pv b4c3 a8a7 d7a7 a1a7 c3d2 e5f4 g3g8 a7a2 d2e1 e6e5 e1f1 a2d2
info depth 34 score cp 0 time 224825 nodes 4894234338 nps 21769084 tbhits 8016833 pv b4c3 a8a7 d7a7 a1a7 c3d2 a7a8 d2e3 e5d5 h2h4 e6e5 h4h5 a8a3
info depth 35 score cp 0 time 236870 nodes 5152373302 nps 21751903 tbhits 8525774 pv b4c3 a8a7 d7a7 a1a7 c3d2 e5d4 d2e2 a7a8 h2h4 a8c8 h4h5 d4e5

[Nordlandia]

Joost: what does your cpu beast say about this modified position

[d]rnbqkbnr/pppp1ppp/8/8/2B1P3/8/PB3PPP/RN1QK1NR b KQ - 0 5

[Joost Buijs]

Joost: what does your cpu beast say about this modified position

[d]rnbqkbnr/pppp1ppp/8/8/2B1P3/8/PB3PPP/RN1QK1NR b KQ - 0 5

The engine has a problem with this position, it's constantly changing it's mind. It switches between c6, d6, Bb4+ and Nf6.
For a long time it thinks Nf6 is best, but after 10 minutes it switches to d6.

Code: Select all

info depth 22 score cp -59 time 16432 nodes 365137483 nps 22221122 tbhits 0 pv d7d6 g1f3 c8e6 b1d2 b8c6 e1g1 d8d7 a1c1 g8e7 c4e6 d7e6 d2c4
info depth 23 score cp -34 time 23518 nodes 520239783 nps 22120919 tbhits 0 pv g8f6 b1c3 d7d6 g1f3 c8e6 c4e6 f7e6 f3g5 d8d7 d1b3 b8c6 g5e6
info depth 24 score cp -56 time 79214 nodes 1751742314 nps 22114049 tbhits 0 pv f8b4 b1c3 g8f6 g1e2 b8c6 e1g1 c6e5 d1b3 b4f8 c3d5 e5c4 b3c4
info depth 25 score cp -79 time 155155 nodes 3421349120 nps 22051168 tbhits 0 pv b8c6 g1f3 d7d6 b1c3 f8e7 e1g1 c8e6 c3d5 e7f6 d1b3 f6b2 b3b2
info depth 26 score cp -67 time 166888 nodes 3678081086 nps 22039218 tbhits 0 pv g8f6 b1c3 f8b4 g1e2 b8c6 e1g1 c6e5 d1b3 a7a5 c3d5 e5c4 b3c4
info depth 27 score cp -67 time 184312 nodes 4061301129 nps 22034925 tbhits 0 pv g8f6 b1c3 f8b4 g1e2 b8c6 e1g1 c6e5 d1b3 a7a5 c3d5 e5c4 b3c4
info depth 28 score cp -74 time 299898 nodes 6616872457 nps 22063743 tbhits 2 pv g8f6 b1c3 f8b4 g1e2 b8c6 e1g1 c6e5 d1b3 a7a5 c3d5
info depth 29 score cp -73 time 520390 nodes 11490413771 nps 22080389 tbhits 51 pv d7d6 d1b3 d8e7 g1f3 d6d5 c4d5 e7b4 b1d2 b4b3 d5b3 b8c6 e1g1

Funny to see that it already shows TB access, it uses Nalimov 6 pieces. I never felt the need to add syzygy.