Out of the Kai 450 positions, 99 not found by SF

[jp]

[d]r1bq1r2/pppn2k1/2Pp1b2/3Pp3/1P2PpP1/2N2P1p/P3B3/R2QNR1K b

Nb6 may also win.

Code: Select all

r1bq1r2/pppn2k1/2Pp1b2/3Pp3/1P2PpP1/2N2P1p/P3B3/R2QNR1K b - - acd 37; ce -178; acs 267; acn 2470000; pv Nb6 a4 bxc6 a5 Nd7 dxc6 Nb8 Qa4 Bh4 Bc4 a6 Nd3 Bg3 Bd5 Rh8 Qc2 Qf6 Rfd1 Kf8 Ne2 Qh4 Rf1 Rh7 Bb3 Rh6 Rg1 Ke7 Rad1 Ke8 Bd5 Kf8 Qc4 Rg6 Rdf1 Rg7 Qc3 Rh7 Bb3 Rh6 Rd1 Be6 Bxe6 Rxe6 Qc4 ; bm Nb6;
r1bq1r2/pppn2k1/2Pp1b2/3Pp3/1P2PpP1/2N2P1p/P3B3/R2QNR1K b - - acd 38; ce -171; acs 294; acn 2715000; pv bxc6 dxc6 Nb6 a4 Be6 a5 Nc8 b5 a6 bxa6 Ne7 Bb5 Qb8 Nc2 Qa7 Qd3 Rfb8 Rab1 Ng6 Nb4 Qc5 Nbd5 Bd8 Qc4 Qa7 Qe2 Nh4 Kh2 Bc8 Qf2 Bxa6 Qxa7 Rxa7 Kxh3 Rba8 Bxa6 Rxa6 Nb5 Rxc6 Nbxc7 Bxc7 Kxh4 Bd8+ Kh5 Rxa5 ; bm bxc6;
r1bq1r2/pppn2k1/2Pp1b2/3Pp3/1P2PpP1/2N2P1p/P3B3/R2QNR1K b - - acd 39; ce -158; acs 320; acn 2955000; pv bxc6 dxc6 Nb6 a4 Be6 a5 Nc4 Qb3 d5 Rd1 Ne3 exd5 Bg8 Nd3 Be7 Qb2 Bd6 Ne4 Qe7 Nxd6 cxd6 b5 Bxd5 Nb4 Bf7 b6 axb6 axb6 Rfb8 b7 Nxf1 Rxf1 Rxb7 cxb7 Qxb7 Qd2 Qb6 Kh2 Qc5 Nd3 Qe3 Qxe3 fxe3 Nb4 Rb8 Rb1 Be6 Kxh3 Bg8 Kg3 Bh7 ; bm bxc6;
r1bq1r2/pppn2k1/2Pp1b2/3Pp3/1P2PpP1/2N2P1p/P3B3/R2QNR1K b - - acd 40; ce -174; acs 451; acn 4166000; pv bxc6 dxc6 Nb6 a4 Be6 a5 Nc4 Qb3 d5 Rd1 Ne3 exd5 Bg8 Nd3 Be7 Qb2 Bd6 Ne4 Qe7 Nxd6 cxd6 b5 Bxd5 Nb4 Bf7 b6 axb6 axb6 Rfb8 b7 Nxf1 Rxf1 Rxb7 cxb7 Qxb7 Qd2 Qb6 Kh2 Ra4 Nc2 Qc5 Bd3 Bc4 Bxc4 Rxc4 Ne1 Qd4 Qxd4 Rxd4 Kxh3 Kg6 Nc2 Ra4 Kg2 d5 Rd1 Ra2 Rd2 ; bm bxc6;
r1bq1r2/pppn2k1/2Pp1b2/3Pp3/1P2PpP1/2N2P1p/P3B3/R2QNR1K b - - acd 41; ce -187; acs 555; acn 5135000; pv bxc6 dxc6 Nb6 a4 Be6 a5 Nc4 Qb3 d5 Rd1 Ne3 exd5 Bg8 Nd3 Be7 Qb2 Bd6 Ne4 Qe7 Nxd6 cxd6 b5 Bxd5 Nb4 Bg8 b6 axb6 axb6 Rfb8 b7 Nxf1 Rxf1 Rxb7 cxb7 Qxb7 Qd2 Qb6 Bd3 Rb8 Na6 Ra8 Qe2 Qe3 Qc2 Qb6 Qc7+ Qxc7 Nxc7 Rc8 Nb5 Rc6 Kh2 Bd5 Kxh3 Kf6 Kg2 ; bm bxc6;
r1bq1r2/pppn2k1/2Pp1b2/3Pp3/1P2PpP1/2N2P1p/P3B3/R2QNR1K b - - acd 42; ce -158; acs 784; acn 7256000; pv bxc6 dxc6 Nb6 a4 Be6 a5 Nc4 Qb3 d5 Rd1 Ne3 exd5 Nxd1 Bxd1 Bf7 Nd3 Qd6 Qc4 Rfd8 Nc5 Be7 Qe2 e4 fxe4 Qe5 Qc4 Bxc5 Qxc5 Rh8 Bf3 h2 b5 Bg6 Qb4 Rad8 Ne2 Rh3 Qd4 Qxd4 Nxd4 Rxf3 Rxf3 Bxe4 Ne6+ Kf6 Nxd8 Bxf3+ Kxh2 Bxd5 b6 axb6 axb6 cxb6 c7 ; bm bxc6;
r1bq1r2/pppn2k1/2Pp1b2/3Pp3/1P2PpP1/2N2P1p/P3B3/R2QNR1K b - - acd 43; ce -124; acs 997; acn 9231000; pv bxc6 dxc6 Nb6 a4 Be6 a5 Nc4 Qb3 d5 Rd1 Ne3 exd5 Bg8 Nd3 Be7 Qb2 Bd6 Ne4 Qe7 Nxd6 cxd6 b5 Bxd5 Nb4 Bf7 b6 axb6 axb6 Nxd1 Bxd1 Rfb8 b7 Rxb7 cxb7 Qxb7 Qd2 Qb5 Be2 Qc5 Bd3 Ra4 Rb1 Qe3 Qxe3 fxe3 Be2 Ra8 Kh2 Rb8 Rb2 d5 Nd3 Rxb2 Nxb2 Kf6 Kxh3 d4 Kg2 Kg5 Kf1 Bg6 Nd3 Bxd3 Bxd3 Kf4 Ke2 ; bm bxc6;
r1bq1r2/pppn2k1/2Pp1b2/3Pp3/1P2PpP1/2N2P1p/P3B3/R2QNR1K b - - acd 44; ce -150; acs 1181; acn 10913000; pv bxc6 dxc6 Nb6 a4 Be6 a5 Nc4 Qb3 d5 Rd1 Ne3 exd5 Bg8 Nd3 Be7 Nxe5 Qd6 Nd7 Qxb4 Qxb4 Bxb4 Nxf8 Kxf8 Nb5 Nxd1 Rxd1 Bxa5 Rd4 Kg7 d6 cxd6 Rxd6 Re8 Bd3 a6 Nd4 Rd8 Nf5+ Kf8 Rf6+ Ke8 Bxa6 Bc7 Rh6 Rd1+ Kh2 Bd5 Rh8+ Kf7 Rh7+ Kf6 Nh4 Ra1 Bc8 Bc4 Kxh3 Ba5 Bf5 Bd5 Bd7 Bb6 Rh5 Rh1+ Kg2 Rg1+ Kh2 Rd1 g5+ Ke5 g6+ Kd6 Kh3 Bd8 Kg4 Rg1+ Kxf4 Bxh4 Rxh4 Rxg6 ; bm bxc6;
r1bq1r2/pppn2k1/2Pp1b2/3Pp3/1P2PpP1/2N2P1p/P3B3/R2QNR1K b - - acd 45; ce -180; acs 1445; acn 13348000; pv bxc6 dxc6 Nb6 a4 Be6 a5 Nc4 Qb3 d5 Rd1 Ne3 exd5 Bg8 Nd3 Be7 Nxe5 Qd6 Nd7 Qxb4 Qxb4 Bxb4 Nxf8 Nxf1 Bxf1 Kxf8 Nb5 Bxa5 Rd4 Bb6 Rxf4+ Ke7 Bc4 Rd8 d6+ cxd6 Re4+ Kf8 c7 Ra8 Bxg8 Kxg8 Re6 d5 Kh2 d4 Kxh3 Rc8 g5 d3 Rd6 Bxc7 Nxc7 Rxc7 Rxd3 a5 Kg4 Kg7 f4 a4 f5 ; bm bxc6;

In your output does ce negative mean Black is better or the side not to move (White here) is better?

[Dann Corbit]

It's EPD so it means the side to move is worse so white is better.

[jp]

[d]5rk1/NpR5/3p1q1b/3Pp2p/4Pn2/5P1b/PPQ2B1P/4N2K w - - 9 29

Lc0 (network 320x24.J13-swa-410000) with 100M nodes gives:

info depth 59 seldepth 111 time 4278801 nodes 100000007 score cp -56 hashfull 218 nps 23234 tbhits 2177196 multipv 1 pv a7c8 f8e8 c8b6 g8h8 f2g3 e8g8 c2f2 f4g6 b6c4 h5h4 c7b7 h4g3 h2g3 g6f4 g3f4 h6f4 c4d6 f6d6 f2h4 d6h6 h4h6 f4h6 b7e7 g8g6 e7e5 h8g7 e5h5 h3f1 h5h4 h6d2 e1c2 f1g2 h1h2 d2g5 c2d4 g5h4 d4f5 g7f8 f5h4 g6h6 h2g3 g2f1 h4f5 h6a6 e4e5 f1d3 f5d4 a6a5 d5d6 a5d5 d4e6 f8f7 e6g5 f7e8 g3f4 e8d7 a2a3 d5c5 g5e4 d3e4 f4e4 d7e6 f3f4 c5c4 e4f3 c4a4 b2b4 e6d5 f3g4 a4a7 g4f5 a7f7 f5g4 f7a7 g4f5 a7f7 f5g4 f7f8 a3a4 d5e6 a4a5 f8b8 a5a6 e6d7 a6a7 b8a8 g4f5 a8a7 f5e4 a7a1 e4d5 a1d1 d5c5 d7e6 b4b5 d1c1 c5d4 c1f1 d4c5 f1f4

I don't know of an easy way to make this Lc0 output readable as pgn format, so I did it manually:

[pgn][FEN "5rk1/NpR5/3p1q1b/3Pp2p/4Pn2/5P1b/PPQ2B1P/4N2K w - - 9 29"] 29. Nc8 Re8 30. Nb6 Kh8 31. Bg3 Rg8 32. Qf2 Ng6 33. Nc4 h4 34. Rxb7 hxg3 35. hxg3 Nf4 36. gxf4 Bxf4 37. Nxd6 Qxd6 38. Qh4+ Qh6 39. Qxh6+ Bxh6 40. Re7 Rg6 41. Rxe5 Kg7 42. Rh5 Bf1 43. Rh4 Bd2 44. Nc2 Bg2+ 45. Kh2 Bg5 46. Nd4 Bxh4 47. Nf5+ Kf8 48. Nxh4 Rh6 49. Kg3 Bf1 50. Nf5 Ra6 51. e5 Bd3 52. Nd4 Ra5 53. d6 Rd5 54. Ne6+ Kf7 55. Ng5+ Ke8 56. Kf4 Kd7 57. a3 Rc5 58. Ne4 Bxe4 59. Kxe4 Ke6 60. f4 Rc4+ 61. Kf3 Ra4 62. b4 Kd5 63. Kg4 Ra7 64. Kf5 Rf7+ 65. Kg4 Ra7 66. Kf5 Rf7+ 67. Kg4 Rf8 68. a4 Ke6 69. a5 Rb8 70. a6 Kd7 71. a7 Ra8 72. Kf5 Rxa7 73. Ke4 Ra1 74. Kd5 Rd1+ 75. Kc5 Ke6 76. b5 Rc1+ 77. Kd4 Rf1 78. Kc5 Rxf4 * [/pgn]

After Lc0's 72... Rxa7, the position is

[d]8/r2k4/3P4/4PK2/1P3P2/8/8/8 w - - 0 1

which is a tablebase draw.

Long before that, SF thinks Lc0's PV is 0.00.

But SF thinks Lc0's 68.a4 is a blunder for White that changes the eval from 0.00 to (depth=50)

-2.29 68...Ke6 69.Kf3 Rc8 70.b5 Rc4 71.a5 Rc5 72.b6 Rxa5 73.Ke2 Rb5 74.b7 Rxb7 75.Ke3 Rb5 76.Kf3 Rb3+ 77.Kf2 Rb4 78.Kf3 Rb5 79.Ke3 Kd7 80.Kd3 Rb3+ 81.Ke2 Rb4 82.Ke3 Rb8 83.Kf3 Ra8 84.Kf2 Ke6 85.Kf3 Ra6 86.Kg3 Ra4 87.Kf3 Kf5 88.Ke3 Ra5 89.Kf2 Rd5 90.Ke3 Rd1 91.Ke2 Ra1 92.Kf2 Rb1 93.Ke2 Rb7 94.Kd3

so maybe Lc0's PV is not correct that deep into the line.

[jp]

After Lc0's 72... Rxa7, the position is
8/r2k4/3P4/4PK2/1P3P2/8/8/8 w - - 0 1
which is a tablebase draw.

Afterwards, I saw tbhits 2177196. Obviously not 7-man then.
Was it 5-man TBs?

[zullil]

After Lc0's 72... Rxa7, the position is
8/r2k4/3P4/4PK2/1P3P2/8/8/8 w - - 0 1
which is a tablebase draw.

Afterwards, I saw tbhits 2177196. Obviously not 7-man then.
Was it 5-man TBs?

6-man Syzygy tables.

[jp]

6-man Syzygy tables.

I see. The final position in the PV has 6 men, so maybe that's why the line stopped there. It's a TB draw.
Maybe other lines with the same first move led to 6-man TB losses, which led to the small negative eval Lc0 attaches to the PV.