Some test positions

[peter]

Your twelfth position, Stockfish seems to switch back and fourth on this one before finally settling on 1.Qf6+1, with an even score of 0.00 after longer analysis:



[d]1r3r2/4bpkp/1qb1p1p1/3pP1P1/p1pP1Q2/PpP2N1R/1Pn1B2P/3RB2K w - - 0 1

That's before move 22.Qf6+ of one of Lyudmil Tsvetkov's legendary games against SF on Louis Zulli's machine played over the forum here.

I've stored a .pgn about that too, but I'd rather not paste it here to avoid unnecessary discussions.

[peter]

On your last position Stockfish announces a mate in 35 after a long search:


[d]r5k1/p1pb2bp/3p3r/P1pPp1p1/2B1Pq2/1R2QPp1/1P4PP/5RBK b - - 0 1

And that's A.Summerscale- J.Gawain 0-1 Brit.CC 2009

[MikeGL]

I have inserted the specific line which requires a zugzwang in order to win

Code: Select all

(7. a4 Kb7 8. Ke3 Be7 9. Kd3 Kc7
10. Kc3 Nb8 11. a5 Kb7 12. Kb2 Nd7 13. Bf2 Nc5  

Why sac the Knight here already?
13...Ka6
After e.g. 14.Kb3 Nc5 already works for Black, 15.bxc5 dxc5 and the fortress holds.
If White chooses other than Kb3, Black can simply move the Bishop, where is your ZZ? White can triangle, but so can Black.

I know you have your own favorite engines where you can enable/disable null move, as I remember on your previous tip to me.

Of course there is to be taken care for ZZ too, that's why I took MateFinder here:

http://www.talkchess.com/forum/viewtopi ... 62&t=65140

About tips, switch off tbs here, they just slowdown and with correct play of Black, there won't be reached any 6men, maybe just to avoid 50 moves rule, but that would be only desperate last tries of White after long runs, that's why the evals stay too high here one more time for so long.
BTW Shredder has 0.00 after 25 moves, as you saw

You are correct, looks like a fortress. I will investigate more and post if something comes up.

regards,
Mike

[Dann Corbit]

I have this record for that one:

[d]r5k1/p1pb2bp/3p3r/P1pPp1p1/2B1Pq2/1R2QPp1/1P4PP/5RBK b - - acd 42; acs 2595; bm Rxh2+; c0 "Aaron Summerscale - Gawain Jones, British Championship, 2009"; c3 "Rxh2+"; ce 1403; id "die.093"; pm Rxh2+; pv Rxh2+ Bxh2 Qh4 Qg1 Rf8 f4 exf4 Rd3 Re8 Re1 Bxb2 Re2 Bf6 Re1 Bd4 Rxd4 cxd4 e5 f3 Re4 gxh2 Qf1 Qxe4 e6 fxg2+ Qxg2 Qxg2+ Kxg2 Ba4 Kxh2 Kg7 Kg3 Kf6 a6 h5 Be2 h4+ Kf3 c5 Bf1 Bb3 Ke4 d3 Bxd3 Rb8 Ke3 Bxd5 Bf1 Bxe6 Bd3 Rb3 Ke2 c4 Bh7 Rb2+ Kd1 Ke5 Bc2 g4 Kc1 Ra2;

[peter]

I have the same main line, Dann, but 15 moves long only

[acase]

Hi Andrew!

Your Ninth position wasn't correct so I skipped on to your tenth postion:


[d]8/2p2kb1/np1p2p1/pN1Pp1Pp/2P1P2P/1P6/P4BK1/8 w - - 0 1
...
1.Nxc7
+- (2.23 ++) Depth: 45/80 00:00:28 1011MN
1.Nxc7 Nxc7 2.Bxb6 Na6 3.Bxa5 Nc5 4.Kf3 Bf8 5.Ke3 Be7 6.Be1 Bd8 7.b4 Nd7 8.Kd3 Bb6 9.a4 Bd4 10.a5 Ke8 11.Bd2 Kd8 12.Be3 Bxe3 13.Kxe3 Nb8 14.Kd3 Nd7 15.Kc2 Kc7 16.Kb3 Kb7 17.Ka4 Ka6 18.Ka3 Ka7 19.Kb2 Nb8 20.Kb3 Ka6 21.Kc3 Nd7 22.Kb2 Nf8 23.Kb3 Nd7 24.Ka4 Nf8 25.c5 dxc5 26.bxc5 Kb7 27.c6+ Kc7

How does your SF continue after 6...Ke8 instead of ...Bd8


I did let SF with full Syzygys play out against Shredder13 without tbs at 120'+30", 12 threads and 8G hash each ponder on, adjudicated after 25 moves:

[pgn]
[Event "Lang 120m+30s"]
[Site "?"]
[Date "2017.09.14"]
[Round "1"]
[White "Stockfish 120917 64 POPCNT"]
[Black "Deep Shredder 13 x64"]
[Result "1/2-1/2"]
[Annotator "2.93;2.68"]
[SetUp "1"]
[FEN "8/2p2kb1/np1p2p1/pN1Pp1Pp/2P1P2P/1P6/P4BK1/8 w - - 0 1"]
[PlyCount "50"]
[TimeControl "7200+30"]

{Intel(R) Xeon(R) CPU X5650 @ 2.67GHz 2666 MHz W=77.6 plies; 13.
867kN/s; 548.479.802 TBAs B=63.0 plies; 13.573kN/s} 1. Nxc7 Nxc7 2. Bxb6 {
[%eval 293,63] [%emt 0:03:30]} Na6 {[%eval 268,52] [%emt 0:09:24] (Sa8)} 3.
Bxa5 {[%eval 293,62] [%emt 0:02:19]} Ke8 {[%eval 273,50] [%emt 0:04:07] (Sc5)}
4. Kf3 {[%eval 293,71] [%emt 0:02:11] (b4)} Kd7 {[%eval 239,53] [%emt 0:03:49]}
5. Be1 {[%eval 293,78] [%emt 0:00:00] (b4)} Kc7 {[%eval 249,59] [%emt 0:08:46]
(Lf8)} 6. b4 {[%eval 293,73] [%emt 0:02:53]} Bf8 {[%eval 242,56] [%emt 0:00:47]
(Kb7)} 7. Bd2 {[%eval 293,74] [%emt 0:02:12] (Ke3)} Be7 {[%eval 235,58] [%emt
0:08:42]} 8. a3 {[%eval 293,85] [%emt 0:00:00] (a4)} Nb8 {[%eval 278,59] [%emt
0:08:30] (Kc8)} 9. a4 {[%eval 293,79] [%emt 0:09:07] (Ke2)} Kb7 {[%eval 264,57]
[%emt 0:06:17] (Sa6)} 10. a5 {[%eval 293,72] [%emt 0:03:07] (Ke2)} Ka7 {[%eval
246,54] [%emt 0:03:49] (Ld8)} 11. Kf2 {[%eval 293,65] [%emt 0:02:27] (Ke3)} Ka6
{[%eval 255,55] [%emt 0:03:39] (Sa6)} 12. Kf3 {[%eval 293,79] [%emt 0:02:28]
(Ke3)} Ka7 {[%eval 0,93] [%emt 0:03:36] (Sd7)} 13. Kg3 {[%eval 293,84] [%emt 0:
03:41]} Kb7 {[%eval 255,55] [%emt 0:03:44] (Sd7)} 14. Kg2 {[%eval 293,86]
[%emt 0:02:41] (Le1)} Nd7 {[%eval 224,55] [%emt 0:03:49]} 15. Kg3 {[%eval 293,
85] [%emt 0:00:00] (Kf3)} Nb8 {[%eval 0,95] [%emt 0:02:13] (Ka8)} 16. Kf2 {
[%eval 293,79] [%emt 0:03:27]} Ka6 {[%eval 0,79] [%emt 0:02:12] (Kc8)} 17. Bc3
{[%eval 293,80] [%emt 0:02:48]} Nd7 {[%eval 220,55] [%emt 0:04:46] (Kb7)} 18.
Kg3 {[%eval 293,77] [%emt 0:02:32] (Ke3)} Kb7 {[%eval 245,53] [%emt 0:04:35]
(Ka7)} 19. Kh3 {[%eval 293,77] [%emt 0:06:21] (Kf3)} Nb8 {[%eval 262,52] [%emt
0:02:54] (Ld8)} 20. Kg3 {[%eval 293,82] [%emt 0:03:30] (Kh2)} Nd7 {[%eval 0,76]
[%emt 0:02:19] (Ka7)} 21. Kf2 {[%eval 293,82] [%emt 0:02:49]} Ka6 {[%eval 0,80]
[%emt 0:01:46] (Ka8)} 22. Kf3 {[%eval 293,85] [%emt 0:02:59]} Kb7 {[%eval 268,
53] [%emt 0:03:56] (Ka7)} 23. Kg2 {[%eval 293,78] [%emt 0:06:53] (Ke2)} Nb8 {
[%eval 283,53] [%emt 0:03:43] (Kb8)} 24. Kf3 {[%eval 293,83] [%emt 0:12:37]
(Kf1)} Nd7 {[%eval 0,78] [%emt 0:01:34]} 25. Kg2 {[%eval 293,84] [%emt 0:00:00]
} Nb8 {[%eval 0,83] [%emt 0:01:27] (Kb8) adjud. Remis angenommen} 1/2-1/2
[/pgn]
"Quote" for evals

Hi Peter!



I think you are correct that this looks like a fortress, as the Stockfish evaluation leveled off at +2.85 and never changed for several iterations.

[acase]

On your last position Stockfish announces a mate in 35 after a long search:


[d]r5k1/p1pb2bp/3p3r/P1pPp1p1/2B1Pq2/1R2QPp1/1P4PP/5RBK b - - 0 1

And that's A.Summerscale- J.Gawain 0-1 Brit.CC 2009

Indeed Peter,

I knew that of course as I've seen it several times here on the CCC.