Page 2 of 2

Re: 4 old zugzwangs where modern engines still have troubles

Posted: Wed Jul 13, 2016 1:37 pm
by mar
Very nice, cheng solves all but last position (struggles a bit with the first).

Re: 4 old zugzwangs where modern engines still have troubles

Posted: Thu Jul 14, 2016 6:29 pm
by Vinvin
One more. SF doesn't find a winning score :
[pgn]
[Event "-"]
[Site "-"]
[White "w"]
[Black "b"]
[Result "1-0"]
[SetUp "1"]
[FEN "rk1b4/p2p2p1/1P6/2R2P2/8/2K5/8/5B2 w - - 0 1"]

1. Rc8+ Kxc8 2. Ba6+ Kb8 3. b7 Bf6+ 4. Kc4 Be5 5. Kc5 Bh2 6. Kd4 Bf4 7. Kd5
Bh2 8. f6 gxf6 9. Ke4 Bg3 10. Kf5 Be5 11. Kg6 Bb2 12. Kf7 f5 13. Ke8 f4 14.
Kxd7 Bc3 15. bxa8=Q+ Kxa8 16. Kc8 f3 17. Bb7# 1-0
[/pgn]

Sting-SF finds a winning score in 6 seconds :

Code: Select all

rk1b4/p2p2p1/1P6/2R2P2/8/2K5/8/5B2 w - - 0 1
Sting-sf-67:
...
 25	00:02	 13 294k	4 479k	 0,00	2.Ba6+ Kb8 3.b7 Bf6+ 4.Kc4 Be5 5.Kd5 Bg3 6.Kc5 Bh2 7.Kd5 Bc7 8.Kd4 Bg3 9.Kc5
 26	00:04	 31 645k	7 777k	 0,00	2.Ba6+ Kb8 3.b7 Bf6+ 4.Kc4 Be5 5.Kd5 Bg3 6.Kc5 Bh2 7.Kd5 Bc7 8.Kd4 Bg3 9.Kc5
 27+	00:04	 44 318k	9 046k	+0,08	2.Ba6+
 27+	00:04	 45 340k	9 104k	+0,16	2.Ba6+
 27+	00:05	 49 478k	9 203k	+1,75	2.Ba6+
 27	00:05	 52 209k	9 211k	+2,02	2.Ba6+ Kb8 3.b7 Bf6+ 4.Kb4 Bd4 5.Kc4 Be3 6.Kd5 Bf4 7.Ke4 Bd6 8.f6 g6 9.f7 Bf8 10.Kf4 Be7 11.Ke5 g5 12.Kf5 Bf8 13.Kxg5 d5 14.Kf5 d4 15.Ke6 d3 16.Kd7 d2 17.bxa8Q+ Kxa8 18.Kc8
 28+	00:06	 56 759k	9 351k	+2,14	2.Ba6+
 28+	00:06	 57 288k	9 367k	+2,26	2.Ba6+
 28+	00:06	 64 660k	9 517k	+8,07	2.Ba6+
 28	00:08	 79 462k	9 775k	+9,37	2.Ba6+ Kb8 3.b7 Bf6+ 4.Kc4 Be5 5.Kc5 Bg3 6.Kd5 Bh2 7.f6 gxf6 8.Ke4 Be5 9.Kf5 Bd4 10.Kg6 f5 11.Kf7 f4 12.Ke7 Kc7 13.bxa8Q Bc5+ 14.Ke8 Kd6 15.Qb8+ Kd5 16.Qxf4 Kc6 17.Qf7 Bb6 18.Qxd7+ Kc5 19.Ke7 Kb4 20.Ke6 Be3 21.Ke5
 29+	00:08	 80 798k	9 769k	+9,49	2.Ba6+
 29+	00:08	 81 881k	9 768k	+9,61	2.Ba6+
 29+	00:09	 90 323k	9 818k	+11,09	2.Ba6+
 29	00:09	 95 750k	9 826k	+11,55	2.Ba6+ Kb8 3.b7 Bf6+ 4.Kc4 Be5 5.Kc5 Bg3 6.Kd5 Bh2 7.f6 gxf6 8.Ke4 Bg3 9.Kf5 Bh4 10.Kg6 Bg5 11.Kf7 Kc7 12.bxa8Q Kb6 13.Qb7+ Kc5 14.Qxa7+ Kd6 15.Qb7 Kc5 16.Qxd7 Kb4 17.Qd3 Kc5 18.Qc4+ Kb6 19.Qb5+ Kc7 20.Qc5+ Kd7 21.Qd4+ Kc7 22.Ke6 Kb8 23.Qb6+ Ka8 24.Qb7+
 30+	00:10	 101 890k	9 857k	+11,67	2.Ba6+
 30+	00:10	 104 080k	9 870k	+11,79	2.Ba6+
 30+	00:11	 119 018k	9 988k	+13,27	2.Ba6+
 30	00:12	 120 856k	10 000k	+12,80	2.Ba6+ Kb8 3.b7 Bf6+ 4.Kc4 Be5 5.Kc5 Bg3 6.Kd5 Bf4 7.Ke4 Bc7 8.f6 gxf6 9.Kf5 Be5 10.Kg6 d5 11.Kf5 Kc7 12.bxa8Q Kb6 13.Qb7+ Kc5 14.Qb5+ Kd4 15.Qb4+ Ke3 16.Qe1+ Kd4 17.Qc1 Bd6 18.Qd2+ Kc5 19.Qc3+ Kb6 20.Qd4+ Kxa6 21.Qxf6 Kb5 22.Qxd6 d4 23.Ke4 d3 24.Kxd3 Ka5 25.Qc5+ Ka6
 31+	00:12	 123 571k	9 983k	+12,92	2.Ba6+
 31+	00:12	 124 394k	9 975k	+13,05	2.Ba6+
 31	00:12	 125 405k	9 965k	+12,88	2.Ba6+ Kb8 3.b7 Bf6+ 4.Kc4 Be5 5.Kc5 Bg3 6.Kd5 Bc7 7.f6 gxf6 8.Ke4 Bd6 9.Kf5 Be5 10.Kg6 d5 11.Kf5 Kc7 12.bxa8Q Kb6 13.Qb7+ Kc5 14.Qb5+ Kd4 15.Qb4+ Ke3 16.Qe1+ Kd4 17.Qc1 Bd6 18.Qd2+ Kc5 19.Qc3+ Kb6 20.Qd4+ Kxa6 21.Qxf6 Kb5 22.Qxd6 d4 23.Qd5+ Kb6 24.Qxd4+ Kb7 25.Qd7+ Ka6 26.Qe7 Kb6 27.Qd7 a6 28.Qd6+ Kb7 29.Ke6 Ka7 30.Kd7
 32+	00:12	 127 437k	9 956k	+13,01	2.Ba6+
 32+	00:12	 128 234k	9 953k	+13,13	2.Ba6+
 32+	00:19	 201 198k	10 209k	+80,73	2.Ba6+
 32	00:19	 201 198k	10 209k	+80,73	2.Ba6+ Kb8 3.b7 Bf6+ 4.Kc4 Be5 5.Kc5 Bg3 6.Kd5 Bf4 7.f6 gxf6 8.Ke4 Bh6 9.Kf5 Bg5 10.Kg6 Bh4 11.Kf7 Bg5 12.Ke7 Kc7 13.bxa8Q Kb6 14.Qb7+ Kc5 15.Qxa7+ Kb4 16.Kf7 d5 17.Qd4+ Ka3 18.Qxd5 Bc1 19.Qd3+ Kb2 20.Qd4+ Kc2 21.Bc4 Bh6 22.Bd3+ Kd2 23.Bf5+ Ke2 24.Bg4+ Ke1 25.Qg1+ Kd2 26.Qh2+ Kd3 27.Qxh6


Re: 4 old zugzwangs where modern engines still have troubles

Posted: Thu Jul 14, 2016 8:13 pm
by peter
Vinvin wrote: [d]rk1b4/p2p2p1/1P6/2R2P2/8/2K5/8/5B2 w - - 0 1

1. Rc8+ Kxc8 2. Ba6+ Kb8 3. b7 Bf6+ 4. Kc4 Be5 5. Kc5 Bh2 6. Kd4 Bf4 7. Kd5
Bh2 8. f6 gxf6 9. Ke4 Bg3 10. Kf5 Be5 11. Kg6 Bb2 12. Kf7 f5 13. Ke8 f4 14.
Kxd7 Bc3 15. bxa8=Q+ Kxa8 16. Kc8 f3 17. Bb7# 1-0
Very well!

But the way to mate seems some moves longer, MateFinder's output after Backward to 10th, here after 10...Bh4 instead of ...Be5:


[d]rk6/pP1p4/B4p2/5K2/7b/8/8/8 w - - 0 1

Analysis by Stockfish MateFinder 64 POPCNT:

11.Kg6 f5 12.Kf7 Bg5 13.Ke8 Kc7+ 14.bxa8Q Kd6 15.Qxa7 Kd5 16.Qxd7+ Ke5 17.Bb7 Be3 18.Qd5+ Kf4 19.Qf3+ Ke5 20.Qxe3+ Kf6 21.Qe7+ Kg6 22.Qh4 f4 23.Be4+ Kg7 24.Qxf4 Kg8 25.Qf8#
+- (#15) Depth: 53/30 00:00:45 2279MN

Re: 4 old zugzwangs where modern engines still have troubles

Posted: Fri Jul 15, 2016 9:56 am
by Vinvin
peter wrote: But the way to mate seems some moves longer, MateFinder's output after Backward to 10th, here after 10...Bh4 instead of ...Be5:
Sure, my line was only to show simply how white mates.