Zitat von
EST
Lt. (Deep)Hiarcs 15.3.1 wird etwa in der nachfolgend dargestellten Endspielstudie (Schwarz am Zug) auch nach langer Rechenzeit und großer Rechentiefe ..Td1? vorgeschlagen und Weiß gewinnt.
Hello Friedhelm,
that is weird
. I gave it several tries, Hiarcs 15.3.1 best move has always been 1...Rxa7 from my experience.
First short run, less than 20 secs, two threads:
Code:
0:19/24-62, 1/18: 1...Rxa7, 37,3M nodes (1,9M nps), 128MB hash, 2 threads
-2.25/23 1...Rxa7 2.Rh8+ Ke7 3.Rh7+ Ke6 4.Rxa7 Rd1+ 5.Kxd1 cxb2 6.Ra6+ Ke7 7.Kd2 b1=Q 8.Ra7+ Kd6 9.Ra6+ Kc5 10.Ra5+ Kc6 11.Ra6+ Kb7 12.Ra4 bxa3 13.Rxa3 Qg1 14.Rb3+ Kc6 15.Rd3 Qxg5+ 16.Kc3 Qg3 17.Kb3 Qf2 18.c3 Qb6+ 19.Kc2 Qc5 20.Kb3 Qb5+ 21.Kc2
+2.04/23 1...Rd1+ 2.Kxd1 Rxa7 3.Rh8+ Kg7 4.Rb8 Ra4 5.Kc1 cxb2+ 6.Kxb2 bxa3+ 7.Kb1 Rc4 8.Re8 Rc3 9.Rxe5 Rxf3 10.Rc5 Re3 11.Ka2 Rxe4 12.Kxa3 Kf7 13.Kb2 Ke6 14.Ra5 Rg4 15.Kb3 Rg3+ 16.Kb4 Kd7 17.c4 Rg1 18.Re5 Kd6 19.Rd5+ Ke6 20.Kc3 Rg4
+4.17/23 1...Kg8 2.bxc3 bxa3 3.Rhg7+ Kh8 4.c4 Rac8 5.Rh7+ Kg8 6.Rag7+ Kf8 7.Rc7 Rxc7 8.Rxc7 Ra8 9.Kb1 Ra6 10.Ka2 Rg6 11.Kxa3 Rxg5 12.Rc5 Kg7 13.Rd5 Rg3 14.Rd3 Kf6 15.Kb4 Rg8 16.c5 Kg5 17.Kb5 Kf4 18.c6
+5.44/23 1...cxb2+ 2.Kxb2 bxa3+ 3.Kxa3 Rxa7+ 4.Rxa7 Rd6 5.Ra5 Re6 6.Kb4 Kg7 7.Kc4 Kf7 8.Rc5 Re8 9.Kd5 Rg8 10.Kxe5 Rxg5+ 11.Kd4 Rg1 12.f4 Kf6 13.Rc6+ Ke7 14.f5 Kd7 15.Rc5 Rg4 16.c4 Rg8 17.Rd5+ Ke7
#2/23 1...Rd7 2.Rxa8+ Rd8 3.Rxd8#
#1/23 1...Rab8 2.Rh8#
#1/23 1...Rdb8 2.Rh8#
...
second run, up to depth 30:
Code:
2:01/31-66, 1/18: 1...Rxa7, 376,3M nodes (3M nps), 128MB hash, 4 threads
-2.31/30 1...Rxa7 2.Rh8+ Ke7 3.Rh7+ Ke6 4.Rxa7 Rd1+ 5.Kxd1 cxb2 6.Kd2 b1=Q 7.Ra6+ Ke7 8.axb4 Qxb4+ 9.Kc1 Qd4 10.Rf6 Qg1+ 11.Kb2 Qxg5 12.Rf5 Qg3 13.Kb3 Qg8+ 14.Kb4 Qb8+ 15.Kc4 Qc8+ 16.Kb4 Qb7+ 17.Kc4 Qa6+ 18.Kb4 Qd6+ 19.Kc4 Qc6+ 20.Kb4 Qb6+ 21.Kc4 Ke6 22.c3 Qd8 23.Kb4 Qd3 24.Kb3 Qd1+ 25.Kb4 Qb1+ 26.Kc4 Qf1+ 27.Kb4 Qa6 28.Rg5 Qb7+ 29.Kc4
+2.31/30 1...Rd1+ 2.Kxd1 Rxa7 3.Rh8+ Kg7 4.Rb8 Ra4 5.Kc1 cxb2+ 6.Kxb2 bxa3+ 7.Kb1 Rc4 8.Rb5 Rc3 9.Rxe5 Rxf3 10.Rc5 Rg3 11.Ka2 Kg6 12.c3 Kh5 13.Kxa3 Kg4 14.g6 Kf4 15.Rc6 Kxe4 16.Kb4 Kd5 17.Kb5 Ke5 18.c4 Kd4 19.Ra6 Ke5 20.c5 Kd5 21.Rd6+ Ke5
+4.85/30 1...Kg8 2.bxc3 bxa3 3.c4 Rdb8 4.Rhg7+ Kh8 5.Rgb7 Rc8 6.Kb1 Rxa7 7.Rxa7 Rxc4 8.Rxa3 Rc6 9.Ra8+ Kg7 10.Re8 Rc5 11.Re6 Kf7 12.Rf6+ Ke7 13.Rf5 Kd6 14.f4 exf4 15.Rxc5 Kxc5 16.Kc1 Kd4 17.g6 Kxe4 18.g7 Ke5 19.c4 Kf6 20.g8=Q Kf5 21.Kd2
+5.79/30 1...cxb2+ 2.Kxb2 bxa3+ 3.Kxa3 Rxa7+ 4.Rxa7 Rd6 5.Ra5 Re6 6.Kb4 Kf7 7.Kc4 Re8 8.Kd5 Rg8 9.c4 Rxg5 10.c5 Ke7 11.Ra7+ Kd8 12.Kd6 Kc8 13.Rf7 Rg8 14.Kxe5 Kb8 15.f4 Re8+ 16.Kd5 Rd8+ 17.Kc6 Rg8 18.e5 Rg6+ 19.Rf6 Rg3 20.f5
#2/30 1...Rd7 2.Rxa8+ Rd8 3.Rxd8#
#1/30 1...Rab8 2.Rh8#
#1/30 1...Rdb8 2.Rh8#
...
and third run, 10 minutes with moves selection to get deeper:
Code:
10:00/39-82, 1/18: 1...Rxa7, 2,2G nodes (3,8M nps), 128MB hash, 4 threads
-2.63/38 1...Rxa7 2.Rxa7 Rd1+ 3.Kxd1 cxb2 4.Kd2 b1=Q 5.axb4 Qxb4+ 6.Kc1 Qd6 7.g6 Qxg6 8.Ra8+ Kg7 9.Ra7+ Kh6 10.Ra3 Qg1+ 11.Kb2 Kg5 12.Rb3 Kf4 13.Rd3 Qc5 14.Ra3 Qd4+ 15.Kb1 Qf2 16.Kb2 Qe1 17.Rb3 Kg3 18.Ra3 Qd2 19.Rd3 Qb4+ 20.Rb3 Qd4+ 21.Kb1 Qd1+ 22.Kb2 Kf4 23.Rd3 Qf1 24.Ra3 Kg5 25.Rd3 Kf6 26.Ra3 Qf2 27.Kc1
+2.59/38 1...Rd1+ 2.Kxd1 Rxa7 3.Rh8+ Kg7 4.Rb8 Ra4 5.Kc1 cxb2+ 6.Kxb2 bxa3+ 7.Kb1 Ra7 8.Rb5 Rf7 9.Rxe5 Rxf3 10.Rf5 Rc3 11.Rf2 Kg6 12.e5 Re3 13.Ka2 Kxg5 14.e6 Rxe6 15.Kxa3 Ra6+ 16.Kb4 Rb6+ 17.Kc5 Rb2 18.Kd4 Ra2 19.Kd3 Ra3+ 20.c3 Kg4 21.Rf8 Kh5 22.Kd4 Ra1 23.Rf5+ Kh6 24.c4
Out of curiosity, I also tried the non-deep exe file, and I got the same correct output.
Are you using a Mac version? (mine is Windows 11).
MfG,
Eric