All five Kingmoves (1.Kb1, Kd1, Kb2, Kc2, Kd2) are mate in 46, (with Houdini 6 foreward/backwards solved)
my pgn see here:
http://www.talkchess.com/forum/viewtopi ... 91&t=65176
Thread: The trojan horse puzzle
"IQ test" for engines
Moderator: Ras
-
- Posts: 1204
- Joined: Thu Dec 25, 2008 9:07 pm
- Full name: Herbert L
-
- Posts: 3621
- Joined: Wed Mar 08, 2006 8:15 pm
- Full name: Jouni Uski
Re: "IQ test" for engines
I tried position with SF matefinder. It has +17,xx evaluation after 5 minutes with 80/110 plies. But main line with some 70 plies has no progress at all
. End position was evaluated as 0,00. Difficult!

Jouni
-
- Posts: 548
- Joined: Sat Aug 17, 2013 12:36 am
Re: "IQ test" for engines
On my machine it fails high at depth 64, and by depth 66 it displays the winning line. Took a little over 10 mins on 1 core, 512MB hash, with 5-man TBs.
Out of curiosity, I re-ran the test with 4 threads and 2GB of hash, and it still had not failed high at depth 71.... It appears to have trouble finding the Na3 path.
I tried 1 thread again with 2GB hash and it sees Na3 for a bit but abandons the path...
4 threads, 16MB hash no luck....
4 threads, 512MB hash no luck
Basically, it seems to be that if it doesn't find the full path (to at least a pawn capture) for the Na3 move while it is in the PV, it gets abandoned, moved down the move ordering list, and subject to much more pruning which makes it even more difficult to find.
But, if the display is showing +17 with "no progress" then the GUI is not displaying the correct line. +17 should mean SF is up at least a queen
.
Out of curiosity, I re-ran the test with 4 threads and 2GB of hash, and it still had not failed high at depth 71.... It appears to have trouble finding the Na3 path.
I tried 1 thread again with 2GB hash and it sees Na3 for a bit but abandons the path...
4 threads, 16MB hash no luck....
4 threads, 512MB hash no luck
Basically, it seems to be that if it doesn't find the full path (to at least a pawn capture) for the Na3 move while it is in the PV, it gets abandoned, moved down the move ordering list, and subject to much more pruning which makes it even more difficult to find.
But, if the display is showing +17 with "no progress" then the GUI is not displaying the correct line. +17 should mean SF is up at least a queen

-
- Posts: 1168
- Joined: Sun Feb 14, 2010 10:02 pm
Re: "IQ test" for engines
OK, I feel the pression that I should add this example to the "IQ test" for engines.
[d]3k4/8/7p/2p1p1pP/1pPpPpP1/1P1P1P2/N7/2K5 w - - (win - gain);
next Sting; old dual core; 1024M Hash
BTW. Sting SF 8.8 solves it too, but is not able to display the correct solution (PV) line.
[d]3k4/8/7p/2p1p1pP/1pPpPpP1/1P1P1P2/N7/2K5 w - - (win - gain);
next Sting; old dual core; 1024M Hash
Code: Select all
info depth 62 multipv 1 score cp 509 nodes 1188545311 nps 1226291 time 969219 pv
c1d2 d8e8 d2c2 e8d7 a2c1 d7e7 c1e2 e7f7 c2c1 f7f8 c1d2 f8f7 e2c1 f7g8 d2c2 g8f7
c2d1 f7f6 d1e1 f6e7 e1f1 e7e8 f1f2 e8e7 f2g2 e7d7 c1a2 d7e7 g2h3 e7f6 a2c1 f6e7
c1e2 e7d6 e2g1 d6e6 h3g2 e6e7 g1e2 e7e6 g2h1 e6e7 h1h2 e7e6 e2c1 e6e7 h2h3 e7d8
h3h2 d8e8 h2h3 e8d8
info depth 63
info currmove c1d2 currmovenumber 1
info currmove c1b2 currmovenumber 2
info currmove c1b1 currmovenumber 3
info currmove c1d1 currmovenumber 4
info currmove c1c2 currmovenumber 5
info currmove a2c3 currmovenumber 6
info currmove a2b4 currmovenumber 7
info depth 63 multipv 1 score cp 509 nodes 1367332010 nps 1231223 time 1110547 p
v c1b2 d8e8 b2b1 e8e7 a2c1 e7f7 b1a1 f7e7 a1a2 e7d7 a2b1 d7e6 c1e2 e6e7 e2g1 e7f
6 b1a2 f6f7 a2a1 f7f6 g1h3 f6e7 a1b2 e7f7 b2c1 f7f6 c1d2 f6f7 d2c2 f7e6 c2d1 e6e
7 d1e2 e7e6 e2f2 e6f7 f2g2 f7e8 h3g1 e8e7 g1e2 e7e6 g2f2 e6e7 e2g1 e7e6 g1h3
info depth 64
info currmove c1b2 currmovenumber 1
info depth 64 multipv 1 score cp 517 lowerbound nodes 1393775104 nps 1231896 tim
e 1131406 pv c1b2
info currmove c1b2 currmovenumber 1
info depth 64 multipv 1 score cp 525 lowerbound nodes 1405287277 nps 1229018 tim
e 1143422 pv c1b2
info currmove c1b2 currmovenumber 1
info depth 64 multipv 1 score cp 537 lowerbound nodes 1408225355 nps 1228750 tim
e 1146063 pv c1b2
info currmove c1b2 currmovenumber 1
info depth 64 multipv 1 score cp 555 lowerbound nodes 1411676455 nps 1228529 tim
e 1149078 pv c1b2
info currmove c1b2 currmovenumber 1
info depth 64 multipv 1 score cp 582 lowerbound nodes 1416572235 nps 1228097 tim
e 1153469 pv c1b2
info currmove c1b2 currmovenumber 1
info depth 64 multipv 1 score cp 623 lowerbound nodes 1434391951 nps 1228534 tim
e 1167563 pv c1b2
info currmove c1b2 currmovenumber 1
info depth 64 multipv 1 score cp 684 lowerbound nodes 1607119785 nps 1234806 tim
e 1301516 pv c1b2
info currmove c1b2 currmovenumber 1
info depth 64 multipv 1 score cp 776 lowerbound nodes 1625516221 nps 1234197 tim
e 1317063 pv c1b2
info currmove c1b2 currmovenumber 1
info depth 64 multipv 1 score cp 913 lowerbound nodes 1630773219 nps 1233419 tim
e 1322156 pv c1b2
info currmove c1b2 currmovenumber 1
info currmove c1d2 currmovenumber 2
info currmove c1c2 currmovenumber 3
info currmove c1d1 currmovenumber 4
info currmove c1b1 currmovenumber 5
info currmove a2c3 currmovenumber 6
info currmove a2b4 currmovenumber 7
info depth 64 multipv 1 score cp 1074 nodes 2281124060 nps 1149398 time 1984625
pv c1b2 d8d7 b2a1 d7d8 a2c1 d8c7 c1e2 c7d7 e2g1 d7e8 a1b2 e8f7 b2b1 f7e8 b1c2 e8
e7 g1h3 e7f8 c2b1 f8e7 h3f2 e7f8 b1c2 f8e7 f2d1 e7e6 d1b2 e6d6 c2b1 d6c6 b2a4 c6
d6 a4b6 d6e7 b1c2 e7e6 c2d2 e6e7 b6d5 e7f7 d2e1 f7e6 e1e2 e6f7 e2d1 f7e6 d5c7 e6
d7 c7b5 d7e6 b5a3 e6f7 a3c2 f7e8 d1e2 e8f7 e2f2 f7f8 c2e1 f8f7 f2g1 f7g8 g1g2 g8
g7 g2h3 g7g8 e1g2 g8h7 g2h4 g5h4 h3h4 h7g8 g4g5 h6g5 h4g5 g8h7 g5f6 h7g8 f6e5 g8
f7 e5f4 f7e6 e4e5 e6e7 h5h6
Maybe, I can't be friendly, but let me be useful.