Tough position for SF from GM game (looks like it should be solvable by the engine)

304 views
Skip to first unread message

leotr...@gmail.com

unread,
Jul 25, 2021, 5:23:40 PMJul 25
to FishCooking
r2q1r2/1b2bpkp/n3p1p1/2ppP1P1/p6R/1PN1BQR1/NPP2P1P/4K3 w - - 0 1

sometimes SF quickly finds it on depth 41, once on 50 and once didn't find on 52.

Ivan Ivec

unread,
Aug 12, 2021, 4:02:32 PMAug 12
to FishCooking
It is solvable by engines (see attached).

Very beautiful queen sacrifice.

c2_position3.png

Rob FRage

unread,
Aug 18, 2021, 4:47:12 AMAug 18
to FishCooking
It is pretty quickly findable.
Ceres using the Leelenstein 15 network finds it at depth 13 for me in around 1 minute 20 seconds (resutls may vary with different GPUs). Ceres using the J94-130 network finds it at depth 12 in roughly the same time. LC0 with the J94-130 network finds it even faster at 20 seconds for me.
Latest stockfish compile (using nn-ac5605a608d6, 8 cores, 16gb hash) from today has a harder time with the position but not as bad as depth 52. It finds it consistently at depth 39 in 1 minute 30 seconds

An Stable

unread,
Aug 18, 2021, 10:04:22 AMAug 18
to FishCooking
I wonder where this position is coming from since it is almost equal to and with mainly the same solution as the famous game between Asauskas and Malisauskas in Vilnius 2004:
1. e4 c5 2. Nf3 d6 3. d4 Nf6 4. Nc3 cxd4 5. Nxd4 a6 6. Bc4 e6 7. Bb3 Be7 8. g4 0-0 9. g5 Nfd7 10. Rg1 Nc5 11. Be3 Nxb3 12. axb3 Nc6 13. Nxc6 bxc6 14. e5 d5 15. Ra4 g6 16. Rh4 c5 17. Rg3 Kg7 18. Qf3 Bb7
and here Asauskas found 19. Qf6 !

Best regards
Andreas
Reply all
Reply to author
Forward
0 new messages