Topic Rybka Support & Discussion / Rybka Discussion / Tactical position baffles all my engs, including winfinder
![[us]](/mwf/v12/flags/us.png "United States")
r1b2rk1/2q1bppp/p2pp3/2n3PQ/1p1BP3/1BN5/PPP2P1P/2KR2R1 w - - 0 1
All engines see that Qh6 is good for white, but there is another way that happened in the game:
1.Bf6. No engine I've tried including Winfinder 2.2 likes this move. After game continuation Rd8, which in fact Rybka 2.3.1 suggests,
1.Bf6 Rd8
r1br2k1/2q1bppp/p2ppB2/2n3PQ/1p2P3/1BN5/PPP2P1P/2KR2R1 w - - 0 2
none of my engines including rybka could find
2.Rg3!! and even after making the move, most engines still don't get it, and play bxc3, as in the game
2...bxc3
r1br2k1/2q1bppp/p2ppB2/2n3PQ/4P3/1Bp3R1/PPP2P1P/2KR4 w - - 0 3
After 3.Rh3, it still takes most engines a long time to realize that h6 doesnt win, let alone even save black, as in the game:
3...h6
4.Qxh6!
Now, engines after a few seconds on average realize that it is mate in a few moves and +-
I've never seen such a tactical position that baffles comps so much..Comments?
The key is that queen sac on h6 which forces mate, even in the position where it is played (after 1. Bf6 Re(d)8 2. Rg3 bc 3. Rh3 h6), some engines like Rybka and Fritz just can't seem to find the sac. It's mate in nine, some engines find it quickly (Hiarcs, Spike), some do not, but to see that idea in the initial position is probably just beyond the horizon of most programs. And I doubt anyone could hold the position after 1.Qh6 anyway.
Black has so many spite checks and ways to sac material to delay the mate (...Nxb3+, ...cxb2+, ...e5 and ...Bg4), and white has sacced so much material, I think it just becomes difficult for a computer to see that far accurately.
What game is this?
Black has so many spite checks and ways to sac material to delay the mate (...Nxb3+, ...cxb2+, ...e5 and ...Bg4), and white has sacced so much material, I think it just becomes difficult for a computer to see that far accurately.
What game is this?
Hi,
2.1o and 2.1d3 both play Rg3 fairly quickly here ....
2.1o and 2.1d3 both play Rg3 fairly quickly here ....
that's interesting, 2.3.1 doesn't find it
![[ph]](/mwf/v12/flags/ph.png "Philippines")
It's because the positions you posted can already be won by alternative moves (that the engine found first) that are enough to win (although not best.)
It's a waste of time to look futher in an already winning position or else they will lose their strength in practical play with time control such as in engine vs engine match.
That's why some engines have option that can be toggled to Analysis-mode if you want more accuracy.
It's a waste of time to look futher in an already winning position or else they will lose their strength in practical play with time control such as in engine vs engine match.
That's why some engines have option that can be toggled to Analysis-mode if you want more accuracy.
the fact that it likes qh6 on move one has nothing to do with how it plays after you input the move bf6. examining how an engine thinks in a different posotion is not a waste of time, many people use engines for analysis rather than just watching them play.
Topic Rybka Support & Discussion / Rybka Discussion / Tactical position baffles all my engs, including winfinder
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill