chess online free

Chess Online Free

Classify and you will unmask!
Nutty name, real passions
[ Sign up | Log in | Guest ] (beta)
sneaky 33 ( +1 | -1 )
If Fritz is so strong why can't it solve this? Here is a real beautiful puzzle to solve. Computers
will have a very tough time with this (the ones I
have can't do it.)

"White to play and win."

White: Qc8, Bd8, Kf6, Pg6
Black: Na8, Qe8, Kg8, Pa5, Pf7, Ph7

I've double-checked the above set-up, this is NOT a
misprint or typo! White wins this position by force in
a most spectacular way.

--Sneaky
kremator 3 ( +1 | -1 )
It could If it had a super poweruful processor.
sirk 11 ( +1 | -1 )
Comp Killer Here's my favorite...
White to play and Draw

White:
Ke3,Bb4, Bc8
Pawns: a3, b3, c4, d5, e5, f4, g3, h2

Black:
Kb6, Rb8, Re8, Bg6,
Pawns:a4, b5, c6 d7, e6, f5, g4, h3

Comps are helpless
coyotefan 2 ( +1 | -1 )
Computers are not helpless Their programmers are.
olympio 2 ( +1 | -1 )
no their hardware is
sneaky 9 ( +1 | -1 )
no their methodology is

Still waiting on an answer, by the way. White to
play and win, absolutely astounding.
caldazar 55 ( +1 | -1 )
Sneaky: My chess engine found the win roughly as fast as I did (I think; I didn't time myself or anything). It's just a question of computing power and how fast a processor you need for the program to calculate a dozen moves (or so) out.

Sirk's problem highlights a problem with chess computers though. With current chess engines, you'd need hardware that could calculate several dozens of moves out for the computer to recognize the draw because current engines calculate in terms of moves rather than overall ideas.
sirk 72 ( +1 | -1 )
The only concievable way that a comp could solve the problem above(under current methodology) is to calculate out 50 moves and find no way to force a win, pawn move or capture. I can't even begin to imagine how long it would take to compute 100 ply with so many pieces on the board.

In my opinion even the best computer's don't really "play" chess, there just highly sophisticated, combination machines. You can spuce them up with opening books, tablebases, etc, you can even make them assigh material values to positional considerations as in the Kasparov-Junior Bxh2 game. but in the end the're just adding up numbers not "playing"

ps. I should add that I found the above problem in a book, so it is not really "my" problem
bogg 6 ( +1 | -1 )
OK, I give up. I came up with 1. Bc7?? Qc8: 2. gf:+ Kh8 3. Be5 but this loses to Qc5. What is the solution?
caldazar 17 ( +1 | -1 )
Bogg... You're on the right track; keep going. The important thing to note is that both Black's back rank and the long diagonal leading to his king are weak.
zdrak 50 ( +1 | -1 )
1.Bc7 Qxc8 2.gxf7+ Kh8 3.Be5 Qc5 4.Bb2! (but not Ba1, see below) Nc7 (to guard e6) - and now black is almost in zugzwang. His queen is at the only vantage point from which to cut off the white king AND guard f8. His knight guards e6. His h-pawn guards g6, and his king is stuck in the corner. So all is left is to move the a-pawn.

5.Ba1 a4 6.Bb2 a3 7.Ba1 (had white played 4.Ba1 it would have been his turn now, and he's lose. But now, it's black's turn) b2 8.Bb2 - and ONLY NOW, does my (old version of) Fritz notice that black is about to be checkmated in a few moves.
After 8...a1=Q 9.Bxa1 the remarkable zugzwang is complete.
bogg 62 ( +1 | -1 )
nice! A bit of the old human horizon effect shown by my analysis. BTW after 3. ... Qc5 it takes Hiarcs only 2 seconds to find the mate and Shredder only 10. Both of these programs do much better on the Nunn tactical test than fritz. There is also a web site, sorry I forget the address, where the webmaster has a set of about 10 positions he uses to evaluate chess programs. When I ran onto this site a few months ago Hiarcs was the only program that had scored points in all of the test cases just nosing out Shredder which frequently got a more correct answer but missed some completely.