Go back
So Good

So Good

Only Chess

T
Mr T

I pity the fool!

Joined
22 Jan 05
Moves
22874
Clock
03 Jan 10
Vote Up
Vote Down

I have been testing out a new fritz with some especially difficult puzzles but it seems to be able to solve every one it is given more or less instantly. I didnt realise the computer was quite THAT good now - has anybody else managed to find a mating attack, or win of material that an engine cannot spot? The quest continues....

h

Joined
25 Apr 06
Moves
5939
Clock
03 Jan 10
Vote Up
Vote Down

Examples plz

greenpawn34

e4

Joined
06 May 08
Moves
43363
Clock
03 Jan 10
Vote Up
Vote Down

Apperently the latest boxes are struggling with this one.

White to play and win.



I'm getting close but keep getting bogged down in lines similiar to this.

I'm sure I have the right idea but have only been working on it off and on for
a few days in spare moments since I saw it on Chessbase.

h

Joined
25 Apr 06
Moves
5939
Clock
03 Jan 10
Vote Up
Vote Down

I bet Ba1 is involved (and probably I've seen the thing before)

greenpawn34

e4

Joined
06 May 08
Moves
43363
Clock
03 Jan 10
Vote Up
Vote Down

Is that why the a-pawn is there to stop a Queen on the a-file taking the Bishop?

I have a line where the Queen is tied to the Bishop and f8 but can simply
advance the a-pawn.

Boxes are taking the a-pawn and heading for a draw.

Post the solution, I don't mind.

h

Joined
25 Apr 06
Moves
5939
Clock
03 Jan 10
Vote Up
Vote Down

Originally posted by greenpawn34
Post the solution, I don't mind.
I don't exactly know the solution it's too long ago, but I'm pretty sure I've seen it. Your initial moves must be correct from there it's either Ba1 or Bb2 and zugzwang-ish

greenpawn34

e4

Joined
06 May 08
Moves
43363
Clock
03 Jan 10
Vote Up
Vote Down

I'm naff at this tempo tossing zugzwang stuff, it's too cute.
I need something to calculate (patterns to recall).

I'm a caveman whose attacks usually need a blunder, not an ice skater.

I'll be miffed at not getting it but not as annoyed as I used to get.
There was a time these things would totally devour me.

Hide the solution - I'll give it another try.

N
10. O-O

Kalispell, MT

Joined
05 Jul 08
Moves
23554
Clock
04 Jan 10
1 edit
Vote Up
Vote Down

Originally posted by greenpawn34
Apperently the latest boxes are struggling with this one.

White to play and win.

[fen]n2Bqk2/5p1p/Q4KP1/p7/8/8/8/8 w - - 0 1[/fen]

I'm getting close but keep getting bogged down in lines similiar to this.

I'm sure I have the right idea but have only been working on it off and on for
a few days in spare moments since I saw it on Chessbase.

. .
white to play
--------------}
1. Qc8 Kg8 2. Bc7 Qxc8 3. gxf7+ Kh8 4. Be5 Qc5[/pgn]
1. Qd6+ Kg8 2. gxh7+ Kh8 3. Bxa5 Qc8 4. Bd2 Nb6


Seems like black is unwound here... the only move not leading to made is Nb6......
Seems pretty conclusive from here.


-GIN

N
10. O-O

Kalispell, MT

Joined
05 Jul 08
Moves
23554
Clock
04 Jan 10
2 edits
Vote Up
Vote Down

Originally posted by Tyrannosauruschex
I have been testing out a new fritz with some especially difficult puzzles but it seems to be able to solve every one it is given more or less instantly. I didnt realise the computer was quite THAT good now - has anybody else managed to find a mating attack, or win of material that an engine cannot spot? The quest continues....
how about this position?



These types of positions break many engines, where the very simple rule
Rooks behind Passed pawns
Seems to be violated. Interestingly enough, some engines will see
Re1 and fail to see Ra1
and instead place useless intermezzo's in this ending,
whereby failing horribly.

The Ra8 problem is maddening, where many engines prune the correct move out of
the tree very early. It seems related to a tier scoring issue, where in the table the
black king has several moves before the value of the position will finally begin to
diminish for black, and thus the engines miss it! This position itself, is really an
extension of the horizon effect; only with tier scoring having produced a violent ripple
on the engines ability to play this ending correctly.

N
10. O-O

Kalispell, MT

Joined
05 Jul 08
Moves
23554
Clock
04 Jan 10
1 edit
Vote Up
Vote Down

Even more maddening:

When engines see wins...in clearly drawn positions!


I'm not sure if all engines fail to see this, but I believe most can't recognize the draw.


-GIN

N
10. O-O

Kalispell, MT

Joined
05 Jul 08
Moves
23554
Clock
04 Jan 10
2 edits
Vote Up
Vote Down

Originally posted by heinzkat
I bet Ba1 is involved (and probably I've seen the thing before)
Indeed... Solved I do believe

1. Qc8 Kg8 2. Bc7 (nod to GP34) Qxc8 3. gxf7+ Kh8 4. Be5 (thanks Heinz!) Qc5 (containment)



And this is where your zugzwang takes place heinz!
5. Bb2 Where blacks greatest delay lies in ...Nc7 6. Ba1 and now the a pawn
is the last delay before mate falls.

EDIT: actually black has slightly more delay left... Nb5!
11. Ke6+ Nc3 12. Bxc3+ Qxc3 13. f8=Q#


-GIN

greenpawn34

e4

Joined
06 May 08
Moves
43363
Clock
04 Jan 10
Vote Up
Vote Down

Did it with big clue from Heinz and thrown a red herring by Nowakowski
who made me think I was the wrong path. (I see he has corrected it).

I kept trying to Zug Black by blocking the a-pawn.
Then it dawned on me - let Black Queen and take it.
After 10.Bxa1 Black is in Zuggers and it's mate.

A wonderful piece of chess composition. Superb.

How quick are R3 & F12 getting this?

M. Matous, 1975

h

Joined
25 Apr 06
Moves
5939
Clock
04 Jan 10
Vote Up
Vote Down

Rybka 3 gives +10,23 after 15 hours, 43 minutes and 38 seconds

h

Joined
25 Apr 06
Moves
5939
Clock
04 Jan 10
Vote Up
Vote Down

What about that three pawns problem? I think it kinda sucked

greenpawn34

e4

Joined
06 May 08
Moves
43363
Clock
04 Jan 10
Vote Up
Vote Down

Originally posted by heinzkat
Rybka 3 gives +10,23 after 15 hours, 43 minutes and 38 seconds
My computer give it III+ after XI Hours XII Miutes VIII seconds.

Ceaser I. It's that old it gives me the time taken in Latin.

Cookies help us deliver our Services. By using our Services or clicking I agree, you agree to our use of cookies. Learn More.