Please turn on javascript in your browser to play chess.
Only Chess Forum

Only Chess Forum

  1. 03 Jan '10 18:54
    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....
  2. 03 Jan '10 18:56
    Examples plz
  3. 03 Jan '10 19:27
    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.

  4. 03 Jan '10 19:29
    I bet Ba1 is involved (and probably I've seen the thing before)
  5. 03 Jan '10 19:38
    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.
  6. 03 Jan '10 19:46
    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
  7. 03 Jan '10 19:54
    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.
  8. Standard member Nowakowski
    10. O-O
    04 Jan '10 00:38 / 1 edit
    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
  9. Standard member Nowakowski
    10. O-O
    04 Jan '10 01:00 / 2 edits
    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.
  10. Standard member Nowakowski
    10. O-O
    04 Jan '10 01:04 / 1 edit
    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
  11. Standard member Nowakowski
    10. O-O
    04 Jan '10 02:31 / 2 edits
    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
  12. 04 Jan '10 12:08
    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

  13. 04 Jan '10 12:23
    Rybka 3 gives +10,23 after 15 hours, 43 minutes and 38 seconds
  14. 04 Jan '10 12:27
    What about that three pawns problem? I think it kinda sucked

  15. 04 Jan '10 12:32
    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.