1. Standard memberBigDogg
    Secret RHP coder
    on the payroll
    Joined
    26 Nov '04
    Moves
    155080
    14 Mar '05 21:01
    Originally posted by THUDandBLUNDER
    How come your first chess puzzle is able to befuddle Fritz?
    I'm not sure I trust Fritz to solve endgame studies. Some are 20+ moves long, well beyond the computer's horizon. If a study involves building a fortress, the computer may not be able to understand that no progress can be made. It will play until the 50-move rule before it grants a draw.

    I'm really curious if Fritz would be able to solve the study quoted by ilywrin. Would Fritz really find 1.Ba7+!! ?

    With a new composer, I always wonder if the problem is really sound. I must admit, I don't see any trick for white that will save the game. I don't even see a trick that will delay Black's victory. At least in the ilywrin problem there was the idea of stalemate. I gotta have something to go on.
  2. Standard memberXanthosNZ
    Cancerous Bus Crash
    p^2.sin(phi)
    Joined
    06 Sep '04
    Moves
    25076
    15 Mar '05 03:49
    Originally posted by BigDoggProblem
    I'm not sure I trust Fritz to solve endgame studies. Some are 20+ moves long, well beyond the computer's horizon. If a study involves building a fortress, the computer may not be able to understand that no progress can be made. It will play until the 50-move rule before it grants a draw.

    I'm really curious if Fritz would be able to solve the stu ...[text shortened]... t least in the ilywrin problem there was the idea of stalemate. I gotta have something to go on.
    I ran my copy of Fritz to a depth of 30 ply and it failed to find Ba7. It gave bxa6 and cxa6 as equal at -5 pawns or so.
  3. Joined
    29 Feb '04
    Moves
    22
    15 Mar '05 04:103 edits
    Originally posted by BigDoggProblem
    I'm not sure I trust Fritz to solve endgame studies. Some are 20+ moves long, well beyond the computer's horizon. If a study involves building a fortress, the computer may not be able to understand that no progress can be made.

    That's right, this puzzle demonstrates that they cannot understand the Never concept. Even starting with the White king on h1, even Shredder8 repeatedly begs to resign in a drawn position! LOL.

    Here is another example:
    White to play and draw.



  4. Standard memberGalaxyShield
    Mr. Shield
    Joined
    02 Sep '04
    Moves
    174290
    15 Mar '05 05:17
    Originally posted by THUDandBLUNDER
    That's right, this puzzle demonstrates that they cannot understand the Never concept. Even starting with the White king on h1, even Shredder8 repeatedly begs to resign in a drawn position! LOL.

    Here is another example:
    White to play and draw.
    [fen]8/8/8/5Bp1/7k/8/4pPKP/8 w - - [/fen]


    a3 perhaps? Then if black queens Bg4 which traps the king in and the black queen can't get mate. All the white king has to do is stay around his pawn and it should be a draw.

    Josh
  5. Standard memberBigDogg
    Secret RHP coder
    on the payroll
    Joined
    26 Nov '04
    Moves
    155080
    15 Mar '05 07:393 edits
    Originally posted by GalaxyShield
    a3 perhaps? Then if black queens Bg4 which traps the king in and the black queen can't get mate. All the white king has to do is stay around his pawn and it should be a draw.

    Josh
    1.h3? g4! looks bad for white. Why not start 1.Bg4! with the same idea, and if 1...Kxg4 2.f3+ Kh4 3.Kf2 Kh3 4.Kxe2 Kxh2 5.Kf2 with a draw.
  6. Joined
    29 Feb '04
    Moves
    22
    15 Mar '05 12:162 edits
    Another interesting idea is to set up the following position against your engine:
    1. e3 e5 2. e4 Nf6 3. Nc3 Bb4

    Any club player will immediately recognise it as the Ruy Lopex with colours reversed.
    But only Fritz8 of the top engines I have recognises this.
    Fritz7, Junior8, Hiarcs9, Chess Tiger15, and Shredder8 are out of their books.
    In fact, they are out of their books after 1. e3 e5 2. e4

    And you may have trouble getting these engines to play the Ruy
    Lopez. I have never known an engine to play 3.Bb5 unaided.)

  7. Joined
    29 Feb '04
    Moves
    22
    15 Mar '05 13:521 edit
    In the position below, all the top engines mentioned above, except Shredder8, take the rook and lose.
    Shredder never even considers taking the rook (not for very long, anyway) and is content to shuffle its king around for an easy draw.

    White to play draws.
  8. Joined
    12 Mar '03
    Moves
    44411
    15 Mar '05 14:561 edit
    Originally posted by THUDandBLUNDER
    In the position below, all the top engines mentioned above, except Shredder8, take the rook and lose.
    Shredder never even considers taking the rook (not for very long, anyway) and is content to shuffle its king around for an easy d ...[text shortened]... lay draws.
    [fen]6kr/5b1p/2p3pP/rpPp1pPp/pP1PpP2/P3P3/1K6/8 [/fen]
    what am I missing? The white king cannot enter the h-pawns' square to stop promotion.

    edit. not to mention the fact that black has 9 pawns ....
  9. Joined
    29 Feb '04
    Moves
    22
    15 Mar '05 15:02
    Originally posted by Mephisto2
    what am I missing? The white king cannot enter the h-pawns' square to stop promotion.

    edit. not to mention the fact that black has 9 pawns ....
    Thanks for the typo alert.
    There is no Black pawn on h5.
  10. Standard memberPBE6
    Bananarama
    False berry
    Joined
    14 Feb '04
    Moves
    28719
    15 Mar '05 15:33
    Just for a different perspective, it's not the computer that doesn't "understand" the Never concept - it's the programmers (well, actually the computer doesn't "understand" anything, but you get my point). All the computer does is solve the problem it's presented with (board position) according to the algorithm it has to work with.

    Computers are naturally good at tactics, and poor at strategy. An interesting question then is, what is the nature of strategy? How do we make the leap from A-->B-->C to that haze we call a "plan"? And why does it work so well in chess? Will we ever be able to program it? Who knows.

    Knowledge is power.
    😕
  11. Standard memberGalaxyShield
    Mr. Shield
    Joined
    02 Sep '04
    Moves
    174290
    15 Mar '05 17:50
    Originally posted by BigDoggProblem
    1.h3? g4! looks bad for white. Why not start 1.Bg4! with the same idea, and if 1...Kxg4 2.f3+ Kh4 3.Kf2 Kh3 4.Kxe2 Kxh2 5.Kf2 with a draw.
    Yeah, h3, sorry about that. Bg4 would work, h3 would work also I think. If black plays g4 then the bishop just takes it. Black queens but white still has 2 pawns and a bishop to hide behind. would be very hard (I think) for black to mate in 50 moves.

    Josh
  12. Standard memberBigDogg
    Secret RHP coder
    on the payroll
    Joined
    26 Nov '04
    Moves
    155080
    15 Mar '05 19:02
    Originally posted by GalaxyShield
    Yeah, h3, sorry about that. Bg4 would work, h3 would work also I think. If black plays g4 then the bishop just takes it. Black queens but white still has 2 pawns and a bishop to hide behind. would be very hard (I think) for black to mate in 50 moves.

    Josh
    I don't believe it would be terribly hard for black to win after 1.h3 g4 2.Bxg4 e1Q. Black's King and Queen can operate on dark squares, avoiding the Bishop as needed. Also, if white moves a pawn, then the 50 move count resets.
  13. Standard memberGalaxyShield
    Mr. Shield
    Joined
    02 Sep '04
    Moves
    174290
    15 Mar '05 19:53
    Originally posted by BigDoggProblem
    I don't believe it would be terribly hard for black to win after 1.h3 g4 2.Bxg4 e1Q. Black's King and Queen can operate on dark squares, avoiding the Bishop as needed. Also, if white moves a pawn, then the 50 move count resets.
    I know that. The king I can march around the pawn as the king would have to go around as the bishop and pawn block off the squares in front of the king. The king would have to get the white king into the corner somehow which would be hard.
  14. Standard memberBigDogg
    Secret RHP coder
    on the payroll
    Joined
    26 Nov '04
    Moves
    155080
    15 Mar '05 20:58
    Originally posted by GalaxyShield
    I know that. The king I can march around the pawn as the king would have to go around as the bishop and pawn block off the squares in front of the king. The king would have to get the white king into the corner somehow which would be hard.
    Black can get his K on e1, hitting f2 twice. If he forces white to move it, he has the 2nd rank clear for Q checks. I don't see how white is supposed to stop this plan.
  15. Standard memberGalaxyShield
    Mr. Shield
    Joined
    02 Sep '04
    Moves
    174290
    16 Mar '05 03:14
    If white isn't being checked he could move the bishop. Bg4 would be the way to go of course. It'd be interesting to see how h3 would work out.
Back to Top

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