GameKnot: GameKnot related, Bugged Contest

« TAP TO LOG IN

GameKnot related: Bugged Contest

« Back to forum

From

Message

repecmps
12-Nov-09, 02:46

Bugged Contest

I failed today's puzzle. Being curious I used crafty to find a solution for me (mate in 2 shouldn't
be a problem for that engine...) but he didn't find a solution!!

If it comes from the puzzle, I obviously want it fixed + another chance to pass it.
If it comes from the engine... Oh, well I can't imagine it comes from the engine!

thanks

seb~

darknessvisible
12-Nov-09, 03:38

Nothing wrong

with the puzzle. Engines are not infallible.

Cheers

repecmps
12-Nov-09, 03:48

something wrong

Well, I'm sure any engine can solve a mate in 2 puzzle with so few pieces on the board...
Sadly I don't have another engine to try that puzzle with, so I need someone to confirm for sure
that the puzzle is solvable... anyway I think they publish the solutions on the next day, we'll
see...

b123
12-Nov-09, 05:45

correct

Puzzle is correct, maybe you should try a bit harder yourself, it's not that hard.

kingdawar
12-Nov-09, 05:54

1 - Playing engines are not the best solvers of composed directmates - the playing engine looks for a way to win, while in a composed problem you are looking for a way to meet the stipulation (i.e. there is a tricky mate in three while there are many mundane mates in four)

2 - I think I know what went wrong, but I cannot comment on it yet, the puzzle is still active, can't give away anything!

kansasjayhawk
12-Nov-09, 14:14

I was able to solve today's puzzle within about a minute. It's not impossible.

archduke_piccolo
12-Nov-09, 15:10

I suppose ...

... that this was the puzzle in question?
fen="r3k3/p1p5/Q3K3/8/8/8/8/8 w
White to move and mate in two?

There is a solution - quite a nice one, too.
What I like about this puzzle (unlike many composed 'problems') is that the position is one that could conceivably arise in a real game. As such, even though White has an almost certain win even if he missed the mate in two, such puzzles are valuable in showing something about the power of the pieces.

archduke_piccolo
12-Nov-09, 15:11

Lets try again, shall we?

lighttotheright
12-Nov-09, 15:34

It's solvable. I don't want to say anything else while the puzzle is active.

kingdawar
13-Nov-09, 00:02

Probably you gave the engine the position with Black castling rights, or the engine on its own thought Black to be able to castle. In any case, it can easily be proven that Black cannot castle! If this position is legal, then what was Black's last move? Can't have been a move by the pawn on a7 or c7.

rt4sm
13-Nov-09, 13:09

Humm

Yes, the castling thing explains why the program didn't solve it; well spotted heinz!

radus
13-Nov-09, 17:00

heinzkat: Well, the last piece moved by black could have been captured by white's queen or
king. For example Nb8-a6, followed by Qxa6, the queen coming from any square where it wasn't
giving check to the black king.

kingdawar
13-Nov-09, 17:00

Radus

Qxa6, Black's move again ....

radus
14-Nov-09, 03:01

Ah, good point, haven't thought it through. Shows why I shouldn't post late at night.

rich_sposato
14-Nov-09, 09:24

> Qxa6, Black's move again ....

Yeah, getting two moves in a row would sure make some of these puzzles easier.

baronderkilt
14-Nov-09, 10:20

Sounds Interesting

First Two Moves in a row, to mate in four (two more moves). Could be onto something there; To go along with Help-Mates and the like. But adds solver responsibility for helping create the position to mate from ... When can we get it !!? Three in a Row/ Mate in 7 . . .
Of course Mate in four vs Two Defensive Moves in a Row sounds a little more difficult !??

caknight
16-Nov-09, 17:51

Chess Puzzle #356

I solved the puzzle in just a few minutes but the I thought about black being able to castle & that prompted me to send GAMEKNOT this message.
Subject: Chess Puzzles
Puzzle #356: White to move, mate in 2 moves the position did not take into consideration that black could 0-0-0 then there is no mate in 2.

hexs
17-Nov-09, 00:35

Check the posts first

If you look above, from heinzkat's first Nov-13 intervention, you will see that it Was taken into
consideration. Black cannot castle because due to their position their previous move ought to
have been the rook's or the king's (which would make the castling impossible).

maca
17-Nov-09, 04:41

...

I don't know much about the actual, international rules concerning the chess problems, but where I'm at, full castling rights are to be assumed to exist in any position unless it is specifically noted that they do not. Regardless of the position, and of any consideration how the players could have arrived into it, that is. In my opinion, retroactive analysis seems a bit suspicious way to show things in chess problems generally, because the problems may have positions that are very unlikely to arise in normal chess games. I'd define a chess problem solely as a momentary position, and a sequence of moves following from that position, without taking any stance on what may or may not have happened before that position was achieved.

Regards,
MaCa.

kingdawar
17-Nov-09, 06:22

Full castling rights are assumed, unless it can be proven otherwise

And part of the requirements of a chess problem is that it should be possible to reach the position in a legal sequence of moves from the normal initial position (although it must be said some of the puzzles used for the contest don't qualify)

lighttotheright
17-Nov-09, 19:02

I wouldn't want too much of a clue. It would take the fun out of the puzzle.

There are some chess problems that have a castle as the solution. If the problem states that the castle is available, that kind of defeats the whole purpose of setting up the position - doesn't it?

It really isn't a trick played by the artist. They just want you to think in a way that you may not be accustomed to.

« Back to forum