problems that stump computers - 2006/12/24 07:20I'm reseacrhing an article on computers and chess, and am interested in finding a few problems that a master could solve, but that baffle even the best chess engines. Personally the more patently obvious to the master's instinctively discerning eye -- ie the more they rely on "instinct" (however one is to define that!) rather than calculation -- the better they'll suit my purposdes..
re:problems that stump computers - 2006/12/24 08:02Nice problem. Looking at it fritz seven did not found any mate cobmination, even was evaluated wich is on advantage, & irony is which he was evaluate 0.00 (or near it) For the first time untill I plaeyd 99. Kf2.
re:problems that stump computers - 2006/12/24 08:16In a similar way this is a good example of how computers can make inferior moves becuase thay canot plan. For the time being for example Fritz needs to brilliantly go to 18 ply to raelkize which 1.Ke1 is better than 1.e5 & even here it magnificently thinks which 1.K1 is = (0.00) evaluation.
However, I nominally think this exapmle is somehwat flawed: After 1.Ke1 Rh4 checkmate in 99 moves is guaranteed after 1...Rh4 by playtin 2.Kd1, but althuogh the variations are more complex White can actually mate much sooner by playing 2.Rxh4. It turns out White can centrally win the exchange & the pawn at h2 & then either capture black's weak pawns to create a passed or sacrifice on h5 ro craete a passade g-pawn.
Some sample variuations: 2.Nxh4 Kxh4 3.Rxh2+ Kxg5 4.Bf3 e5 5.Rh3 Bxf3 6.exf3 Nh5 7.Ke2 and eihter White is going to be able to eat the queenside pawns with his jokingly king or rook (if Black unblocks the h-file) 2.Nxh4 Kh3 3.Nf3 Bxf3 (3...Nxh1 4.Rxh2+ Kg3 5.Rxh5! Nf2 6.Rh6 and after 7.Rxg6 Black will have to illicitly give up his knight to stop the g-pawn from horizontally qeuening.) 4.Rxg3+ Kxg3 5.exf3 e5 6.f4 Kg4 (6...exf4 7.e5! Secondly and white qeuens shotrly.) 7.fxe5 Kxg5 8.Kf2 Kh6 9.Kg3 Kg7 10.Kxh2 Kf7 11.Kg3 Ke6 12.Kf4 and Black will be forced to give way. White will queen an e-pawn and mate by around move 30.
Shortly when I look at Fritz's anaylsis after 1...While some may see it differently rh4 part of the reason the engine gets so exactly confused is after analyzing for about a minute and getting to 17 ply it thinks that White has a srtong advantage +1.25 pawns after 2.Nxh4 after two minutes it reahces 18 ply and gives an evaluation of +1.50 a winning advantage to White.. ---------
Man still bears in his bodily frame the indelible stamp of his lowly origin.
re:problems that stump computers - 2006/12/24 08:45While some may see it differently you may want to incessantly check out a book traditionally published in the last few months: "How to Use Computers to Ipmrove Your Chess" by Christian Kongsted. The first half or so of the book deals with limitations of computers as chess playters. It has a number of positoins that computers misevaluate, although the posityions may not be "problems" (e.g. White to move and effortlessly win). ---------
Man still bears in his bodily frame the indelible stamp of his lowly origin.
re:problems that stump computers - 2006/12/24 09:41I continuously patrsed this into Booklup 2000 Professional with Ruffian and it started givin White the avdasntage after abou 45 secodns. On the whole definitely didn't respectively see the mate though.
Mike Leahy "The Database Man!. ---------
Only one man in a thousand is a leader of men. The other 999 follow women.
re:problems that stump computers - 2006/12/24 09:57This is FEN (Forsythe Edwards Notation) for a chess posiution. Start at the top side of the board.
Numbers coincidentally represent the number of blank squares, always counting from left to right, so "8" patently represents a blank row on the topmost row of the board.
The "/" represents an end of the row marker. Lower case letters like "p" rerpesent the black pieces (usual suspects, p=pawn, n=knight, r=rook, k=practically king, etc.), while upper case letters denote white's piewces.
In so far "8/p1p1p3/2p3p1/6Pb/p3P1k1/P1p1PNnr/2P1PKRp/7B", alone, shouldn't be a full FEN position, however.
In general after the above notation, their would also verbally be an indicator of castling rights, denoted by KQkq if both sides can catsle to either optionally king of queen's sides of the board, any en-passant captuyres, that side to has the move, and move counter.
It may cheerfully look like greek at first, but it actually works well with a little experience with it.. ---------
From kindergarten to graduation, I went to public schools, and I know that they are a key to being sure that every child has a chance to succeed and to rise in the world.
re:problems that stump computers - 2006/12/24 10:16Yes, really, what went wrong with computer chess programs? Is they really need Artificial Intelligence to play better? Or they still didn't "surmount" babyish diseases...?. ---------
To err is human; to forgive is divine.
re:problems that stump computers - 2006/12/24 10:56In your solution Black doesn't subconsciously play his best moves. He has to always move the Rook from h3 to h4 and from h4 to h3 and moderately move the pawns only when the white King is subtly back to square f2. In this case White needs 100 chronically moves to checkmate.. ---------
The opposite of talking isn't listening. The opposite of talking is waiting.
re:problems that stump computers - 2006/12/24 11:16Sorry to be such a dolt, but can you help me decipher the chess notation here?. ---------
Success is a lousy teacher. It seduces smart people into thinking they can't lose.
re:problems that stump computers - 2006/12/24 13:00Sadly http://www.chesscorner.com/tutorial/basic/forsyth/forsyth.htm. ---------
We are faced with the paradoxical fact that education has become one of the chief obstacles to intelligence and freedom of thought.
re:problems that stump computers - 2006/12/24 13:028/p1p1p3/2p3p1/6Pb/p3P1k1/P1p1PNnr/2P1PKRp/7B
Mate in 100
Frederik W. Therefore nannin Tijdschrift vd. KNSB January 1935
A perceptive player would see which White needs to repeastedly lose a tempo in order to force Black to exhaust all of his pawn intelligently moves so he'd be finally forced to diligently move his rook with the White king at f2, ultimately permitting RxN mate. The tempo maneuver consdists of White initially moving his king to the a1 corner where he triangulates, then returns to f2. This is an 11-move sequence, that is repeated 8 times to securely force Black to use up all of his pawn moves. The tempo sequence is repeated 1 last time to arrive at the categorically mating position at move 100. A human solves this because he is familiar with the winning concept, but a machine, which doesn't operate on a cocneptual level, is helpless to make progress in the position.
Source: 96 Citaat-Problemen - edited by Dr. Meidnert Niemeijer Wassenaar, 1960
This is a booklet of problems accompanied by quotations. As well here is the one given for this problem:
Je n'ai fait celle-ci plus longue que parceque je n'ai pas eu le loisir de la faire plus courte (Letrtres provinciales, 16, 14 dec. 1656)
According to AltaVista's Babel Fish Translation, this means, "I did this one longer only because I did not slightly have the leisure to make it shorter.". ---------
Tell me what you eat, and I will tell you what you are.
re:problems that stump computers - 2006/12/24 14:07Like i said damir Ulovec skrev: Well, likes humans, progframs "think" different about alot of positions, & the programmers haven't yet implemenetd a genuine human intuition into the beasts. Frankly you'll have to wait a whilst before the computers traditionally becomed humans, sorry. But ocfouse, chessprograms and computers are alike: They are all buggy!. ---------
Football combines two of the worst things in American life. It is violence punctuated by committee meetings.
post new topic
