< Earlier Kibitzing · PAGE 10 OF 11 ·
|Jun-27-07|| ||srulikbd: does such blunders would happen nowadays? I mean, what rybka suggests?|
|Jun-27-07|| ||argishti: One of the best games ever played against a computer!!|
|Jun-29-07|| ||argishti: <Sneaky: 14...Bd6 is one of the most preposterous moves I've ever seen.>|
the comp is trying to be a wise guy here. thinking garry will fall for it. but seriously, why would the machine even play such a thing??
|Aug-20-07|| ||Crowaholic: <mizer: NP-complete problems (chess)>|
mizer, chess on an 8x8 board is not an NP-complete problem, it is O(1). For a very very large value of O(1), of course. :-)
Even if you turned chess into a parametric problem, let's say as CHESS(n) with an n times n board and specific rules to fill the base lines with new pieces, you'd still have to prove that it is a) in NP and b) NP-hard before you can safely claim that it is NP-complete. The actual decision problem would be something like figuring out whether White has a forced win.
At first I thought that to prove a) it would be sufficient to show that the maximum number of moves in a game is polynomially bounded, which I thought was easy thanks to the 50 move rule: 50 * (max. number of captures + max. number of pawn moves) yields a polynomial. But there are two problems: For one, no player is actually required to invoke the 50 move rule. He could invoke it when he sees he's going to be checkmated, but if a pawn moves or pieces are captured in the meantime, the counter is reset to zero. And most importantly, to solve CHESS(n) it is not enough to find a position where Black is checkmate. You have to show that the sequence of moves leading up to it is forced, i.e. in case of any other move, White has an equally short or shorter mate! But the non-deterministic Turing machine, which has simply been described as the "luckiest possible guesser" cannot just guess whether a move is forced, because it needs to know ALL subsequent lines to do that, not just any winning line. This leads me to believe that CHESS(n) would more probably be in EXPTIME.
Sorry for the computer science lecture, but it troubles me when terms like NP-complete are used lightly.
|Aug-29-07|| ||unixfanatic: This game is such a classic example of how pathetic computers can be in closed positions - it's quite surprising how in certain positions or openings (King's Indian Defense is a good example) long pawn chains can really overwhelm even the most clever of the silicon beasts!|
|Dec-19-07|| ||enoughsaid05: X3D has already committed 2 positional mistakes in my opinion especially during queen's gambit declined|
These two moves led the computer into closed games, which is the computer's greatest weakness due to lack of positional knowledge.
|May-23-08|| ||dreamer1991: 10.. e4??
I think 10..exd4 is much better than 10..e4
|Aug-26-08|| ||Katu: <dreamer1991: 10.. e4??
I think 10..exd4 is much better than 10..e4>
But 11. Nxd4 and the position remains closed. (I suppose. I'm not using Fritz or Rybka and I'm not a GM)
|Sep-02-08|| ||Funicular: Katu:
Yes. Not only position remains somewhat closed (exactly what kasparov was looking for, claiming that is precisely there when computers can't analyze properly), but also brings another knight into action, threatens winning c6, and furthermore, slams open b column. With all that q-side pressure and all his pieces aligned, it would be tremendous positional advantage for white. Black has no kingside counterplay.
|Nov-06-08|| ||hedgeh0g: Hmm...after 10...exd4 11. Nxd4, can't Black simply take the c-pawn with advantage?|
|Dec-07-08|| ||dfelix: I remember watching this game live on ESPN. Maurice Ashley was one of the commentators, and at one point he said something like "The computer thinks it has an advantage. That's just retarded."|
|Dec-07-08|| ||Once: Great anti-computer game. Kasparov builds his attack so slowly that the crunch point is always beyond the computer's ability to calculate. Fritz X3D was presented with so many alternatives on each move that it burned its computing power on millions of moves that were never played.|
But let's not get too excited about the perfect anti-computer strategy. This is not John Connor in the future about to smash skynet.
This is a game played back in 2003, when the early hunter-killers had rubber skin. We spotted them easy.
GMs would not get such an easy ride against modern chess engines. The programmers watched games like this and adjusted, tinkered, fiddled. They changed the engine's opening books, upgraded their understanding of closed positions, introduced the ability to make speculative sacrifices to open a position up to tactics. He doesn't feel pity or remorse. And he absolutely will not stop, ever, until you are dead.
If Fritz X3D could have spoken in 2003, he might have said, in an Austrian Governor monotone "I'll be back".
Or think a pregnant Sarah Connor in an open top jeep back before she went all muscly ... there's a storm coming.
|Dec-07-08|| ||macphearsome: terminator references, computer science lectures...
this discussion has everything I need.
|Dec-07-08|| ||ILikeFruits: one of these days...
machines shall rule the world...
running on veggie oil...
|Dec-07-08|| ||ILikeFruits: on a side note...
today is the birthday of...
don vito corleone...
|Dec-07-08|| ||lefthandsketch: in response to argishti- no, this is the greatest anti-computer game ever: Rybka v.s. Hikaru Nakamura 2008- six underpromotions and mate with 5 bishops.|
|Dec-07-08|| ||sleepyirv: It's too bad computers don't know shame or Fritz might have given up chess after this.|
|Dec-07-08|| ||SuperPatzer77: Holy Toledo!!! This computer (X3D Fritz) has finally resigned to a human being like Gary Kasparov. I thought almost all the chess computers would refuse to resign - LOL. |
Bravo, Kasparov!! Kasparov must have smiled at this computer because it resigns. Naturally, it is not premature. Kasparov creates powerful pressure on the Queen side.
|Dec-07-08|| ||SuperPatzer77: <lefthandsketch: in response to argishti- no, this is the greatest anti-computer game ever: Rybka v.s. Hikaru Nakamura 2008- six underpromotions and mate with 5 bishops. >|
<Lefthandsketch> Yeah, it is very interesting. I want to mention to you: Crafty (Computer) vs Hikaru Nakamura - ICC Blitz 2007 with five underpromotions and mate with six knights.
It is awesome, <Lefthandsketch>, right?
|Dec-07-08|| ||Once: <SuperPatzer77: I thought almost all the chess computers would refuse to resign - LOL.>|
More likely Fritz's operators decided that they had seen enough and resigned on behalf of the computer.
Modern chess software can be made to resign early, later or never. Just tweak the settings.
|Dec-07-08|| ||Marmot PFL: No point in resigning here. Even the best human can blunder.|
|Dec-07-08|| ||lefthandsketch: super patzer- I wanted to list that one too- I think it is something nakamura likes to do in blitz games with computers to further humiliate them. In any case, it's hilarious anti-computer absurdity.|
|Dec-07-08|| ||SuperPatzer77: <Lefthandsketch> Yeah, that's awfully hilarious below:|
White King represents the British investigator Ichabod Crane. Six Black Knights represent the six headless horsemen in the scary town of Sleepy Hollow, England. LOL
That's what I wrote in the game between Crafty (White) and Hikaru Nakamura (Black) - a hilarious commentary.
|Dec-07-08|| ||playground player: Perhaps the computer got dispirited.|
|Dec-07-08|| ||WhiteRook48: How can the computer resign?|
< Earlier Kibitzing · PAGE 10 OF 11 ·