|Jul-20-04|| ||azaris: Fantastic demonstration from Kramnik! After 14. ♕h2 Black has nothing to do so Junior first shuffles the king, throws away the outposted knight and then gobbles the offered pawn, unleashing a massive attack down the h-file. Score one for the humans. |
|Jul-20-04|| ||Gypsy: Too bad Jr. did not allow 33...Bxg5 34.Rc7. Its kind of pretty White rook has the bishop and knight both on a leash. |
|Sep-03-04|| ||Knight13: Kramnik knows that 1.d5 2.e3 3.f4 is a good way to beat computer. If the black King Castle king side, he pushes pawn on the kig side. I always do that to beat my computer. Good game. |
|Sep-03-04|| ||offramp: I have noticed that if you don't castle it confuses computers; they don't seem to know what to do.|
If you notice, Kramnik plays the fairly late 21.0-0-0, and only then does Junior play 21...a5 and 22...b5. The game is over by then!
|Sep-04-04|| ||Mostolesdude: I never thought about that offramp, I have to try it against my program ;) |
|Jun-09-05|| ||acdc: <<offramp>: I have noticed that if you don't castle it confuses computers; they don't seem to know what to do.
If you notice, Kramnik plays the fairly late 21.0-0-0, and only then does Junior play 21...a5 and 22...b5. The game is over by then!>|
Interesting anti-computer idea. I will try that too. What do you think the reason for this oddity is? Does the computer get confused because it does not know which side of the board to commit its pieces to?
|Jun-09-05|| ||Medusa: I agree with the idea of avoid castling, see this other game|
Kasparov vs X3D Fritz, 2003
|Jul-12-05|| ||fgh: "Junior Kramnik". Serious pun potential.|
|Jul-12-05|| ||sharpnova: <acdc> i believe that is the correct reason. it is why i try to delay castling as long as possible against computers..|
|Aug-29-05|| ||SnoopDogg: Kingside attacks are very difficult even today to defend for computers. This is a clear example.|
|Sep-15-05|| ||Runemaster: I seem to remember somewhere on this site discussion of a player (not Arno Nickel) who claimed excellent results against computers with the sort of technique we see here from Kramnik. |
Computers seem to find it hard to visualise the consequences of allowing the 'h' file to be opened onto their castled king, even if they gain a piece by doing so. The mating attack is evidently beyond their 'horizon', although perhaps now the strongest computers like Hydra don't have these sorts of weaknesses.
|Apr-13-07|| ||argishti: What a sharp game by kramnik!! he crushed the machine.|
|May-28-09|| ||WhiteRook48: very well played by Kramnik|
|Jun-10-09|| ||SirChrislov: <Runemaster post from '05>:|
I think you're reffering to: Ernest F Pecci
|Jun-10-09|| ||SirChrislov: Pecci, in his book 'A psychiatrist matches wits with fritz' refers to this set-up with 4.f4 as the "barrage position." quite effective.|
|Jun-10-09|| ||Kinghunt: This is just the Stonewall. It worked well enough here, but overall I don't think it's suitable for high level play. I don't think you would ever find this played in say, a world championship match.|
|Aug-04-09|| ||WhiteRook48: 2 e3 was kind of stupid|
|Mar-24-10|| ||birthtimes: Why not play this in a world championship match?! A human would most likely not defend as well as a strong computer...and probably not many grandmasters have read Pecci's book, except Kasparov...|
|Mar-24-10|| ||Alphastar: <birthtimes> because the computer totally misplayed the (late) opening in this game, by castling kingside and then locking the center completely, thus inviting a huge kingside attack by white. In fact I would say that after 10. Rg1, the position is already won for white in a strategic sense.|
Even a master wouldn't make such large positional mistakes in the opening..
And this is ofcourse a typical anti-computer opening.
|Mar-24-10|| ||birthtimes: Yes, this is a typical anti-computer opening, however, if Black, a human, does play the natural and normal book move ...c5, I don't think it would be particularly wise to also castle long, especially if one additionally wants to exchange the pawn on c5 for the d4 pawn.|
Even more critical, is that in this database, there are only 28 games listed from 1893-2008, after 1. d4 d5 2. e3 Nf6 3. Bd3 e6 4. f4!
So this suggests that there just haven't been enough games played by humans to really evaluate how good of a strategy that these anti-computer moves would be when played solely by humans.
Perhaps another Bronstein will soon appear on the horizon, and after reading Dr. Pecci's book, and studying this game by Kramnik and similar ones by Kasparov, will be able to adapt them not only to computer logic, but also to human logic...
|Mar-24-10|| ||birthtimes: Here is another human vs. computer Stonewall game that is played more "humanly" by the computer, and yet eventually White's kingside attack breaks through again to victory...|
Yusupov vs REBEL, 1997
|Mar-25-10|| ||birthtimes: And here's yet another famous game of grandmaster vs. machine in a Bird/Stonewall type setup with wonderful annotations and revealing glimpses into a grandmaster's mind...|
B Alterman vs Deep Fritz, 2000
|Jul-26-11|| ||juan31: Buena onda que un Gran Maestro gane a un programa, quiere decir que a pesar que el programa pueda "pensar" miles de jugadas en segundos, el cerebro de un Maestro sigue siendo superior Gracias Maestro Kramnik|
|Jun-23-17|| ||Nova: Is this Kramnik's famous Stonewall beats computer game??|
If so this should be a GOTD! Pun suggestions? "Stonewall Kramnik"
|Jan-23-18|| ||RookFile: The immortal rookfile game! lol|