< Earlier Kibitzing · PAGE 2 OF 2 ·
|Jul-23-03|| ||Pianzola: El enroque es más sencillo para enfrentar a una computadora que 10.g4
Si se hubiera jugado 23.Qh4 Bf5 hubiera sido bueno para las negras.
32.Rh5 fue el error decisivo. Correspondía 32. Kg6+ Kg7 33. Kf4 intentando repetir jugadas.
35. Cd5 fue mejor que 35. Txh7 Cb3+ que no servía. |
|Nov-07-03|| ||Hittrain: I have a question re: this opening (I'm a beginner). What is the reasoning behind the entire opening and why make the move 6. Qc2? thanks |
|Nov-07-03|| ||Benjamin Lau: It puts pressure on the key h7 and e4 squares. The pressure on the h7 square will become significant once white plays Bd3 and/or black castles short. Putting pressure on e4 is important because it is a key central square. |
|Nov-07-03|| ||Bears092: Also, if black plays dxc4 sometime, a rook would probably me more useful on d1 than a queen. |
|Jan-08-05|| ||lopium: ahajajahaha!! aulero, you just make me laugh. Are you sure Kasparov style is only strong against humans? Or you will say it can be strong against computers... anyway, i'm not sure his style is just impressive. Because if it were true, every humans who can think very deep in chess could beat him.
same for Alekhine style... What about Morphy's one? eh? |
|Jan-13-05|| ||Whitehat1963: Where does 32. Ng6 lead? |
|Jan-13-05|| ||beatgiant: <Whitehat1963: Where does 32. Ng6 lead?>
Maybe the idea is to go for a repetition with 32. Ng6+ Kg7 33. Nf4 Kh8. If Black tries to avoid this with 33...Re8, then 34. Qg2+ Kf8 35. Ng6+ hxg6 36. Qxg6 looks like it works. |
|Jan-13-05|| ||Whitehat1963: <beatgiant> I like your line, but what about 32. Ng6+, Kg7 33. Ne7? Obviously, black can't play 33...Nxe7?? (34. Qxh7++). And if 33...Qxe7, 34. Qxf5, keeping the one-pawn advantage for black. But does black have something better in response to 32. Ng6+, Kg7 33. Ne7? |
|Jan-13-05|| ||Whitehat1963: Then again, I just realized, that like in the game, 32. Ng6+, Kg7 33. Ne7 is best met by 33...Nxe4. |
|Jan-13-05|| ||beatgiant: <Whitehat1963: Then again, I just realized, that like in the game, 32. Ng6+, Kg7 33. Ne7 is best met by 33...Nxe4.>
That would allow 34. Rxh7+ Kf8 35. Ng6+. I think Black should be satisfied with 33...Qxe7 keeping the one-pawn advantage. |
|Jan-15-05|| ||Whitehat1963: What does Crafty think of 32. Ng6+? |
|Jan-15-05|| ||crafty: 32. g6+ g7 33. f4 h8 = (eval 0.00; depth 14 ply; 250M nodes)|
|Aug-31-06|| ||argishti: garry was abit to attackish. he should have just played chess instead of pushing for the win.|
|Jun-05-07|| ||Hot Logic: Well Kramnik has implied in interviews that his secret to beating Kasparov was not being psychologically intimidated by 'the Beast'. |
Now we have seen that Kramnik's chess, which some might say is more 'correct' and others might say (misguidedly) is 'boring'. Doesn't triumph over the computers either.
Yet there are normal grandmasters who specialise in beating computers, something which superGMs aren't able to do at the moment. So computers aren't 'superior' just yet, they still can't 'think'.
Back to Kasparov he is not just all style and no substance, I imagine he got most of his weaknesses squeezed out of him in his matches with Anatoli Karpov. What remained was the greatest world champion in chess so far.
|Jun-05-07|| ||the bad bishop: <Hot Logic> who are the GMs who specialise in beating computers and do a better job than the Super GMs?
I am interested as many people refer to GMs who have systems to "beat" computers.
Who are these people and what are their records against the strong programs (not early 90s machines) the Super GMs play against?
(I should note computer assisted correspondence chess does not fit the bill of beating a computer.)|
|Jun-05-07|| ||JointheArmy: <the bad bishop>
You could start here and here
Ernest F Pecci
|Jun-05-07|| ||the bad bishop: <join the army> Thanks - I have always wondered about these games for several reasons:|
1. Pecci for example shows 20 games in the database - how many losses where there that were not recorded (because they do not serve the purpose of showing a computer can be beaten)? Was it 1, 10 , 100 or 1000?
2. What were the conditions under which the games were played (time controls etc etc)
3. How would these guys go against Rybka or Deep Fritz.
Don't get me wrong, I applaud anyone who can beat a computer - even one of these earlier versions.
But I am skeptical about the concept as it is expressed by <Hot Logic> - "GMs who specialise in beating computers". I am not criticisng <HL>, just expressing a view that there seems to be an impression that there are people out there even today who can "beat" computers with zany openings when guys like Kramnik cannot.
My suspicion (I have no proof - just my good sense and judgment) is that if Pecci sat down with classical time controls against Rybka in a 21 game match he would probably go down 21 - 0 or close to it. He would therefore not, in my book, be a "computer killer".
|Jun-07-07|| ||JointheArmy: <the bad bishop> Well if your looking for modern games on sufficient hardware.|
S Gramlich vs Fritz, 2006
|Jun-08-07|| ||the bad bishop: <JTA> thanks again. |
Are you saying that he is in the class of <normal grandmasters who specialise in beating computers> as <Hot Logic> expressed it. Obviously he is not a GM, but does he <specialise in beating computers>.
I take no issue with the fact that computers CAN be beaten - my query relates solely to whether there are people who specialise in doing so (which I would think means they have a record better than say a 25% success rate) and have a better chance than a Super GM.
<JTA> I welcome your thoughts on that specific proposition.
|Jun-08-07|| ||JointheArmy: <the bad bishop> No I'm not saying that. Though it doesn't mean I don't agree with it. I just don't know enough information to comment.|
However, I think its possible that such people exist. On ICC, they have GM's that beat computers at bullet, blitz and that's all they seem to play. I don't know if they'd have a better chance than a Super GM, but there hasn't been enough computer vs. Super GM matches for me to conclude accurately. Hopefully they wouldn't all turn out like Hydra vs. Adams.
|Jun-18-07|| ||the bad bishop: <JTA> My feeling is that at classical controls there is no one who can regularly beat the best computers. As such, there are no "specialists" who can outperform the Super GMs. Just my view but the power of computers at classical controls can now easily dismantle "anti computer systems".|
At the blitz and bullet controls you referred to I think the equation changes somewhat. As computers don't "think" and rely on brute calculation to get the "answer" - you will naturally have a better chance of beating them if you limit their processing time. It is here that I think "anti computer" systems still probably have some chance - the idea being move the game to a position where the computer is left to do the maximum number of calculations (ie. consider the maximum number of positions) and it may then suffer. I guess some programmers out there could add their 2 cents here.
But, over time, both the increase in processing speed and the "smarter" programming of these computers to only consider the "best" lines will, I fear, lead to these shorter controls providing no sanctuary for any player - be they a Super GM or <normal grandmasters who specialise in beating computers>.
Its interesting and sad in a way.
|Aug-18-07|| ||Paulo Fernando: they both play very well|
|Aug-19-07|| ||contra: Not a good game by Gazza. One pawn is all it takes for a computer to win. Deep Junior's Qb8-d6 manoeuvre was splendid.|
|Sep-07-09|| ||Garech: <JTA> and <the bad bishop> - interesting discussion. I believe there must be GM's out there who no specialise in beating engines, although I do not know of any specifically Nakamura certainly comes to mind; he has beaten many engines with classic anti-computer strategy; keeping the game closed, exchange sacrifices etc. What's brilliantly amusing about those games too is that because the engines refuse to resign he often mates with six or seven minor pieces promoted from pawns just to humiliate the machine. |
As for "zany openings" with regard to beating computers - the strategy there is to go out of book early, which gives yourself better chances as engines have huge opening databases. Also aiming towards a closed game is advantageous and this may well dictate move choice in the opening against computers and make the openings seem unusual or "zany" - but they have a very logical strategy in mind. (It should also be strongly noted that even the best engines are often crushed by Super GMs when they do not have an openings book based solely on human research - not to mention endgame tablebases).
<Just my view but the power of computers at classical controls can now easily dismantle "anti computer systems">
Are you sure about this? Have all of the engine weaknesses been rectified, e.g. the fifty move rule weakness that Nakamura expoited to defeat Rybka?
I agree with you that there is nobody out there who can regularly beat the best engines. However, lets not forget that with perfect play chess is always a draw. With thier massive calculating abilities engines make a very small amount of mistakes (I daren't even call them blunders) so the fact that they are ever beaten at all is something in itself. I think there are many GMs out there who could force a draw in almost every game against the best engines out there. In a way that should almost be the idea against the best computers.
|Dec-16-12|| ||wordfunph: "Nothing is new. Humans make mistakes, computers are bad, and their programmers are arrogant."|
- Garry Kasparov (after losing game 3)
< Earlier Kibitzing · PAGE 2 OF 2 ·