This has to be a very weak computer NOT the program.........

[pichy]

Computer at blitz should be almost unbeatable, this is the reason why I do NOT consider the high rating of the top 5 programs to be taken too seriously when matched versus a top GM that knows how to block a position:roll:

PS: Unless programs are programmed to sacrifice a pawn or two in order to break through the position a great program like Rybka is still too materialistic vs a GM and at the end it runs out of time and suffer positionally and give up material. Notice how GM Nakamura was down-trading pieces with the program knowing that Rybka was blocked with the pawns and at the end it will not sacrifice a pawn or two to break the position
http://www.chessgames.com/perl/chessgame?gid=1497429

Here are two more matches versus Rybka:
http://www.chessgames.com/perl/chess.pl ... pid2=10084

I have quickly looked over the first game you presented, it has nothing to do with weak hardware or software but everything to do with how to beat these programs.

He knows how.

I have something to add, in the beginning back in the 80's when programs did not have opening or ending databases they were very weak programs, later besides adding databases, more knowledge was added and chess programs became stronger and stronger. A program that does not have the knowledge of sacrificing a pawn or two specially when it is ahead in order to break through a fortress of pawns barrier it is still a weak program when matched against a strong human that knows how to block a position. Therefore, Rybka is still a weak chess program.

PS: I should NOT say that Rybka is a weak chess program, it is the strongest program available, but most strong human GMs have more chess knowledge than rybka, and blocking and being able to unblock a position any strong human GM or even a master is capable of doing it when they are ahead in material.

No problem, put the contempt at 3.00 and Rybka will be going to crush you.

Kai

Yes, there is a problem, rybka does not have as much chess knowledge as GM Nakamura nor Anand. It is NOT just with a block position take a look at this game.

http://www.chessgames.com/perl/chessgame?gid=1497426

[pichy]

Computer at blitz should be almost unbeatable, this is the reason why I do NOT consider the high rating of the top 5 programs to be taken too seriously when matched versus a top GM that knows how to block a position:roll:

PS: Unless programs are programmed to sacrifice a pawn or two in order to break through the position a great program like Rybka is still too materialistic vs a GM and at the end it runs out of time and suffer positionally and give up material. Notice how GM Nakamura was down-trading pieces with the program knowing that Rybka was blocked with the pawns and at the end it will not sacrifice a pawn or two to break the position
http://www.chessgames.com/perl/chessgame?gid=1497429

Here are two more matches versus Rybka:
http://www.chessgames.com/perl/chess.pl ... pid2=10084

I have quickly looked over the first game you presented, it has nothing to do with weak hardware or software but everything to do with how to beat these programs.

He knows how.

I have something to add, in the beginning back in the 80's when programs did not have opening or ending databases they were very weak programs, later besides adding databases, more knowledge was added and chess programs became stronger and stronger. A program that does not have the knowledge of sacrificing a pawn or two specially when it is ahead in order to break through a fortress of pawns barrier it is still a weak program when matched against a strong human that knows how to block a position. Therefore, Rybka is still a weak chess program.

PS: I should NOT say that Rybka is a weak chess program, it is the strongest program available, but most strong human GMs have more chess knowledge than rybka, and blocking and being able to unblock a position any strong human GM or even a master is capable of doing it when they are ahead in material.

No problem, put the contempt at 3.00 and Rybka will be going to crush you.

Kai

Yes, there is a problem, rybka does not have as much chess knowledge as GM Nakamura nor Anand. It is NOT just with a block position take a look at this game.

http://www.chessgames.com/perl/chessgame?gid=1497426

[pichy]

Computer at blitz should be almost unbeatable, this is the reason why I do NOT consider the high rating of the top 5 programs to be taken too seriously when matched versus a top GM that knows how to block a position:roll:

PS: Unless programs are programmed to sacrifice a pawn or two in order to break through the position a great program like Rybka is still too materialistic vs a GM and at the end it runs out of time and suffer positionally and give up material. Notice how GM Nakamura was down-trading pieces with the program knowing that Rybka was blocked with the pawns and at the end it will not sacrifice a pawn or two to break the position
http://www.chessgames.com/perl/chessgame?gid=1497429

Here are two more matches versus Rybka:
http://www.chessgames.com/perl/chess.pl ... pid2=10084

I have quickly looked over the first game you presented, it has nothing to do with weak hardware or software but everything to do with how to beat these programs.

He knows how.

I have something to add, in the beginning back in the 80's when programs did not have opening or ending databases they were very weak programs, later besides adding databases, more knowledge was added and chess programs became stronger and stronger. A program that does not have the knowledge of sacrificing a pawn or two specially when it is ahead in order to break through a fortress of pawns barrier it is still a weak program when matched against a strong human that knows how to block a position. Therefore, Rybka is still a weak chess program.

PS: I should NOT say that Rybka is a weak chess program, it is the strongest program available, but most strong human GMs have more chess knowledge than rybka, and blocking and being able to unblock a position any strong human GM or even a master is capable of doing it when they are ahead in material.

No problem, put the contempt at 3.00 and Rybka will be going to crush you.

Kai

Yes, there is a problem, rybka does not have as much chess knowledge as GM Nakamura nor Anand. It is NOT just with a block position take a look at this game.

http://www.chessgames.com/perl/chessgame?gid=1497426

[Uri Blass]

Computer at blitz should be almost unbeatable, but GM Nakamura got lucky or this has to be a very old and weak P.C. with Rybka

http://www.chessgames.com/perl/chessgame?gid=1497429

This game is not against rybka3 because it is from 15.3.2008

Rybka3 can find 182.Ka3 that is enough to draw and the opening 1.g4 is not good for white

Edit:182.Ka3 probably also lose the game and I will check the losing mistake that may be earlier and if rybka3 can avoid it easily.

The other games that they give are from 2007.

[Uri Blass]

Computer at blitz should be almost unbeatable, this is the reason why I do NOT consider the high rating of the top 5 programs to be taken too seriously when matched versus a top GM that knows how to block a position:roll:

PS: Unless programs are programmed to sacrifice a pawn or two in order to break through the position a great program like Rybka is still too materialistic vs a GM and at the end it runs out of time and suffer positionally and give up material. Notice how GM Nakamura was down-trading pieces with the program knowing that Rybka was blocked with the pawns and at the end it will not sacrifice a pawn or two to break the position
http://www.chessgames.com/perl/chessgame?gid=1497429

Here are two more matches versus Rybka:
http://www.chessgames.com/perl/chess.pl ... pid2=10084

I have quickly looked over the first game you presented, it has nothing to do with weak hardware or software but everything to do with how to beat these programs.

He knows how.

It seemed that you did not look carefully in the first game.

1)1.g4 d5 2.h3 is not good for white and it is not a good way to play for a win and rybka got a drawn position.

2)182.Re1 is clearly a losing move that better hardware or software can avoid.

182.Ka3 is a drawing move

[d]8/4bp2/p1p3p1/Pp1n1kPp/1PpPpP1P/4P2R/1K6/1R6 w - - 0 182


Edit:I did not look carefully and the losing mistake of white was earlier(182.Ka3 lead to repetition but black can win material and probably the game but I believe that good software or hardware can avoid the losing mistake and I will give more analysis)

Edit:It seems that
181.Rg3 could be enough to draw when rybka played the blunder 181.Kb2

The reason for my earlier mistake is that white could force repetition of the position after 181.Kb2 Be7 so I assumed that the draw evaluation for rybka for 182.Ka3 is correct.

[d]8/5p2/p1pb2p1/Pp1n1kPp/1PpPpP1P/K3P2R/8/1R6 w - - 0 181

[Terry McCracken]

Computer at blitz should be almost unbeatable, this is the reason why I do NOT consider the high rating of the top 5 programs to be taken too seriously when matched versus a top GM that knows how to block a position:roll:

PS: Unless programs are programmed to sacrifice a pawn or two in order to break through the position a great program like Rybka is still too materialistic vs a GM and at the end it runs out of time and suffer positionally and give up material. Notice how GM Nakamura was down-trading pieces with the program knowing that Rybka was blocked with the pawns and at the end it will not sacrifice a pawn or two to break the position
http://www.chessgames.com/perl/chessgame?gid=1497429

Here are two more matches versus Rybka:
http://www.chessgames.com/perl/chess.pl ... pid2=10084

I have quickly looked over the first game you presented, it has nothing to do with weak hardware or software but everything to do with how to beat these programs.

He knows how.

It seemed that you did not look carefully in the first game.

1)1.g4 d5 2.h3 is not good for white and it is not a good way to play for a win and rybka got a drawn position.

2)182.Re1 is clearly a losing move that better hardware or software can avoid.

182.Ka3 is a drawing move

[d]8/4bp2/p1p3p1/Pp1n1kPp/1PpPpP1P/4P2R/1K6/1R6 w - - 0 182


Edit:I did not look carefully and the losing mistake of white was earlier(182.Ka3 lead to repetition but black can win material and probably the game but I believe that good software or hardware can avoid the losing mistake and I will give more analysis)

Edit:It seems that
181.Rg3 could be enough to draw when rybka played the blunder 181.Kb2

The reason for my earlier mistake is that white could force repetition of the position after 181.Kb2 Be7 so I assumed that the draw evaluation for rybka for 182.Ka3 is correct.

[d]8/5p2/p1pb2p1/Pp1n1kPp/1PpPpP1P/K3P2R/8/1R6 w - - 0 181

They were bullet games..that has to be taken into account.

[Uri Blass]

considering the fact that the games are bullet games mean that with more time the computer could find a better move and it also means that with better hardware or better software the computer could find a better move.

The mistake of rybka at move 181 is not something that you need a deep search to avoid and slightly bigger depth thanks to better hardware can certainly help.

[BubbaTough]

Man....are people seriously complaining that the computer did not have enough time to defend against the human? I think its pretty well established that adding more time helps the human more than the computer. Yes Yes, Rybka 8 running on a 1024 core cluster would play some different moves....but the bottom line is that while that American GM may have wasted his time messing around on the internet instead of serious study and serious competition...he is one bad dude when playing against a computer.

-Sam

[AdminX]

Man....are people seriously complaining that the computer did not have enough time to defend against the human?
-Sam

[Uri Blass]

<snipped>

Man....are people seriously complaining that the computer did not have enough time to defend against the human? I think its pretty well established that adding more time helps the human more than the computer.

It was correct in the past but chess programs improved and I am not sure if it is correct today.

Humans today may have almost no chance against the computer by thinking and the best chance of them may be by memorizing some lines that work against the computer(the lines do not have to be a full game and first 20 moves to get a position that they know to defend against the computer may be enough).

It is easier to collect more data to predict the computer moves at blitz and not at long time control(for the simple reason that you can get in the same time more computer games at blitz)

It means that humans who use part of their training to memorize opening
lines that they can use against the computer may score better against computers at blitz and not at long time control(when the computer can play a move that they are not prepared for it because they prepared only for the move that the computer plays at blitz).

Uri