Man vs chess engines,the endless issue....

[Uri]

Honestly I think top GMs are still much better than Deep Rybka 3 which is currently the strongest program. Our software has limitations. It's not perfect. That's why there is room for new improvements.

I say this because computers play weak in the openings without their opening book. Computers don't understand openings. They play the opening entirely from memory but without 'understanding' it. Even in the endgame computers sometimes make weak moves.

Computers also underestimate the factor of mobility and space advantage on the board. This is an area which is still very difficult for computers. How to place the pawns correctly as to give the pieces optimal mobility and how to correctly manouver in closed positions is also very difficult for computers because it involves deep strategic factors.

Also computers are not parallel enough to match the human mind extreme complexity and parallelism.

[mhull]

Honestly I think top GMs are still much better than Deep Rybka 3 which is currently the strongest program. Our software has limitations. It's not perfect. That's why there is room for new improvements.

I say this because computers play weak in the openings without their opening book. Computers don't understand openings. They play the opening entirely from memory but without 'understanding' it. Even in the endgame computers sometimes make weak moves.

Computers also underestimate the factor of mobility and space advantage on the board. This is an area which is still very difficult for computers. How to place the pawns correctly as to give the pieces optimal mobility and how to correctly manouver in closed positions is also very difficult for computers because it involves deep strategic factors.

Also computers are not parallel enough to match the human mind extreme complexity and parallelism.

But computer stamina is more than enough to overcome these human advantages. Humans can no longer win a match against strong programs.

[ArmyBridge]

Honestly I think top GMs are still much better than Deep Rybka 3 which is currently the strongest program. Our software has limitations. It's not perfect. That's why there is room for new improvements.

Of course Rybka 3 have several weakness but a Human GM has too much weakness than Rybka or Naum, Fritz Shredder etc etc. A GM would be tired, boring, sad, and get nervous, when reach a tactical position sweats and get stress, the program only search eval and play that´s all

I say this because computers play weak in the openings without their opening book. Computers don't understand openings. They play the opening entirely from memory but without 'understanding' it. Even in the endgame computers sometimes make weak moves.

After after hundreds of Chess develop a GM would play opening without any Book, so it can take advantage of these tons knowledge, but is unfair comparison, what do you think that will happen if , for say, Anand play Rybka in Fischer Random chess??...I would not bet a cent for the human to plays better the opening than Rybka

Computers also underestimate the factor of mobility and space advantage on the board. This is an area which is still very difficult for computers. How to place the pawns correctly as to give the pieces optimal mobility and how to correctly manouver in closed positions is also very difficult for computers because it involves deep strategic factors.

wow are you sure about that????? could you please give some example of this?, Guy ,have you replay trough some Junior games? or Thinker plays? if it is not active play I don´t know what could be

Also computers are not parallel enough to match the human mind extreme complexity and parallelism.

Sorry but the only thing that the program must do is to calculate and evaluate, is not necessary to solve mathematical problems or philosophical problems to beat to any GM
New years Regards!!!

[Uri Blass]

Honestly I think top GMs are still much better than Deep Rybka 3 which is currently the strongest program. Our software has limitations. It's not perfect. That's why there is room for new improvements.

The fact that our software is not perfect is known but the same is correct for top GM's.

They are not perfect and have limitations so this is not a reason to think that GM's are better than rybka.

I say this because computers play weak in the openings without their opening book. Computers don't understand openings. They play the opening entirely from memory but without 'understanding' it. Even in the endgame computers sometimes make weak moves.

Top humans are also going to play weak in an opening that they did not memorize.

part of their advantage relative to other humans is simply that they memorize the opening better.

What is understanding of the opening?
humans know that moves are good thanks to calculations of lines
that human did in the past(part of it is with the help of computers).

If we talk about endgames then top humans also sometimes play weak moves in the endgame so again no superiority for humans.

Computers also underestimate the factor of mobility and space advantage on the board. This is an area which is still very difficult for computers.

This is nonsense.
If computers underestimate mobility then programmers can fix the problem easily by increasing the mobility score.

computers sometimes evaluate wrong mobility for both sides(underestimate or overestimate) but the same can be correct also for top humans and I see no proof that top humans are better.

How to place the pawns correctly as to give the pieces optimal mobility and how to correctly manouver in closed positions is also very difficult for computers because it involves deep strategic factors.

everything is basically tactics if you search deep enough and again I know no proof that top humans are better.

Also computers are not parallel enough to match the human mind extreme complexity and parallelism.

I think that it is the opposite and the human mind is not complex enough to match the extreme complexity of the computer.

A computer can memorize hundreds of millions of chess positions in the hash table and the human mind cannot do it and it is not a question of speed because even if you give humans infinite time they have a problem of having a finite memory that is clearly smaller than the computer's memory.

Uri

[Uri Blass]

I think that the effect of faster hardware is overrated when it comes to how well a top engine performs against humans. After all, engines are vastly better than humans in tactics either way, and in the rare situations where humans prevail it is because of strategy and not tactics. Would Kasparov have noticed a difference between a 12 and 16 ply searcher? On the other hand, would he have noticed the difference between Rybka 1.0's eval and Rybka 2.3.2a's eval?

computers play better positional moves when they get bigger depth so I think that kasparov could notice a big difference between 12 and 16 ply search of the same program.

Uri

[Dr.Wael Deeb]

Even though I am not a professional chess player,I have tens of wins against chess engines rated around 2400 Elo and below and a lot of draws against chess engines rated over 2600 elo,but....never the less,I think that the computer had already dominated the chess scene and there is nothing we can do about it....

[George Tsavdaris]

It depends. Some GMs know how to exploit computer weaknesses better than others and so are better against computers.

The long time control results of the last years, and not your opinion, show the opposite of what you are saying.
I see only computer wins or draws. If you can show me a win by a human.....

It's really difficult to say who is better.

It's really easy to say: Computers!
In Chess.

If we speak about creativity, positional move and strategic play then humans by far.
But Chess unfortunately for humans is less about creativity and strategic play and it's more about not making small(or bigger) tactical mistakes here and there and being consistent ALL the time.

[George Tsavdaris]

Honestly I think top GMs are still much better than Deep Rybka 3 which is currently the strongest program.

Honestly i think this is totally wrong.
Results show this. IF you don't _want to accept them and stick to your subjective opinion then no problem.

And i'm speaking about playing Chess. Not about creativity, making positional moves, strategic play etc, where humans are better by far.

I'm speaking about playing Chess which includes tactical play where humans pale beside computers. And practical Chess as it seems is much more about not making tactical mistakes and being strong tactically, than having exceptional strategic plans.

Honestly I think top GMs are still much better than Deep Rybka 3 which is currently the strongest program.
Our software has limitations. It's not perfect. That's why there is room for new improvements.

Correct but completely irrelevant to the first paragraph.

I say this because computers play weak in the openings without their opening book. Computers don't understand openings.

This is a huge misunderstanding of computers Chess.

•Firstly computer Chess is NOT engine Chess which is the program only. It's the engine plus the opening book plus the endgame tablebases.

•Secondly programmers from the start of computer programming found a specific and very effective way to handle opening. That is not to try to make the engines to understand opening, but to put a data table with opening statistics(form humans or computers) for the computer to read it and play according to it.
It worked and it become the way of handling opening in the computer programs.
There is no reason to change it and accuse engines that don't understand openings. This is how they have been designed!

So you can't accuse computers that can't understand openings as computers as i've said are the engines plus the opening books plus the EGTBs. So computers actually play the opening very good.

They play the opening entirely from memory but without 'understanding' it.

You make a basic mistake here by thinking anthropocentrically.
•First: Why computers have to understand anything? They just have to play it.
•And secondly: Why to define "understand" in an anthropocentric way?

Even in the endgame computers sometimes make weak moves.

They make but in the last years this has become very rare.
And of course GMs and top GMs made weak moves too in the endgame, so what you say makes no sense about the comparison of humans-computers.

Computers also underestimate the factor of mobility and space advantage on the board. This is an area which is still very difficult for computers. How to place the pawns correctly as to give the pieces optimal mobility

I don't really understand why you say all these?
I disagree anyway.

and how to correctly manouver in closed positions is also very difficult for computers because it involves deep strategic factors.

This is the only one i agree with you.

[Dr.Wael Deeb]

And as George said and as I've always said,we are yet to witness a positive human result against the current chess engines....yet to prove this human strategical superiority over the chess engines....because when a comp plays the human,I only see shiny floors after it were wiped with human bodies

[mhull]

And as George said and as I've always said,we are yet to witness a positive human result against the current chess engines....yet to prove this human strategical superiority over the chess engines....because when a comp plays the human,I only see shiny floors after it were wiped with human bodies

This is true, because while GMs understand chess better than programs, programs are much more accurate at tactics and never get tired, which has proven to be the more important advantage in the winning of matches.

As you say, even the strongest humans are from this day forward, forever consigned to the lower end of the mop handle.