Will computers ever solve chess?

Sort:
fburton
BettorOffSingle wrote:

I've already used engines to prove that chess is a forced draw.  I'm sure others have found the proof as well. 

Has the proof been published anywhere? How did you get over the horizon effect, or does it not matter in your proof?

Elroch
Pawnghost wrote:

Might as well ask if computers have solved billiards. Program a robotic arm to shoot the best angles in pool. People will still prefer to play each other. Chess was also created to be played against our fellow man. Computers can't appreciate the beauty or risk of a well played match. A computer cannot solve why chess captures the mind as it does.

I want to see a tennis playing robot.

Elroch

Go is a much tougher game. But there has been dramatic progress recently. The next step will be a match against the world go champion, Lee Se-dol, starting March 9.

troy7915
s23bog wrote:

Computers will not "solve" the current form of chess until brute force is possible.

  You mean until engines can calculate all the possibilities that exist, like they did with checkers. I don't think that day will ever come, considering how astronomical that number is.

troy7915
s23bog wrote:

What prevents computers from being able to examine every possible position?  What are the limitations of the hardware and/or software?

  If that number is so astronomical that it surpasses the number of atoms in the universe( Kasparov and before him Petrosian has said it), then it truly seems impossible, doesn't it?

RoepStoep
s23bog wrote:

I think it boils down to how much people want to invest in solving something like chess.  IBM poured in millions with Deep Thought, and Deep Blue (et al?).  If trillions were available, a solution would be found.

Read this thread again, the reasons this is not true have been mentioned numerous times. How is more money going to solve the billions of years of calculation times required (hook up all the computers in the world and you still need more time than the earth is supposed to exist), and where are you going to buy more storage room than the estimated amount of atoms in the universe? The problem is time and space, not money.

RoepStoep

Hmm, I don't get what you mean at all, a truce? What would be the point of the game if two players were to call a truce? And I don't see at all what you mean about forgetting the coordinates either, they are just a means of expressing what move or square you're talking about, used equally by computers and humans. Would you like people to point to the knight or say they move the bishop on the far right to that square next to the black queen or something? I don't see how that would change anything, let alone be an improvement...

Re_Kong

Deciding if chess is either solvable or not is a maths matter, not a computer science one.

 

If chess are solvable, sooner or later computers will find an algorithm for white to win every game, if they are not, computers will find an algorithm for black to draw every game.

 

Just think about this: chess is a finite game.

Sedlescombe

I have never understood the angst around computers and chess.  Are the achievements of Usain Bolt diminished by the fact that I can drive my car faster than he can run? 

RoepStoep
Re_Kong wrote:

Deciding if chess is either solvable or not is a maths matter, not a computer science one.

 

If chess are solvable, sooner or later computers will find an algorithm for white to win every game, if they are not, computers will find an algorithm for black to draw every game.

 

Just think about this: chess is a finite game.

The problem is not whether chess is solvable, because as you say it is a finite game so of course it is solvable. The problem is not the algorithm to achieve such a strong solution either, we have that already: look at every possible white move, then look at every possible black reply to each move, and continue until you have exhausted all possible games.

The problem is twofold: the amount of positions is so astronomical that the time to examine them all is unrealistic, and the amount of space required to store the results of this massive calculation is simply impossibly large, if you could store one position per atom, all the atoms of the earth give you only a tiny fraction of the total space required.

Re_Kong
RoepStoep ha scritto:
Re_Kong wrote:

Deciding if chess is either solvable or not is a maths matter, not a computer science one.

 

If chess are solvable, sooner or later computers will find an algorithm for white to win every game, if they are not, computers will find an algorithm for black to draw every game.

 

Just think about this: chess is a finite game.

The problem is not whether chess is solvable, because as you say it is a finite game so of course it is solvable. The problem is not the algorithm to achieve such a strong solution either, we have that already: look at every possible white move, then look at every possible black reply to each move, and continue until you have exhausted all possible games.

The problem is twofold: the amount of positions is so astronomical that the time to examine them all is unrealistic, and the amount of space required to store the results of this massive calculation is simply impossibly large, if you could store one position per atom, all the atoms of the earth give you only a tiny fraction of the total space required.

 By solvable i mean "white wins every game", and that can't be said for sure right now.

 

You brought up a good point with the "storage thing", I hadn't thought about that.

RoepStoep
Re_Kong wrote:

 By solvable i mean "white wins every game", and that can't be said for sure right now.

 

You brought up a good point with the "storage thing", I hadn't thought about that.

 Ok, the game is solvable for sure, but indeed we don't know the result of such a solution until it has been done

Chef-KOdAwAri

it's universally accepted that, once solved, the best line will be a draw.

RoepStoep
Lenudan wrote:

it's universally accepted that, once solved, the best line will be a draw.

 It is (almost) universally assumed that it will be a draw

RoepStoep
s23bog wrote:

Someone asked what I meant by a truce.  Imagine a position with all of the white pieces gathered around the black king, and all of the black pieces gathered around the white king, and neither player in check.

 

Now, how do you play to that with someone else?

 I mainly wondered what the point of a truce would be, as the game is about destorying / checkmating the opponent

troy7915
RoepStoep wrote:
Lenudan wrote:

it's universally accepted that, once solved, the best line will be a draw.

 It is (almost) universally assumed that it will be a draw

  Yes, assumed. An accepted belief.

RoepStoep

It would be funny if chess turned out to be a win for black, with perfect play being the mainline Berlin. Kramnik would go down history as the greatest genius in chess

SmyslovFan

Apart from a few trolls found on a few corners of the internet, almost every chess player accepts that chess is a draw with best play. There may never be an absolute proof that will satisfy internet trolls, but for the rest of us, chess is a draw. That's depressing because it means that our greatest wins are due to errors made by our opponents. 

The upshot of this is that in order for elite players to win, they must encourage their opponents to make mistakes. They play confusing and dangerous lines in order to have better winning chances. But in doing so, they also risk losing. Human chess is far more interesting than computer chess, not least because computer chess is getting so that less than 20% of all computer games are decisive, but also because humans play more interesting variations!

troy7915
SmyslovFan wrote:

Apart from a few trolls found on a few corners of the internet, almost every chess player accepts that chess is a draw with best play. There may never be an absolute proof that will satisfy internet trolls, but for the rest of us, chess is a draw. That's depressing because it means that our greatest wins are due to errors made by our opponents. 

The upshot of this is that in order for elite players to win, they must encourage their opponents to make mistakes. They play confusing and dangerous lines in order to have better winning chances. But in doing so, they also risk losing. Human chess is far more interesting than computer chess, not least because computer chess is getting so that less than 20% of all computer games are decisive, but also because humans play more interesting variations!

  It's not about satisfaction, but logic. 'Draw' is a belief, not a fact. To confuse a belief with a fact is illogical.

SmyslovFan

Troy, the belief is based on all the available evidence, which is quite plentiful. If the belief were based on nothing, the trolls could keep on trolling. Well, even though the belief that chess is a draw is based on substantial evidence, and even though every professional chess player accepts that chess is a draw, the trolls will keep on trolling anyway.