Upgrade to Chess.com Premium!

Is Chess on the verge of being solved?


  • 3 years ago · Quote · #61

    fburton

    madhacker wrote:

    Also, and I've raised this before, "solving" chess wouldn't make the computer play a great game. Far from it. Think of this: the computer is about to play its first move. Should it play 1. e4 or 1. a4? Which is better? After all, it knows that they both lead to a draw. Why should one be better than the other? It might play 1. a4.

    1.e4 and 1.a4 will both lead to a (huge) variety of wins, losses and draws, depending on how the opponent plays. The 'better' move may then be the one that offers more opportunities for winning, or perhaps more opportunities for the opponent to make mistakes.

    madhacker also wrote:

    A computer which solves the game would play rubbish moves until its position was bad but not beyond saving on best play, they play perfect moves to save it. In other words, it would never lose, but probably wouldn't win too many either!

    In what sense would the moves be rubbish if it could eventually save the game? The position would only look iffy in terms of our naive evaluation functions.
    However, as you say, this is very hypothetical!
  • 3 years ago · Quote · #62

    erixoltan

    Chess will be solved when someone invents the quantum computer, but not before that. A quantum computer would be able to crunch through the entire combinatorial explosion of chess to determine whether a position is a win, loss or draw.  Fortunately there are tremendous technical obstacles to building such a machine, so the game will retain some mystery for the foreseeable future!

    A computer that has solved the game would not play bad moves in a position where multiple non-losing moves exist and there is no move that wins by force. 

    Think about it.  A programmer smart enough to write code that solves the game of chess will also be so stupid that they can't figure out how to select reasonable moves in a drawn position?  It seems very unlikely to me. Off the top of my head I can think of several algorithms that would fix that issue.

  • 3 years ago · Quote · #63

    Tomba31

    Nah... Chess is one of those games where the possibilities are endless and no one nor computer can play the game and expect to know if its going to win or lose just by looking at a board and its first move.

  • 3 years ago · Quote · #64

    pawnzischeme

    I once solved chess using an abacas, t-square, bubble level, micrometer, hammer, ice tongs, miner's lamp, catcher's mask and a rubik's cube; but I lost my notes on how I did it, and I lost my t-square.

  • 3 years ago · Quote · #65

    GIex

    Ryan390 wrote:
    damongross wrote:

    Not only will chess not be solved in the foreseeable future, this thread will not be completed in the foreseeable future.


    Amen to that..


    Yep. And God help us until then. Laughing

     

    The matter is that it's not up to the computer or the human player's skills. It's up to their time management efficiency. If the programmer has enough time to go through all the variations that the computer does, to apply the evaluation algorithms that the computer applies, and so on, the result would be the same. But since he's not able to do so, he would assign those duties to the computer, and he will concentrate on creating the algorithms' instead - what the computer can't do, no matter how much time it is given. Simply because a computer does nothing it is not designed to do (except for crashing). Which is the reason computers will never be better than humans at chess understanding.

     

    That's why human players concentrate on thinking about principles and strategy - to save time. This way they needn't go through all the possible variations a move will lead them to if they know that the move will lead to a better position. As Fischer put it, "Don't worry about finding the best move. Just try to find a good move." The player's task is not to predict the future, but to make sure what he plays is consistent. That's why chess, in addition to not going to be solved soon, needn't be solved soon, at least not for the purpose of playing it in a better way.

  • 3 years ago · Quote · #66

    pathfinder416

    pawnzischeme wrote:

    I once solved chess using an abacas, t-square, bubble level, micrometer, hammer, ice tongs, miner's lamp, catcher's mask and a rubik's cube; but I lost my notes on how I did it, and I lost my t-square.


    This guy down the street from me had some spare time during a backyard barbeque and showed very elegantly that the micrometer is redundant. Or was it the ice tongs? He didn't write it down.

  • 3 years ago · Quote · #67

    madhacker

    fburton wrote:
    1.e4 and 1.a4 will both lead to a (huge) variety of wins, losses and draws, depending on how the opponent plays. The 'better' move may then be the one that offers more opportunities for winning, or perhaps more opportunities for the opponent to make mistakes.

    I meant they both lead to a draw on best play. The problem with using the "most paths which lead to a win" approach to choosing between different drawn positions, is that the computer would aimlessly exchange pieces, on the basis that almost all replies the opponent can make result in a win for the computer (apart from the obvious one - the recapture). That's why I suggested the lost positions database approach as a different possibility.

    I'm a software developer and I've given this some thought (sad, I know!) I can't think of a better option than the database one. I'm all ears though.

  • 3 years ago · Quote · #68

    Ziryab

    pawnzischeme wrote:

    I once solved chess using an abacas, t-square, bubble level, micrometer, hammer, ice tongs, miner's lamp, catcher's mask and a rubik's cube; but I lost my notes on how I did it, and I lost my t-square.


    I can loan you my t-square.

  • 3 years ago · Quote · #69

    jonnyjupiter

    I have a rubiks cube and a hammer

  • 3 years ago · Quote · #70

    Ubik42

    pathfinder416 wrote:

    But I also programmed on the TRS-80 Model I in the late 1970's, which made me feel like a god.


     Me too, I wonder how many of us are left.

  • 3 years ago · Quote · #71

    helltank

    Chess will never be solved because to solve chess you have to have perfect play.

    And because of the incredible variety of responses to any reasonable opening move by white, each of which has their pros and cons, there is no such thing as perfect play.

    Also, you would first have to resolve the Hypermodern vs Classical battle, which is virtually impossible.

  • 3 years ago · Quote · #72

    GIex

    The matter is that many people expect that if chess can be solved, it will be solved into a singular variation. But proving that is itself not much easier than solving chess.

    I believe there is no single solution to chess, but a variety of solutions. What is common between them is called strategy - those are concepts that need to be implemented so that the opponent won't have straight advantage. And as long as applying strategy works well, you needn't solve chess any further.

  • 3 years ago · Quote · #73

    Ryan390

    Everyone's pretty much repeating everything that's already been said. Laughing

    It's an interesting subject, but I think if it were possible, I wouldn't want it to happen as it would ruin the game. Yell

  • 3 years ago · Quote · #74

    melvinbluestone

    The answer is: Yes, chess is on the verge of being solved. Probably in a few months.......... Right before the November elections..... . . .   .    .   .

  • 3 years ago · Quote · #75

    ivandh

    Everybody is just restating what has been argued before.

  • 3 years ago · Quote · #76

    wishiwonthatone

    I've never said this before.

  • 3 years ago · Quote · #77

    pathfinder416

    Anything worth doing must be worth overdoing.

  • 3 years ago · Quote · #78

    Ubik42

    People keep going around in circles.

  • 3 years ago · Quote · #79

    fburton

    And adding short sentences which don't actually say anything useful to an already worn out topic.

  • 3 years ago · Quote · #80

    fburton

    madhacker wrote:
    fburton wrote:
    1.e4 and 1.a4 will both lead to a (huge) variety of wins, losses and draws, depending on how the opponent plays. The 'better' move may then be the one that offers more opportunities for winning, or perhaps more opportunities for the opponent to make mistakes.

    I meant they both lead to a draw on best play. The problem with using the "most paths which lead to a win" approach to choosing between different drawn positions, is that the computer would aimlessly exchange pieces, on the basis that almost all replies the opponent can make result in a win for the computer (apart from the obvious one - the recapture). That's why I suggested the lost positions database approach as a different possibility.

    I'm a software developer and I've given this some thought (sad, I know!) I can't think of a better option than the database one. I'm all ears though.


    Not sad at all - if nothing else, this stuff is fun to think about. Smile

    "I meant they both lead to a draw on best play."

    Has this been proved? And at what point does a position change from being "drawn on best play" to "lost on best play"? For all we know, 1.e4 might be lost on best play!


Back to Top

Post your reply: