Chess - Play & Learn


FREE - In Google Play

FREE - in Win Phone Store


Will computers ever solve chess?

  • #3621

    @elroch :

    point 4 : where comes the info from from ? and who says that better energy sources will never be available if the current ones are not sufficient ? E = mc2 and there is a lot of m ... 

    point 5: who says quantum computers are the best possible computers even in the future ?


    Personally I think that it would be possible to proof that the brute force way of calculating chess is possible or not, given the number of possibilities in chess compared to theoretical storage possibility of information. I am very unsure about "smart" calculation ...

  • #3622

    Thanks @Elroch, concise and to the point.

    Now a sensible person might disregard the other 3619 posts, and this silly thread might be washed away -- until the next breathtaking advance is announced by the Tech Prophets. 


    Nobel Laureate Herbert Simon predicted computers would beat the World Chess Champion  in about 10 years.  Instead , it took 40 years.

     A "solution" might take 400 years, or never.  happy.png




    This thread is testament to the fact that we can IMAGINE a solution to chess.

    No more or less.  Just live with it.


  • #3623

    I don't think a chess 'solution' needs a 'stupendous amount of energy'. A chess game is basically over after about 30 moves. Someone makes an error and they lose. There is no coming back. There are more moves but the person who made the error slowly or quickly loses. It's complicated for the first 12 - 30  moves, then gets less complicated as the pieces are removed.  People try to make it a battle that has infinite solutions, but it's not. It has a limited number of solutions.  

  • #3624

    It astounds and baffles me that people think that if a "mistake" is made, that the resulting positions no longer need to be looked at for the endeavor of "solving chess".


    I don't think a forced win has been found for queen odds.  That could be a bit dated,


    Really, nothing should be discarded.  Essentially, all positions should be solved, not just the starting position.

  • #3625
    Elroch wrote:

    FAQ for Will computers ever solve chess?


    4. Is there any chance that computers will ever solve chess?

    The laws of physics (uncertainty principle relating time resolution and energy and information theory giving a finite amount of information per bit at a given temperature) mean that any conventional program that could solve chess would have to use a stupendous amount of energy for a very long time (more than a planet can provide). So, the answer "no" is justified.


    Well stated, and I pretty much agree with most of it.

    Take note that on item 4 however you correctly qualify the statement by saying "conventional" programs. This is an important qualifier because the number of unconventional solving algorithms outnumbers the number of chess games. In other words, the scope of solving strategies outnumbers the problem. Btw, for people interested in this topic, it's also interesting to note that this same question is discussed at math forums, and the possibility of solving chess within a few decades has not been ruled out.🤠

  • #3626

    My wording was not good. I meant program running on a conventional computer.

    After I wrote this, I realised there was another point that needed to be made. The physical reason that conventional computers require a certain amount of energy per operation is that they erase information. The erasure of one bit requires a minimum amount of energy which is proportional to the temperature.

    However, there is a class of (mainly theoretical) computers that use completely reversible operations studied by Bennet (and showing von Neumann to be wrong!). These computers have no minimum energy per operation. However, the question is whether the storage requirements of such a computer used to solve chess could be kept under control. Eg see Bennet and Landauer's SciAm article, The fundamental physical limits of computation and other papers like Ultimate physical limits to computation by Seth Lloyd.


  • #3627

    There is no need to do the same thing over and over and over again until chess is solved.  It is possible to have different approaches to solving chess and not have the methods interfere with one another.


    I met a homeless man while I was attending UNT.  His name is Gregory Scott Holcomb.  He said a few things to me that I have remembered throughout the decades.  One of them was:  "There are 3 types of people.  There are those who learn from the mistakes of others, others who learn from their own mistakes, and then there are those who never learn."


    If the mistakes are tossed out, then there is nothing left to learn.

  • #3628
    vickalan wrote:

    It's also interesting to note that this same question is discussed at math forums, and the possibility of solving chess within a few decades has not been ruled out.🤠

    They probably haven't ruled out P=NP either, which would make this problem easier.


Online Now