Will Chess ever be "solved"?

I love how people just throw out definitive statement yanked out of some dark hole. First, to solve for win, you only need to find one example, not search the totality of all possible moves. One line, that’s it. To prove draw you need to exhaust all possibilities, which seems hopeless, but who really knows. Maybe one day the computer power will exist. Maybe. Doubt it, but maybe. That’s a lot of possible moves. We’re talking running all the computers in the world running until the heat death of the universe would just scratch the surface.

"Piú che 'l doppiar de li scacchi s'inmilla"

Dante, Paradise XXVIII, 93

More than the doubling of chess it enthousands itself.

That was more than 700 years ago and talking about the number of angels in paradise.

The game is already solved when there are 7 or less pieces. No matter where the pieces are the computer knows the outcome. Search up tablebases. 

Here is a paper that shows how checkers is proved to be a draw

https://www.researchgate.net/publication/231216842_Checkers_Is_Solved

The same method is applicable to chess once computers and memory have advanced more, likely before the end of this century.

 

That's a big no. As said in the paper you linked, checkers have roughly 500*10^20 possible position. This number is ridiculously small number compared to chess complexity (See https://en.wikipedia.org/wiki/Shannon_number , the game tree complexity is about 10^120). Since there is only ~10^80 atom in the universe, there is no way a classical computer can ever completely solve chess.

There is a small possibility that quantum computer solves chess, but to date it seems hard to make  such an algorithm even assuming that we build powerful enough quantum computers in the future.

I cant tell if you are joking, because of how wrong that statement is.

 

10^120 is not the number of positions, but move sequences, i.e. games. The number of possible positions is much smaller. 10^43. Meanwhile, the visible universe has grown, increasing the number of atoms.

Even so, it remains true that with current data storage capabilities, if computers could solve chess, they would not be able to save the solution.

It just depends

Yes.

Hi I am new here.

When checkers was solved, professional checkers tournaments took a huge hit and they still haven't recovered. 

Even though professional checkers starts with 4 premoves to keep things interesting, kind of the same idea as chess960, and that style of play has not been solved nor can it be solved in the foreseeable future, it still pretty much killed checkers.  

It isn't just a line. The bots would have to search literally all the possible combination of moves to all their responses until the end of the game. Also, we know a chess game can last for a very long time - according to Wikipedia the longest game lasted up to almost 270 moves. But indeed, I agree that people are throwing statements as they are completely truthful out of nowhere. Many specialists have commented about this possibility, and I'd say the outcome of this thread is really different.

Is everyone here a chess specialist and a supercomputer genius? I don't get it. In the words of many specialists themselves, chess won't be solved in a long time, if ever. I don't know why everyone is deliberately affirming that chess is going to be solved. What the hell does "solving" even mean? Finding a line that wins or draws the game with perfect play? That wouldn't even have such a huge impact, in my opinion. I mean, engines already draw and beat themselves so many times, so what would even be the point?

@immaculate if you happen upon one winning line for white, the number of alternate that you have to double check is made significantly more manageable.

As for Optimissed theory that there can be some kind of mathematical proof instead of checking all lines, this is 100% speculation. There is no reason to believe that such a proof exists, and even less of a reason to believe that some genius mathematician would spend his or her life trying to solve something so pointless in order to put an end to dumpster fire threads that everyone rolls their eyes at.

Chess has been solved for all six piece endings. Eventually it will be solved entirely but the solution will be very very large.

     Chess theoretically CAN be "solved". It's just a matter of checking every possible first move by white against every possible reply by black through every possible variation. This is impossible for a human to accomplish. It will take several giant advances in computer technology to make it possible, and even then it's unlikely that anyone would wish to spend the necessary time.

     Even should the solution be found it wouldn't affect human play. Very few people would be able to memorize all the variations needed to force any win that was found, and all an opponent need do is make a few second-best moves to take the game out of the prepared lines.

Derek, you clearly haven’t read through the thread and don’t understand the practical mathematical impossibility of solving for that many positions given our current understanding of how one would create such a database... and as of 2018 we have a tablebase for seven. Six requires 3.7 trillion positions. Seven requires 424 trillion. Eight would require 38 quadrillion and more than 5 petabytes of storage... for just eight pieces. Adding a few more pieces and we don’t even have words for the numbers or the number of bytes. To say with certainty that it will be solved is silly.

No it is silly to say that it will never be done. What practical mathematical impossibility would exist in the future?

It’s not a simple matter of doubling... we’re talking jumps in orders of magnitude. Very soon we run out of atoms on the planet on which to store the data. Soon after that we run out of atoms in the Milky Way. Eventually we run out of atoms in the visible universe on which to store data. Even if we could make the database, how on earth do you do a search? The numbers are so big that they are beyond comprehension.

And if you say Moore’s Law or quantum computers, you don’t understand just how huge these numbers are. Getting to 8 would be a tremendous accomplishment. Getting to 10 is barely comprehensible, and we would still have 22 pieces to go, each step with more and more mind-bogglingly large increases. Your faith that it will be solved eventually is naive.

People, to solve chess it is not necessary to look at all possibilities.
If you can prove that the Grünfeld is a draw, then you do not have to prove that the Queen's Gambit is a draw as well.
If you can prove the Berlin is a draw, then you do not have to prove that the Petrov or the Sicilian are draws as well.
We now have 7 men table bases. For a full proof it is not necessary to have 32 men table bases. as with checkers it is possible to calculate forward towards a table base with a sufficient number of men.
We get +1 man in table bases every 10 years. By the end of this centrury it may be possible to solve chess. To get an idea: 80 years ago in 1940 the first computer as used in the Manhattan project consisted of vacuum tubes and had low speed and low memory capacity. We know where we stand now with speed and memory. 80 years from now speed and memory may suffice to solve chess.