Is Chess on the verge of being solved?

The problem with brute force solving is that it will require immense improvement in computer technologies in order a chess engine to be able to go through all possible game variations to the end while spending a reasonable time for that task - say, a time that it can function uninterrupted - several years with good maintainance maybe. After all, if it needs a time to calculate long enough that it is likely to simply crash or wear out at some moment, what to do then?

With the current capabilities of chess engines, solving chess is therefore impossible. There are many researches of the speed that computers' performance develops with time though, such as the Moore's Law and others, and also many researches about the ammount of information that has to be processed (the Shannon Number tends to give a good picture of the scale of that ammount, and those estimations have been going on, complemented and refined since the very beginning of computer chess and they continue now too).

But even the most optimistic of scenarios of computer improvement related to chess solving result in time estimations that are way beyond what's reasonable to rely on. Chess is, fortunately or unfortunately, not going to be solved, at least not in a similar way.

ok, i have been found out.

yes, chess is entirely solveable.

i wrote the manuscript and it is in my desk but fide threatened my life if it was published.

FIDE wouldn't mind chess becoming solved. It's not within human capabilities to use that kind of information anyway, so it won't influence human play.

Solving chess is more of a challenge that makes people dream about, and a stimulation for developing both engines' and humans' play, and a topic to fill up forums with. In short, a carrot on a stick for chess enthusiasts.

To benefit from knowing the "exact" solution of chess, you would need to know all the possible lines. Say, you and your opponent play "perfectly" up to move 32, then your opponent makes a mistake. What next if you don't remember the refutation?

No hate. I have lots of empathy towards those who are inferior to me.

Consider NIM .a game, with unlimited number of "piles" and "sticks"/ yet 

a computer can solve any state instantly. Chess is vastly more difficult, but

finite and until P <> NP is proven  we best not project exponential curves

as the  obvious limit of computation - especially when newer forms of

computing (like quantum or molecular or analog computing or ???? ) may

be developed.

you miss the point. the human mind will not able to solve all chess solutions.

machines may be able to since chess is ultimately a game of geometry in levels of time and thus entirely solvable by a machine.

SO END THIS THREAD OR REMAME IT TO "WHEN WILL CHESS BE SOLVED BY MACHINES ?" 

I have always considered the "argument from huge numbers" to be quite valid. That is, that the number of chess positions is so huge that even if we used 1 atom per position it would take vast masses of material to save all the positions.

But. A big but. Lets assume we have million unique atoms, molecules or whatever entities. Those entities as singles can hold about million bits of information, right? 1 per entity. But what about if we form a bit of information also from doubles of those entities, for example two of them next to each other? And then groups of three and so on. Then, with optimization which is a real pain in the neck, heh, we could arrive to enormous numbers. I fist thought we could get to a 1,000,000! which is a 1,000,000 times 999,999 times 999,998 and so on but I don't think that is the right answer but never the less it's a down right mean chunk of information.

(By the way, isn't dna storing data like this? Nature always shows us the way. ;))

Even if chess was "solved" by computers, it would always be a great game to be played by humans. So who cares?

sorry,

i solved it...

i have put the answer in a deli sandwhich at 42 eldgrige street, new york city...

the first one to go in and accuse the waiter behind the counter of being a "soup nazi" is is the one who gets the sandwhich ( sorry, it is a ham sandwich )...

...SEINFIELD...

Poor attempt at being funny in my opinion. Care to talk about the subject?

actually I read somewhere that with the current rate at which processor calculation speed is increasing, chess will be solved within a century or so. then they will do go. I think I read this in a New York Times article.

Moores law will break down.  It's not perpetually sustainable.

There was a time that checkers seemed out of reach, and computers solved it several years ago. It took dozens of computers seventeen years: http://www.sciencemag.org/content/317/5844/1518.abstract

It's an entirely different scale of problem though.

Moores law has already broken down - the processing speed of modern processors is capped and is so so that we get the optimal waveform without generating excess heat. Now, we just pile on more processors and hope that works.

 Chess will eventually be solved, as with any zero-sum game.

Indeed, but in computing technology a century is almost unimaginable. Pruning algorithms may reduce the work.

I don't think it will without a fundamental paradigm shift in computing like a breakthrough in quantum computing.  Even then, only weakly solved.

You can't solve chess by pruning.  It needs to be exhaustive by definition.

To put the difference in scale between checkers and chess in perspective, if you had computers fast enough to solve checkers in 1 second instead of 17 years, and if you had been trying to solve chess at that speed since the big bang occurred, you would still not have made any significant progress in evaluating all the positions that would have to be evaluated to solve chess perfectly.

I do believe though that there will come a time within the next few decades when humans will have almost nothing to contribute to chess relative to computers, when centaurs will be beaten 100% of the time by equal strength computers without a human component, and any contribution of the best human-playing chess player in the world would just serve to impair the computer's performance. That will be a sad day, but barring some kind of a civilizational disaster that knocks us back to the middle ages or prevents us continuing to make progress, it will certainly happen if people keep spending time and resources to improve the abilities of chess playing programs.

When you compare Deep Blue to the average mobile device running Shredder, it becomes very difficult to comprehend the limits of computing 25 years from now. That chess might be weakly solved a bit over a century from now is not unimaginable. But to think that we are anywhere close to eight or nine piece tablebases in the next decade simple fails to grasp the scale of the problem.