Moore's Law is dead, and Finegold still thinks it will take 1000 years. And yet, Steinitz was the one who said that chess was a draw with best play. The next 42 pages promise to become interesting.
True or false? Chess will never be solved! why?
Sort:
Currently, they are up to 7-man tablebases and that takes up 140,000 GB of space. Imagine what a 32-man tablebase would require? It will eventually be solved as computing power increases, but it's not going to be in anyone's lifetime that is currently alive right now. Likely less than 1000 years though.
Remember that in addition to the storage requirements, the computational requirements also grow exponentially to generage each successively larger tablebase. I think 1000 years is a pipe dream, but then I also think it will not be done in 100,000 years.
Is chess really just a problem to be solved? I thought there were often a variety of move choices and options of styles to play.
I don't think you fully appreciate the magnitude of the problem.
Indeed, all we have to do is IMAGINE that technology will (somehow) solve the problem (sometime) in the future.
That's So EASY. Problem Solved !! 
King_Geehm wrote:
Technology says yes. To put it simply: if it can be broken down into mathematics, it can be solved.
Those who doubt there’s enough storage in the world should consider the fact that it doesn't necessarily all have to be stored in one place. Also consider that as technology advances, the physical size of storage devices are becoming smaller while still growing in data capacity. Then there’s also the factor of compression algorithms becoming more and more efficient at shrinking data.
There are roughly 10^43rd legal positions in the game of chess. There are about 10^50th atoms in this planet. So, it might technically be feasible to store a database able to hold all possible chess positions with best move as long as we reached some magical future tech state where we can store a bit of data in only a few hundred atoms, and we somehow could transmute the planet into pure storage.
Somehow, I am doubting it will ever be "solved" in the purest of senses.
rdecredico wrote:
TheGrobe wrote:
The Knight's Tour is a completely different problem, you can't simply extrapolate from it to the game of chess. Why not just say that since Tic Tac Toe is solved to be a draw that chess must also be?
Chess is not insoluble. The Knights tour proves there are many solutions that are all viable. I did not claim nor state it proved chess was solvable. I claimed the wise learn from it that there are many paths to perfection in chess.
Tables bases.
The key to solving chess.
Ah, trolling in yet another thread... keep up the good work.
King_Geehm wrote:
Or, we could imagine that math and the magnitude of supecomputers dont exist. thats even easier. Technology already IS well on its way to solving the problem, no imagination necessary. I'm sorry that I dont have the time or inclination to break it down to a 3rd grade level for you.
It isn't about electronics and technology... it is about the sheer size of the problem. Not every problem can be solved by brute force.
King_Geehm wrote:
Technology says yes. To put it simply: if it can be broken down into mathematics, it can be solved.
If this were true, your bank account is as good as hacked, as well as many other things that rely on RSA cryptography. The problem is "simply" one of breaking a large number into its prime factors, for which multiple algorithms exist.
King_Geehm wrote:
It’s one thing to fear what you don’t understand but to live in complete denial is just ignorant.
King_Geehm wrote:
Of course you like to throw around numbers but (...)
Yeah. Our belief in Shannon's entropy blinds us to the obvious truth that computer science trends are bound to continue forever. Our belief that mathematical questions have answers expressed in numbers blinds us to the fact only winners with a positive attitude make progress.
falsh . you have 200 $ you need tv and pc , what will you buy , tv , you save mor money or pc you spend all. now you can take pown and knight for an knight but you need to trade queen . chass is great simulated game . but last crushing yor oponent is solved thing.
King_Geehm wrote:
EricFleet wrote:
King_Geehm wrote:
Or, we could imagine that math and the magnitude of supecomputers dont exist. thats even easier. Technology already IS well on its way to solving the problem, no imagination necessary. I'm sorry that I dont have the time or inclination to break it down to a 3rd grade level for you.
It isn't about electronics and technology... it is about the sheer size of the problem. Not every problem can be solved by brute force.
You see limits but top level scientists with knowledge far beyond any that you or I can possibly fathom, do not. Fortunately, they do not share your "stop it, that’s just impossible" attitude. Of course you like to throw around numbers but you forget to:
Please show me a quote from one of these 'top level scientists' who say chess can be solved.
· ` Exclude the number of illegal positions in that number. And then we can subtract the vast number of positions that are clearly losing positions that need no further analysis.
The number already includes legal positions. As far as losing positions, they must be solved too or else chess is not solved.
· ` consider the fact that as of now, a terabyte of data can be stored on a device of only 1x1 inch of size. They’re only getting smaller and compression algorithms are getting more efficient.
There is a limit to compression. Information theory has proven this. And there is a limit to storage density regardless of the technologies invented. Much like the speed of light, there are hard limits of nature.
· ` Accept that the game can be (is) broken down into mathematical equations and coincidentally, computers are increasingly fast as solving such equations. They’re only getting faster
Of course they can be broken down. They already have been. And I've been in professional IT for 20 years and programmed for fun for 35, so I am aware of this. Also be aware that there are fundamental limits to how fast computers can get.
· ` Fail to realize another ~5M games have been collected for analyzing since the time of my last post
And you fail to realize the 5 Million is a near infintesimal amount compared to 10^43. Even if you each person in the world had a supercomputer that calculated and stored the results of 5 Million games each microsecond, the job would not be finished before the death of the universe.
I am fully aware of the power of technology and how fast it has improved. You fail to realize just how big 10^43 is. Now, if someone can figure out a clever way of avoiding turning this into brute force calculation, maybe we'll solve it. Until then we are stuck with algorithms that evaluate positions based on principles rather than an absolute solving of the game.
And is that really so bad? I don't think so.
Its one thing to fear what you don’t understand but to live in complete denial is just ignorant.
Irontiger wrote:
King_Geehm wrote:
Technology says yes. To put it simply: if it can be broken down into mathematics, it can be solved.
If this were true, your bank account is as good as hacked, as well as many other things that rely on RSA cryptography. The problem is "simply" one of breaking a large number into its prime factors, for which multiple algorithms exist.
King_Geehm wrote:
It’s one thing to fear what you don’t understand but to live in complete denial is just ignorant.
King_Geehm wrote:
Of course you like to throw around numbers but (...)
Yeah. Our belief in Shannon's entropy blinds us to the obvious truth that computer science trends are bound to continue forever. Our belief that mathematical questions have answers expressed in numbers blinds us to the fact only winners with a positive attitude make progress.
I've been in IT for a long time and have been amazed at the amount of progress that has been made. My first computer had 4k of RAM and a tape drive. Today I have a terabyte hard drive and more processing power than I could have ever imagined.
Your (sarcastic) point is true. The historical rate of progress cannot continue forever. In fact, we've already seen CPU speeds level off. We have seen more cores using less power (which is probably more useful in most cases), but overall processing power is not going up at the same rate as in the 80's, 90's and 00's. Somehow folks can only look at a chart and extrapolate wildly about how things will be in a century. WHile I wish it were true that things would continue to improve at the same rate forever, it simply is not true.
Much like the human population leveled off due to natural constraints, so will processing speed. Whether it levels off entirely in 10 years or 100 or 1000, it doesn't matter. The number of positions in chess mean that we will not use brute force to solve it entirely. And isn't this a good thing? Do we really want a 32 piece endgame tablebase so that we know for sure that a variation is good or not?
I'm sure I won't convince the True Believers in technology who worship it but don't understand it. But it is the truth.
King_Geehm wrote:
EricFleet wrote:
blhabidyblah blah... I live my life in doubt and fear of technological advances.... blhabidy blahblah.... I use a thesaurus to look up words like "infintesimal" to make it seem like im saying something really inteligent.... blahbidy blah blah... totally unrealted topic about Shannon's entropy... blah blah, blahbidy blah... right?
Considering I earn a living from technology, your first statement is false.
And considering I misspelled the word in question, I don't think I looked it up :)
And Shannon's entropy wasn't me... you might want to reread the posts. But no, it isn't unrelated. It refers to a limit on how much information can be stored in a message. It helps us define how much we can theoretically store in a given space.
And while your "blah blahs" are clearly the height of intelligent discourse, I do respectfully argue that you just might be wrong and possibly uninformed.
King_Geehm wrote:
Irontiger wrote:
If this were true, your bank account is as good as hacked...
I hate to be the one to break it to you but there has been an increasingly high number of "unbreakable" securities being broken lately; from websites, to banks, to government systems.
The hackers, if given enough time, can and do reverse engineer any code that is made. Their time is limited due to the authorities having systems in place to monitor and catch them. But surprisingly, chess does not have such authorities trying to stop them. So what was your comaprison again...? Oh yea, there is none.
But hey, that Shannon's entropy thing almost made you seem smart. If only it had some bearing on this topic.
This problem has a million dollar bounty, if you are so sure it can be solved go ahead and collect:
http://en.wikipedia.org/wiki/Millennium_Prize_Problems#P_versus_NP
Sep 10, 2014
0
#797
of course a significant portion of the 10^43 positions are trivial. the seven piece tablebases were released about five years ahead of several predictions because they excluded certain trivial positions, such as five queens of the same color. Such a reduction of the problem also made it possible for these tablebases to fit into a mere 140 TB of storage. You could put them on your hard drive.
Regarding that "if it can be broken down into mathematics, it can be solved"-thing, it's unfortunately false. Or should I say fortunately? Anyway, it's a fascinating subject. If you're interested there's a ton of books about this, but a good start might be this article.
Ziryab wrote:
of course a significant portion of the 10^43 positions are trivial. the seven piece tablebases were released about five years ahead of several predictions because they excluded certain trivial positions, such as five queens of the same color. Such a reduction of the problem also made it possible for these tablebases to fit into a mere 140 TB of storage. You could put them on your hard drive.
You have quite the hard drive... And even if 99.99% are trivial (I doubt it is nearly that high), you still have 10^39 posions left :)
Depends on your definition of "solved"