Damn glad to met you. One day wonder?
Sorry. It's two days. And you've already snagged a WGM friend. 
True or false? Chess will never be solved! why?
Sort:
you're cute.
To those who say chess is a draw:
"Solving" a game is a concept from game theory, which is a branch of mathematics. Therefore it is the standards of mathematics, rather than the standards of common sense or everyday thinking, that apply to the answer. In math, if you haven't proven something, you don't get to regard it as true. The statement that chess is a draw has not been proven with mathematical rigor; therefore it is a conjecture, which is a word more polite than "speculation" that mathematicians use for something they suspect, perhaps very strongly, to be true but have not proven. Even the unanimous opinion of masters and grandmasters that chess is a draw does not mean that a game theorist can consider the game solved, any more than the widespread opinion of mathematicians that, say, the Poincare conjecture was true counted for anything until it was proven in the 2000's. That's just how math is.
Furthermore, people with a mathematical background are usually very reluctant to accept as true anything that could in principle be proven if it were true but hasn't been. This is why so many of us who know next to nothing about chess reject the "knowledge" of chess experts on this question.
To those who argue that chess cannot be solved because [some number] is a very large number, your belief that storing that much data or examining that many positions is necessary to solve the game is also conjecture as far as I know.
Sorry, numbnuts. String Theory doesn't apply. We are not all Universalists.
Do your homework. And learn to write more concisely.
Math lost its certainty almost a hundred years ago. Duh?
cjoev wrote:
To those who argue that chess cannot be solved because [some number] is a very large number, your belief that storing that much data or examining that many positions is necessary to solve the game is also conjecture as far as I know.
We're talking about having a 32-men database, ie, the evaluation of all positions.
This needs storage of at least the answer.
It wont fit on a normal home PC size hard drive. But what if you spread it across 5 or 6?
You're going to need at least seven.
Just like a giant 7-layer peanut butter club sandwich. Yummy.
I've never heard of this peanut butter club before, but I'm in.
Irontiger wrote:
cjoev wrote:
To those who argue that chess cannot be solved because [some number] is a very large number, your belief that storing that much data or examining that many positions is necessary to solve the game is also conjecture as far as I know.
We're talking about having a 32-men database, ie, the evaluation of all positions.
This needs storage of at least the answer.
I'm not so sure. The answer is so large that you can only look at a very small portion of it at a time. So what's the difference between storing the whole thing and having an algorithm that can produce the part you want to see on demand? The naive algorithm for evaluating a chess position does not take an unreasonable amount of space to run, just an unreasonable amount of time. Space vs. time is the classic trade-off in computer science. It certainly doesn't seem likely that any approach to solving chess can succeed with both a reasonable amount of space and a reasonable amount of time, but that's conjecture.
So maybe 2 hard drives with a search speed of an hour?
Rather than instant recall with 7.
cjoev wrote:
Irontiger wrote:
cjoev wrote:
To those who argue that chess cannot be solved because [some number] is a very large number, your belief that storing that much data or examining that many positions is necessary to solve the game is also conjecture as far as I know.
We're talking about having a 32-men database, ie, the evaluation of all positions.
This needs storage of at least the answer.
I'm not so sure. The answer is so large that you can only look at a very small portion of it at a time. So what's the difference between storing the whole thing and having an algorithm that can produce the part you want to see on demand? The naive algorithm for evaluating a chess position does not take an unreasonable amount of space to run, just an unreasonable amount of time. Space vs. time is the classic trade-off in computer science. It certainly doesn't seem likely that any approach to solving chess can succeed with both a reasonable amount of space and a reasonable amount of time, but that's conjecture.
Arguably this would qualify as a strong solution but how would you know it without having the 32 man tablebase to confirm?
cjoev wrote:
Irontiger wrote:
cjoev wrote:
To those who argue that chess cannot be solved because [some number] is a very large number, your belief that storing that much data or examining that many positions is necessary to solve the game is also conjecture as far as I know.
We're talking about having a 32-men database, ie, the evaluation of all positions.
This needs storage of at least the answer.
I'm not so sure. The answer is so large that you can only look at a very small portion of it at a time. So what's the difference between storing the whole thing and having an algorithm that can produce the part you want to see on demand? The naive algorithm for evaluating a chess position does not take an unreasonable amount of space to run, just an unreasonable amount of time. Space vs. time is the classic trade-off in computer science. It certainly doesn't seem likely that any approach to solving chess can succeed with both a reasonable amount of space and a reasonable amount of time, but that's conjecture.
Fair enough, but the timescales become completely unreasonable then.
At least you wouldn't get much cheating if it took people 6 million years to check the tablebase.
TheGrobe wrote:
cjoev wrote:
Irontiger wrote:
cjoev wrote:
To those who argue that chess cannot be solved because [some number] is a very large number, your belief that storing that much data or examining that many positions is necessary to solve the game is also conjecture as far as I know.
We're talking about having a 32-men database, ie, the evaluation of all positions.
This needs storage of at least the answer.
I'm not so sure. The answer is so large that you can only look at a very small portion of it at a time. So what's the difference between storing the whole thing and having an algorithm that can produce the part you want to see on demand? The naive algorithm for evaluating a chess position does not take an unreasonable amount of space to run, just an unreasonable amount of time. Space vs. time is the classic trade-off in computer science. It certainly doesn't seem likely that any approach to solving chess can succeed with both a reasonable amount of space and a reasonable amount of time, but that's conjecture.
Arguably this would qualify as a strong solution but how would you know it without having the 32 man tablebase to confirm?
Perhaps I misunderstand the definition of a strong solution, but what is there to confirm? You'd know the algorithm is correct because it would be simple enough to check by hand, and whatever position you wanted to know about, you'd run the algorithm starting from there. Basically, the algorithm is the 32-man tablebase, highly compressed.
But you'd have to check every position to fit your version of solved? Might take a while.
cjoev wrote:
Perhaps I misunderstand the definition of a strong solution, but what is there to confirm? You'd know the algorithm is correct because it would be simple enough to check by hand, and whatever position you wanted to know about, you'd run the algorithm starting from there. Basically, the algorithm is the 32-man tablebase, highly compressed.
By this definition, chess is solved right now. Launch Rybka or Houdini with a calculation depth of one billion moves, and wait.
Irontiger wrote:
cjoev wrote:
By this definition, chess is solved right now. Launch Rybka or Houdini with a calculation depth of one billion moves, and wait.
I'm not claiming that. Clearly a "solution" that takes prohibitively long to query should not count. But what if there exists some brilliant new algorithm, different from the well-known tree search and much faster, based on some strange but powerful insight into the structure of chess that we have not yet discovered, that can solve a single position very quickly? I'm not saying such a thing exists, I'm just saying that (as far as I know) no one has proved that it doesn't.
which pig said that ?
feel free to be "relevant, friendly or nice".