Elroch wrote:
You can make good batter with bitter as well.
Better batter by bitter benefits from butter, or don't bother.
Chess will never be solved, here's why
Sort:
PDX_Axe wrote:
It is a given that in time a powerful enough computer will be created that can solve chess, literally finding a path to a win from every possible position. And really, who cares? Unless you are Hans Nieman {just kidding, maybe} you don't use an engine when you are playing chess against an opponent. You have not solved chess, and never will, which for most of us is why we play. Not for the mathematics, but the art and beauty that can be found in the game. It is an addiction, our joy and pain simultaneously. Most of this thread is pointless argument in my opinion, but feel free to differ. Some people just get off on contention.
It's not actually a given for anyone that understand computers and the numbers involved. Currently, if you spent the entire wealth of the planet on computers and applied them exclusively to the problem it would still take millions of years.
Sep 20, 2022
0
#3984
In 1999 I purchased a brand new Imac, one of those colorful ones. It came with a whopping 25mb of ram, and I got the max upgrade to 150mb. That was a state of the art computer in 1999. Today in 2022 it's a joke, a mere 23 years later. It won't take millions of years, it might not even take a hundred. People really don't understand time, and how quickly our technology is progressing faster and faster. That chess will be solved by a computer IS a given. However you don't use computers in otb play, unless you are cheating or the rules change. Until they start putting micro computers into our heads to assist our brains the technology will be irrelevant, and so will the fact that a computer might have solved chess.
Sep 20, 2022
0
#3985
Storage is irrelevant, in 2000 desktop processors had 4 million transistors, they now have 13 billion; super computing at this point is off the scale.
But the thing is quantum computing is many orders of magnitude above that too, because they can operate with many outcomes and many assumed initial states, something like chess will be facile. I actually think chess is solvable with traditional processing*, but with quantum computing, it will be trivial.
Some of the simulations of quantum computing could be used to solve something like chess...they can simulate upwards of 5000 cubits currently, I don't think that is far off.
PDX_Axe wrote:
In 1999 I purchased a brand new Imac, one of those colorful ones. It came with a whopping 25mb of ram, and I got the max upgrade to 150mb. That was a state of the art computer in 1999. Today in 2022 it's a joke, a mere 23 years later. It won't take millions of years, it might not even take a hundred. People really don't understand time, and how quickly our technology is progressing faster and faster. That chess will be solved by a computer IS a given. However you don't use computers in otb play, unless you are cheating or the rules change. Until they start putting micro computers into our heads to assist our brains the technology will be irrelevant, and so will the fact that a computer might have solved chess.
Well then.
In 1999 when you bought your Imac, I was director of system software development for an ISP. My computer was quite a bit better than yours, as well (actually both the PC, and the Sun Microsystems server I was running at my desk were quite a bit more robust).
That is all irrelevant, though. You need to grasp what 10^44 means. Engines have never played through ~99.999999999999999999999999999% of possible positions, and they are not even capable of evaluating perfect play if they did.
The storage required would take at minimum our solar system's entire asteroid belt in materials even with the very best estimates for quantum computing, which, by the way, currently cannot process a problem like solving chess as the technology sits now.
So, unless space flights take a rather giant leap forward pretty quick...you're out of luck even getting started with the initial ore refining in the century you have given yourself.
P.S. Moore's Law is failing to hold true any longer, and silicon is hitting its physical limits, so don't expect it to make any big comebacks, either. If you don't know what Moore's Law is...then you have really set sail on a rudderless ship. Time to head back to harbor.
Sep 20, 2022
0
#3987
PDX_Axe wrote:
People really don't understand time, and how quickly our technology is progressing faster and faster.
The problem is the reverse, you don't understand how technology is progressing or how it compares to the complexity of chess. First of all, it's well known that the rate at which processors are improving isn't as fast. Moore's law runs into multiple problems. A big one being we can't miniaturize transistors anymore without running into quantum effects.
Anyway, a common error among ignorant people seems to be they think chess is "really" hard and computers are "really" fast, so they assume it's roughly equivalent, but it's not.
crocodilestyle1 wrote:
Storage is irrelevant, in 2000 desktop processors had 4 million transistors, they now have 13 billion; super computing at this point is off the scale.
But the thing is quantum computing is many orders of magnitude above that too, because they can operate with many outcomes and many assumed initial states, something like chess will be facile. I actually think chess is solvable with traditional processing*, but with quantum computing, it will be trivial.
Some of the simulations of quantum computing could be used to solve something like chess...they can simulate upwards of 5000 cubits currently, I don't think that is far off.
...another person that really needs to read this and the other various threads on the subject.
It's already been established that quantum computers *cannot* work on solving chess as they sit now. Quantum computers currently cannot store and read back any intermediate results, because those reads are destructive. The entire solving of chess would have to be done in one pass. It's not like traditional computing where you can output intermediate steps and refer to them later. Quantum computing is more like a self contained black box, where you pass the input data, then the answer is worked on via matrices flipping states. Reading those matrices destroys them in the process, so you can only do it once.
Think of it like The Imitation Game. Turing's computer just sits there and rotates cylinders endlessly until an answer dings at the end...or, it doesn't. You can't know what it has accomplished until it is done. If you stop it in the middle, you have nothing comprehensible.
There could perhaps be some sort of hybrid computing done with quantum computers that interact with more traditional storage...but then you run right back into the problem of materials requirements being even larger to the point where a single solar system is not going to cut the mustard.
Tell you what...when humanity has built a space elevator and is routinely pushing asteroids to Lagrange points for mining, then we should revisit this issue 
.
There are already dozen and dozens of pages posted debunking these exact arguments. Here's my own back of the envelope calculation from 2018:
https://www.chess.com/forum/view/general/will-computers-ever-solve-chess?page=162#comment-36272882
Sep 20, 2022
0
#3989
crocodilestyle1 wrote:
Storage is irrelevant, in 2000 desktop processors had 4 million transistors, they now have 13 billion; super computing at this point is off the scale.
4E6 vs 13E9 is only 3 orders of magnitude... of course that's quite a lot if your measuring stick is linear, but it's an imperceptible change compared to what's needed to solve chess.
It's like removing 1 bucket of water from the ocean and calling your friend on another continent to see if they noticed, then you remove 1000 buckets and call them again to see if they noticed. There's no difference.
Sep 20, 2022
0
#3990
btickler wrote:
There are already dozen and dozens of pages posted debunking these exact arguments.
Yeah by people who don't understand how quantum computing scales against conventional computing.
Where are the the electrons in a Hydrogen molecule? Come now, it is the simplest molecule that there is....
If you asked me that 24 years ago, I would have been as confused as you are now (well maybe not quite so confused....I guess you are probably googling what an electron is...but anyway, you get the point.) In the modern world the function of where an electron is in molecular hydrogen can be done on most people's mobile phone.
A qubit isn't like a bit, branch prediction can give you extraordinary results in computing with bits, in the quantum world you can get whatever is the predicate from the substrate. I actually think that standard computation can solve chess, quantum computing at a scale slightly beyond what we have now will do so trivially.
crocodilestyle1 wrote:
Yeah by people who don't understand how quantum computing scales against conventional computing.
Where are the the electrons in a Hydrogen molecule? Come now, it is the simplest molecule that there is....
If you asked me that 24 years ago, I would have been as confused as you are now (well maybe not quite so confused....I guess you are probably googling what an electron is...but anyway, you get the point.) In the modern world the function of where an electron is in molecular hydrogen can be done on most people's mobile phone.
A qubit isn't like a bit, branch prediction can give you extraordinary results in computing with bits, in the quantum world you can get whatever is the predicate from the substrate. I actually think that standard computation can solve chess, quantum computing at a scale slightly beyond what we have now will do so trivially.
It's you that does not understand. Quantum computing is only applicable to subsets of traditional computing problems, and solving chess is not one of them. Cryptography, weather prediction, and various "chaos" problems are.
Already blew the last guy that put forth your premise out of the water awhile back...heck, maybe it was you. He tried to go on and on about how quantum computers were already running Python, etc. and could tackle every problem traditional computers do, and I had to show him the reduced instruction set quantum computing actually supports that was covered in his own supporting links, and go through the exact same explanations of quantum computers and the problem of destructive reads, etc.
You're just laboring under the delusion that solving chess is a smaller problem than other problems people talk about undertaking with current and imminent technologies, because solving a relatively straightforward game seems like it should be trivial. But it isn't, it's far, far larger than any dataset humanity has ever dreamed of tackling.
You might as well say we're going to colonize Andromeda with FTL driven ships. Yeah, maybe someday, but making a prediction about it now is basically a call to magic ala Clarke's Law.
Sep 20, 2022
0
#3992
btickler wrote:
crocodilestyle1 wrote:
Yeah by people who don't understand how quantum computing scales against conventional computing.
Where are the the electrons in a Hydrogen molecule? Come now, it is the simplest molecule that there is....
If you asked me that 24 years ago, I would have been as confused as you are now (well maybe not quite so confused....I guess you are probably googling what an electron is...but anyway, you get the point.) In the modern world the function of where an electron is in molecular hydrogen can be done on most people's mobile phone.
A qubit isn't like a bit, branch prediction can give you extraordinary results in computing with bits, in the quantum world you can get whatever is the predicate from the substrate. I actually think that standard computation can solve chess, quantum computing at a scale slightly beyond what we have now will do so trivially.
It's you that does not understand. Quantum computing is only applicable to subsets of traditional computing problems, and solving chess is not one of them. Cryptography, weather prediction, and various "chaos" problems are.
Already blew the last guy that put forth your premise out of the water awhile back...heck, maybe it was you. He tried to go on and on about how quantum computers were already running Python, etc. and could tackle every problem traditional computers do, and I had to show him the reduced instruction set quantum computing actually supports that was covered in his own supporting links, and go through the exact same explanations of quantum computers and the problem of destructive reads, etc.
You're just laboring under the delusion that solving chess is a smaller problem than other problems people talk about undertaking with current and imminent technologies, because solving a relatively straightforward game seems like it should be trivial. But it isn't, it's far, far larger than any dataset humanity has ever dreamed of tackling.
You might as well say we're going to colonize Andromeda with FTL driven ships. Yeah, maybe someday, but making a prediction about it now is basically a call to magic ala Clarke's Law.
Hahahahahaha!
Yeah, quantum computing is just confined to solving a few problems that traditional computing can't get to....
Yeah....I can certainly see how hundred of billions of an analogy indeterminate juxtapositions will fail where a few million ones or zeros fail - why do us precious few even persevere with such a pointless enterprise?
Oh, please tell us more about your steam powered master machines.
(You don't even have the vaguest idea of the scale of the concept do you! Do you know what an electron is? Do you know what a neutrino is? [Do you know what science is?])
@5084
"All those who have been involved in the peer-reviewed research in this field understand that "all opposition" includes all legal moves." ++ Then it would have been worded 'against all legal moves'. Oppose = to strive against, resist
"heuristics that can have exceptions"
++ No.
'Other things being equal, any material gain, no matter how small, means success'
Is a perfect heuristic. It has two parts. Is there material gain? Are other things equal? If both are yes, then the position is won. The first part is easy and a computer can do it in a split second. The second part is trickier. For many positions even grandmasters wlll be unsure if all other things are equal or not. Then more calculation is needed. Fater 1 e4 e5 2 Ba6? white loses material. All other things are equal. White loses.
Another heuristic: pieces are better in the center. 1 Nf3 does more for the center than 1 Nh3. Thus 1 Nh3 opposes less to achieving the draw than 1 Nf3.
Another heuristic: except in certain conditions endgames with opposite colored bishops are draws. Now in this game https://www.iccf.com/game?id=1164259 they have reached an endgame with opposite colored bishops. The conditions for it being a win are not fullfilled. Thus it is a draw. The ICCF grandmasters agreed on a draw. The calculation for the solution of chess can halt there too and the good assistants can adjudicate it a draw. The calculation could go on until a 3-fold repetition, but that would be pointless.
"the solution of checkers was not completed by assuming all positions where one side was two pieces up was decided" ++ Likewise for chess positions a pawn, piece, rook or queen up should not be assumed to be decided. The good assistants must first verify that all other things are equal. If not then more calculation is necessary as in the case of sacrifices that may or may not be correct. 1 e4 e5 2 Ba6? is no such sacrifice. All other things are equal. It needs no grandmaster to see that. So 1 e4 e5 2 Na6? loses for white, no calculation to checkmate in 82 is needed.
Allis even solved Connect Four independently by only 7 rules: 7 heuristics. He noticed to connect 4 you first must connect 3: a big threat. He also noticed to connect 3 you first must create 3 connect 2. Then he also noticed the importance of the central column. Then he noticed that the 7*6 size of the board had tempo consequences favoring rows 1, 3, and 5. His whole strategy to solve Connect Four consists of 7 rules or heuristics.
Solving chess should be a combination of brute force and heuristics.
Pure brute force takes too long. It involves pointless calculations like 1 e4 e5 2 Ba6 to checkmate in all variations or like the opposite bishop ending continued until a 3-fold repetition.
Pure heuristics are not yet possible, though that is the way chess is taught to future grandmasters. Chess books are not filled with variations to memorise, but with heuristics to apply.
crocodilestyle1 wrote:
Hahahahahaha!
Yeah, quantum computing is just confined to solving a few problems that traditional computing can't get to....
Yeah....I can certainly see how hundred of billions of an analogy indeterminate juxtapositions will fail where a few million ones or zeros fail - why do us precious few even persevere with such a pointless enterprise?
Oh, please tell us more about your steam powered master machines.
(You don't even have the vaguest idea of the scale of the concept do you! Do you know what an electron is? Do you know what a neutrino is? [Do you know what science is?])
Oh look, it's the "Inconceivable!" guy from Princess Bride...
Just because you watched some NOVA special with Neil DeGrasse Tyson about the observed position of electrons and some fun stuff about quantum entanglement tossed in doesn't mean you know jack. And here, you clearly don't. Post some numbers, fly boy, or write us some code using an existing quantum computer's fully supported instruction set that would even evaluate who wins the simplest pawn race.
@5103
"quantum computing is just confined to solving a few problems that traditional computing can't get to"
++ If chess is ever strongly solved, then it will be by a quantum computer. Existing quantum computers can run any Python program. It is thinkable that a quantum computer generates an 8-men table base from the existing 7-men table base. Then it could go on.
tygxc wrote:
@5103
"quantum computing is just confined to solving a few problems that traditional computing can't get to"
++ If chess is ever strongly solved, then it will be by a quantum computer. Existing quantum computers can run any Python program. It is thinkable that a quantum computer generates an 8-men table base from the existing 7-men table base. Then it could go on.
...so it was you . Figures.
Made it easier to find though...here's the thread:
https://www.chess.com/forum/view/general/quantum-computer-will-it-hurt-chess-theory?page=1
Sep 20, 2022
0
#3997
btickler wrote:
crocodilestyle1 wrote:
Hahahahahaha!
Yeah, quantum computing is just confined to solving a few problems that traditional computing can't get to....
Yeah....I can certainly see how hundred of billions of an analogy indeterminate juxtapositions will fail where a few million ones or zeros fail - why do us precious few even persevere with such a pointless enterprise?
Oh, please tell us more about your steam powered master machines.
(You don't even have the vaguest idea of the scale of the concept do you! Do you know what an electron is? Do you know what a neutrino is? [Do you know what science is?])
Oh look, it's the "Inconceivable!" guy from Princess Bride...
Just because you watched some NOVA special with Neil DeGrasse Tyson about the observed position of electrons and some fun stuff about quantum entanglement tossed in doesn't mean you know jack. And here, you clearly don't. Post some numbers, fly boy, or write us some code using an existing quantum computer's fully supported instruction set that would even evaluate who wins the simplest pawn race.
Don't need to post numbers little sweety, you clearly think that quantum computing is on linear with standard computing - which kind of indicates that you don't know about either standard computing and certainly not quantum computing.
When was the last time you went to a university? (Actually, may I ask, have you ever been to a university? Do you know what a university is?)
crocodilestyle1 wrote:
Don't need to post numbers little sweety, you clearly think that quantum computing is on linear with standard computing - which kind of indicates that you don't know about either standard computing and certainly not quantum computing.
When was the last time you went to a university? (Actually, may I ask, have you ever been to a university? Do you know what a university is?)
1. It's spelled "sweetie".
2. Saying "on linear" is a clumsy and incorrect usage. Clearly, I *don't* think they are "on linear" though, because I just got done telling you that quantum computers are most useful for a specific subset of problems. For the observant, this means that I do not consider the two technologies to be on the same scale, or to be "linear" on such a scale.
3. You're still talking, but have not produced a single premise or argument yet.
Sep 20, 2022
0
#3999
btickler wrote:
crocodilestyle1 wrote:
Don't need to post numbers little sweety, you clearly think that quantum computing is on linear with standard computing - which kind of indicates that you don't know about either standard computing and certainly not quantum computing.
When was the last time you went to a university? (Actually, may I ask, have you ever been to a university? Do you know what a university is?)
1. It's spelled "sweetie".
2. Saying "on linear" is a clumsy and incorrect usage. Clearly, I *don't* think they are "on linear" though, because I just got done telling you that quantum computers are most useful for a specific subset of problems. For the observant, this means that I do not consider the two technologies to be on the same scale, or to be "linear" on such a scale.
3. You're still talking, but have not produced a single premise or argument yet.
Yeah you think the mathematics used by quantum computers is like junior school mathematics.....how sweet.
I spelled 'sweetie' as I did, because the 400 terms I used to describe someone who has not studied mathematics at school were rejected by the 'abuse system' of the chess.com forum.
crocodilestyle1 wrote:
Yeah you think the mathematics used by quantum computers is like junior school mathematics.....how sweet.
I spelled 'sweetie' as I did, because the 400 terms I used to describe someone who has not studied mathematics at school were rejected by the 'abuse system' of the chess.com forum.
1. No, I don't think anything like that. If anything, quantum computers would be more like linear algebra, given the matrices, so late high school or community college . If I were going to ascribe a concept involving "junior" to anything, it would probably be you, not a quantum computer.
2. Your admission of not being able to express yourself without being rejected repeatedly by the forum filters does not seem that surprising.
Have you tried to grok the quantum computing thread I linked above yet? Since you are seemingly so much better versed on quantum computers than I am, I was waiting eagerly to see your rebuttals to what I laid out about the QCL instruction set, the decoherence issues, etc.
Nah, I agree. Most of it is pointless.