Originally posted by BowmannComputers have solved chess when there are five or fewer pieces remaining. They are working on six pieces. Bowmann's few trillion years may be close to the mark for twenty piece positions.
Perhaps you forgot that this thread was about solving chess, not building better machines to play it.
In order for computers to solve it without considering all possibilities, programmers must solve it first. Pruning is only as effective as the knowledge behind the algorithm.
Do the math. A google is a small number by comparison.
Yes it's possible.
It's possible to travel at speeds close to the speed of light, but getting the energy to do that is almost impossible right now.
I'm not aware of the tree of chess, but it probably means branching each move into more possible moves, and those into more.
And if you look at it like that, yes it is all just numbers and numbers can be calculated.
It just needs a hell of a lot computational power.
So kasparov said computers lack intuition so they'll never win? He's an idiot! Chess is numbers!
So do I think one day we will have that kind of technology? Yeah I do.
Originally posted by adoosC'mon, folks. There are people with the math and computer skills, as well as access to computational power. Some of them have presented part of this work in accessible format on the world wide web. Take a look at http://www.cs.berkeley.edu/~flab/chess/chess.html before you utter nonsense.
And if you look at it like that, yes it is all just numbers and numbers can be calculated.
It just needs a hell of a lot computational power.
After 5 1/2 moves, there are 2097651003696806 possible games. It has been estimated that after 40 moves, there are 10^120 possible games. Extend the game beyond 40 moves, and you should understand that God plays chess because she finds it a challenge.
As Adoos said, chess is numbers. He is probably the only other person here who understands computers, quantum computers, and their potential. And then you throw out 10^120. Well, if you knew anything about quantum computing or had read what I just said, i.e. a 500 qubit computer could do 2^500 calculations per second. A 1000 qubit computer could do 2^1000 calculations per second. 2^1000 is approximately 2^300, way more than the 10^120 that you find to be such a huge number. And all this in one second. God forbid you run it for more than a second, or you have more than one computer. Hopefully, this makes you understand its potential computational power.
Originally posted by Balla88Did you check Lausey's profile?
As Adoos said, chess is numbers. He is probably the only other person here who understands computers...
Everything else you say indicates that you seriously overestimate the potential of a "quantum computer" and misunderstand the theory behind such machines.
Originally posted by BowmannI have done work on the quantum theory behind quantum computers and I have good friends who work in the field. I did not overestimate their potential at all. Please check your facts before you write complete nonsense. Since when is Lausey a quantum theory experts. Quantum computers will be able to solve NP-complete problems. Like chess, the time required to solve these problems increases greatly with added complexity. It is the exact same concept. Their brute force power is immense. They will be able to factor numbers, like those used in RSA encryption, not in 8,400 MIPS years (what it would take a regular computer), but in 3.5 hours.
Did you check Lausey's profile?
Everything else you say indicates that you seriously overestimate the potential of a "quantum computer" and misunderstand the theory behind such machines.
Originally posted by Balla88For the record, I do have a BSc in Applied Physics. Although not a quantum theory expert, I do have a limited understanding of quantum physics.
I have done work on the quantum theory behind quantum computers and I have good friends who work in the field. I did not overestimate their potential at all. Please check your facts before you write complete nonsense. Since when is Lausey a quantum theory experts. Quantum computers will be able to solve NP-complete problems. Like chess, the time required to ...[text shortened]... A encryption, not in 8,400 MIPS years (what it would take a regular computer), but in 3.5 hours.
We are talking in terms of solving chess in this thread. In other words, the ability to determain the best next move in any position. Hence you would have to have the storage capacity to achieve this (which could not even be practically done on a quantum level).
As for a brute force calculation using quantum computers for each move. I tend to find that your figures seem rather far fetched. You are saying that a two fold increase in quantum computer storage (500 qubits to 1000 qubits) achieves a 1^500 fold increase in processing power (2^500 flops to 2^1000 flops). Also, what do you mean by 2^1000 is approximately 2^300?
Originally posted by lauseyevidently he was about to say that 2^1000 is approximately 10^300 being bigger than 10^120.
For the record, I do have a BSc in Applied Physics. Although not a quantum theory expert, I do have a limited understanding of quantum physics.
We are talking in terms of solving chess in this thread. In other words, the ability to determain the best next move in any position. Hence you would have to have the storage capacity to achieve this (which ...[text shortened]... ng power (2^500 flops to 2^1000 flops). Also, what do you mean by 2^1000 is approximately 2^300?
The misunderstand of Adoos is that the 10^120 consits of not only one information but several. If we could compute the singel determining number per game we would be far advanced. So the Quantum computer could probably enumerate the games, but that is far from building up the tree.
My five cents. Diploma in Chemistry and not really quantum computer specialist
Originally posted by lauseyYes, I am saying that that 500 qubit increase would increase the power from 2^500 flops to 2^1000 flops. That's the the way quantum computers work, exponential increase in power. That's why they would be able to solve NP-complete problems.
For the record, I do have a BSc in Applied Physics. Although not a quantum theory expert, I do have a limited understanding of quantum physics.
We are talking in terms of solving chess in this thread. In other words, the ability to determain the best next move in any position. Hence you would have to have the storage capacity to achieve this (which ...[text shortened]... ng power (2^500 flops to 2^1000 flops). Also, what do you mean by 2^1000 is approximately 2^300?
Naturally, this discussion hinges on the power of computers, but let's imagine that there is a computer of inifinte power, that can "solve" chess, brute force, in an instant...
Somehow, at the start of a game, I cannot imagine it saying White checkmates Black in 185 moves... but I can imagine it saying 0.00.
Wouldn't there be just one perfect game? And isn't the solution to chess a draw?
Only when "errors" are made would the solution flip from draw to white win to black win.
I guess, having the solution to chess would be just as useful as having the solution to a crossword puzzle. Something you can rate yourself against once the game is over. It shouldn't detract at all from the joy of playing.