Not all of the positions are very meaningfull or dificult.
Is it possible to solve chess?
Sort:
Maybe in the year 3000-4000 something, super computers will be invented that can store that much information. Chess will eventually be solved but not in this century.
Dec 30, 2009
0
#4
Yes but how do you store that much information? Even if you save one position to as small as an atom...
Maybe this answer is momentarily above our heads. If our ancestors from 1000 years ago would see today how we live... computers etc... I bet they wouldn't believe their own eyes.
There is another thread about this with over 100 comments.
Just the recent ones!
http://www.chess.com/forum/view/general/theory-of-perfect-line?quote_id=3228748
http://www.chess.com/forum/view/general/can-chess-be-solved-by-computers?page=1
http://www.chess.com/news/computers-have-solved-checkers-is-chess-next
http://www.chess.com/forum/view/general/solving-chess
New question please.
Dec 30, 2009
0
#8
Kallatroh wrote:
The number of positions is in the neighbourhood of the number of atoms in the universe. How can you store all that information? Is it even physically possible?
Actually far higher.
10^120 chess (likely higher)
10^83 atoms (likely lower)
OTOH, Rybka 4.0 is close to being capable of never losing with the White pieces.
Maybe you don't need to store every position to solve chess. Maybe you just need a formula or something, that can calculate the best move for every position. Would be extremely hard to develop that formula, but even if it will never happen, we can't say that it's impossible. Maybe the formula exists, just that it's too large and too complicated to be discovered.
veclock wrote:
Maybe you don't need to store every position to solve chess. Maybe you just need a formula or something, that can calculate the best move for every position. Would be extremely hard to develop that formula, but even if it will never happen, we can't say that it's impossible. Maybe the formula exists, just that it's too large and too complicated to be discovered.
Yeah but to test the formula to 100% accuracty you'd have to have it solve every position -- again of course it's possible it's just very unlikely.
Also note that solving each position is also useless without some kind of index to organize it all, which would also take up a lot of space hehe.
Ziryab wrote:
OTOH, Rybka 4.0 is close to being capable of never losing with the White pieces.
Where have you read/heard that? Presumably it would lose to another engine with a higher ELO rating?
Go actually looks VASTLY more complex than chess (10^761 !!). I wonder if Go will also one day be "solved"?
Dec 30, 2009
0
#12
chessoholicalien wrote:
Ziryab wrote:
OTOH, Rybka 4.0 is close to being capable of never losing with the White pieces.
Where have you read/heard that? Presumably it would lose to another engine with a higher ELO rating?
Go actually looks VASTLY more complex than chess (10^761 !!). I wonder if Go will also one day be "solved"?
It is an interesting claim by the author of Rybka.
http://chessok.com/download/A%20Conversation%20With%20Vasik%20Rajlich.wmv
Rajlich also mentioned that he intends to devote attention to Go or poker in two years.
Computer Go players can barely keep up with beginner to moderate level chess players. It is simply much too complex.
I would like to see a strong algorithm based program (in chess or Go). Hopfully Go is complex enough that they turn to algorithms and impress us with some strong play.
Dec 30, 2009
0
#15
orangehonda wrote:
I would like to see a strong algorithm based program (in chess or Go). Hopfully Go is complex enough that they turn to algorithms and impress us with some strong play.
Rajlich said in the interview that developing algorithms is what he most enjoys doing. He stated that for the next eighteen months to two years his goal is to keep Rybka at the top of the rating lists, then he plans to work on poker or Go. Poker seems to draw many chess enthusiasts away becuse there's lots of money to be filched.
Ziryab wrote:
orangehonda wrote:
I would like to see a strong algorithm based program (in chess or Go). Hopfully Go is complex enough that they turn to algorithms and impress us with some strong play.
Rajlich said in the interview that developing algorithms is what he most enjoys doing. He stated that for the next eighteen months to two years his goal is to keep Rybka at the top of the rating lists, then he plans to work on poker or Go. Poker seems to draw many chess enthusiasts away becuse there's lots of money to be filched.
Your earlier comment prompted my post -- I downloaded the interview but it's very long and I haven't watched it yet :), glad to hear he enjoys algorithms -- I'm not interested in poker, hope he goes to Go.
I'm no expert at chess.
Just 17... Trying to get past A Levels 
But, I believe to solve chess, you have to be good at statistics.
If I had three letters. ABC. This can be rearranged 6 ways:
ABC ACB BAC BCA CAB CBA
Simplified; 6! which is 3x2x1 = 6
Since there are 8x8 squares = 64 squares in total.
64! = 1.26886932 × 1089
Which is a ridiculously high number. To find all the possible movements, this will take around 2-3 years depending on how fast you count the number of ways. Once you have overcome this; You will have managed to memorise every single possible chess game that ever happened or could happen. Then create a programme where the computer will be able to play you using the process of elimination.
Until then...... We humans are useless!
TheLoveBelow wrote:
Since there are 8x8 squares = 64 squares in total.
64! = 1.26886932 × 1089
That's not exact. The total number of permutations with ABC is indeed 3! = 6 (by the way, you wrote 6! which is 6x5x4x3x2x1 = 720), but the total number of permutations with a chess board is not 64! for the simple reason that there are not 64 unique states on a chessboard; there are far less, since empty squares are equivalent, as are similar pieces of the same color (with a limitation on bishops). In addition, there are limitations as to which squares pawns can be located, and there's a number of illegal positions, the number of which is nigh impossible to calculate.
Against computers, play tight, forget experiments, eventually they´ll mess it up!
I suspect that quantum mechanics and quantum computers will easilly solve chess. I am trying to read some articles about it.
http://www.google.pt/search?hl=pt-PT&safe=off&rls=com.microsoft:*:IE-SearchBox&&sa=X&ei=W4riTL-nOIqW4AaLwPStBA&ved=0CCEQBSgA&q=quantum+mechanics+chess&spell=1
The number of positions is in the neighbourhood of the number of atoms in the universe. How can you store all that information? Is it even physically possible?