@7782
"define what it means to solve chess?"
'weakly solved means that for the initial position a strategy has been determined to achieve the game-theoretic value against any opposition' - Prof. van den Herik
Chess will never be solved, here's why
Sort:
@7778
"If there is the capability of exhaustively solving chess for 7 pieces - one would think that it's theoretically possible to solve it for 32 pieces"
++ That would be strongly solving Chess to a 32-men table base, but that would require all 10^44 legal positions, which would take billions of years and 10^44 bit of storage.
Jan 27, 2023
0
#5744
Optimissed wrote:
MEGACHE3SE wrote:
optimissed its one thing to call someone's bs, but its another thing entirely to start bullying and being all around cringe. u r starting to be the latter.
Some people try to bully. As I remember, you were the one who boasted that he refuted tygxc's nonsense all by himself and that you weren't surprised that others couldn't. That take a certain amount of cringeworthiness to accomplish. You were, in fact, trying to bully everyone here. A number of people have been trying to bully me for quite a while and I've decided that enough is enough UN to adopt the strongest stance because by now I'm quite sure that those who have been criticising my ideas and insisting that they are right and I'm wrong are doing so incorrectly. They make too many mistakes, don't seem to understand the subject anymore than I do and place far too much faith in experts who are half a century out of date. Yet they won't accept any criticism at all.
Ive went quite a few pages back and I don't see anyone actually explaining where txgxc's math was wrong, although feel free to point it out. simply saying 'its 10^34" isnt enough. it needs to be explicitly why u cant take the square root of that. all of this 'strong vs weak' solving is a red herring, and, if anything those definitions as to what constitutes 'strong' vs 'weak' solving should have been accepted because those definitions actually hurt txgxc's position.
you are also doing a tu quoque fallacy with regards to bullying, but i digress.
@7775
"your 10^17 calculation relies on a pre existing algorithm that creates the best move for ____ on any given position"
++ No, the calculation does not rely on any evaluation algorithm to find the best move.
Many here still do not understand that.
It relies on the 7-men endgame table base and on its absolute evaluations draw / win / loss.
If a calculated series of moves reaches a 7-men endgame table base draw,
then that validates all black moves as fit to achieve the game-theoretic value of the draw.
As for the white moves, all reasonable alternatives need to be explored, not just one.
Jan 27, 2023
0
#5746
tygxc wrote:
++ No, the calculation does not rely on any evaluation algorithm to find the best move.
yeah you do need the best move, literally by definition.
how are you supposed to figure out what black's position is then if you dont have whites move?
@7796
"I don't see anyone actually explaining where txgxc's math was wrong" ++ It is right
"simply saying its 10^34 isnt enough" ++ The 10^34 comes from Gourion's 10^37, multiplied by 10 to allow 3 or 4 queens, and divided by 10,000 as a random sample of 10,000 Gourion positions showed none that can result from optimal play by both sides.
"why u cant take the square root of that"
++ We can. That is the essential difference between weakly and strongly solving: one versus all black moves.
"all of this 'strong vs weak' solving is a red herring"
++ The difference is essential and makes all the difference between 5 years and billions of years, but many here still do not understand that. Checkers has been weakly solved, not strongly.
"those definitions as to what constitutes 'strong' vs 'weak' solving should have been accepted"
++ Indeed, definitions by the world authority on game solving should be accepted.
@7798
"you do need the best move"
++ No, I do not. For the one black move the justification comes from reaching a 7-men endgame table base. For white all reasonable moves need to be explored, not just one best move.
Jan 27, 2023
0
#5749
tygxc wrote:
++ We can. That is the essential difference between weakly and strongly solving: one versus all black moves.
how do you figure out which move it is?
Jan 27, 2023
0
#5751
tygxc wrote:
@7796
"I don't see anyone actually explaining where txgxc's math was wrong" ++ It is right
"simply saying its 10^34 isnt enough" ++ The 10^34 comes from Gourion's 10^37, multiplied by 10 to allow 3 or 4 queens, and divided by 10,000 as a random sample of 10,000 Gourion positions showed none that can result from optimal play by both sides.
"why u cant take the square root of that"
++ We can. That is the essential difference between weakly and strongly solving: one versus all black moves.
"all of this 'strong vs weak' solving is a red herring"
++ The difference is essential and makes all the difference between 5 years and billions of years, but many here still do not understand that. Checkers has been weakly solved, not strongly.
Checkers had a simplification algorithm/position. chess, at the moment, does not. without a simplification algorithm/position, weak versus strong solving are, for our purposes, essentially identical.
you also misread my post. the "simply saying 10^34" is the proposed response to your math error, which I was explaining is not sufficient in disproving it.
Jan 27, 2023
0
#5752
Optimissed wrote:
tygxc wrote:
@7775
"your 10^17 calculation relies on a pre existing algorithm that creates the best move for ____ on any given position"
++ No, the calculation does not rely on any evaluation algorithm to find the best move.
Many here still do not understand that.
Still do not understand? After what explanation? I don't see one.
An evaluation algorithm is an absolute essential and none exists that's fully reliable.
It relies on the 7-men endgame table base and on its absolute evaluations draw / win / loss.
That's an absolute equivalent to calculating all the way to checkmate. Since the seven man table base was an easy project compared with this, then logically, relying on calculating to that doesn't save any time worth speaking of. Also, it's such a massive amount of calculation that it's necessary to make evaluations on lines in order to cut most of them out and so an algorithm is an absolute necessity.
You've changed your story entirely, since a couple of months ago. You were relying on the Stockfish algorithm, which is obviously incapable of the job but still you were relying on it.
If a calculated series of moves reaches a 7-men endgame table base draw,
then that validates all black moves as fit to achieve the game-theoretic value of the draw.
As for the white moves, all reasonable alternatives need to be explored, not just one.
But you wouldn't even know if they were the best (optimal) moves. How could you possibly untangle them?
See? u can refute without being a dingus.
btw i am also sometimes a dingus so i will try to not get defensive if u call me out in the future. i cant criticize ur tone without putting myself open to criticisem
Jan 27, 2023
0
#5753
In order for chess to be solved, an extremely strong engine needs to analyse an incredible 10^120 different games (estimation), a number higher than the number of atoms in the universe.
Secondly, you never know if the present engine's recommendations are truly the absolute 'best' moves that are possible. Anyway, the progress chess engines have made is rather scary. But, I don't know as chess with less than 7 pieces has already been solved, thanks to tablebases. We really should hope that chess never gets solved.
@7801
"how do you figure out which move it is?"
++ For black take the top 1 Stockfish engine move after enough calculation time and do not worry if it is correct or not. If it is correct then it will ultimately lead to a 7-men endgame table base draw and that validate all black moves as fit to achieve the game-theoretic value of a draw.
If it is not correct, then some white move will lead to a 7-men endgame table base loss.
Likewise for white take the top w Stockfish engine moves. As previously calculated the table base exact move will be among the top w = 4 engine moves at 17 s/move on a 10^9 nodes/s engine with 1 error in 10^20 positions, i.e. no error in the 10^17 relevant positions.
Jan 27, 2023
0
#5755
"Likewise for white take the top w Stockfish engine moves"
okay but what if all of the top stockfish moves are wrong?
Jan 27, 2023
0
#5756
tygxc wrote:
@7801
"how do you figure out which move it is?"
++ For black take the top 1 Stockfish engine move after enough calculation time and do not worry if it is correct or not. If it is correct then it will ultimately lead to a 7-men endgame table base draw and that validate all black moves as fit to achieve the game-theoretic value of a draw.
If it is not correct, then some white move will lead to a 7-men endgame table base loss.
Likewise for white take the top w Stockfish engine moves. As previously calculated the table base exact move will be among the top w = 4 engine moves at 17 s/move on a 10^9 nodes/s engine with 1 error in 10^20 positions, i.e. no error in the 10^17 relevant positions.
you are also making the assumption that black is not inherently winning, that hasnt been proven.
@7806
"for chess to be solved, an extremely strong engine needs to analyse an incredible 10^120 different games (estimation), a number higher than the number of atoms in the universe."
++ No. There are only 10^44 legal positions, of which 10^17 are relevant to weakly solving chess. There are between 10^29241 and 10^34082 different Chess games.
"you never know if the present engine's recommendations are truly the absolute 'best' moves"
++ We do know. If a black move leads to a 7-men endgame table base draw, then it is good enough to achieve the game-theoretic value of a draw.
As for white moves we need to explore all reasonable moves that oppose to the draw.
"We really should hope that chess never gets solved."
++ If it is desirable or not is a different question. Ask the Checkers, Nine Men's Morris, Connect Four and Losing Chess players, whose favorite games have been weakly solved.
Jan 27, 2023
0
#5758
ironically your assumptions of correct gameplay outcomes arent even connected to your core errors, but i feel it is important to point them out regardless.
" If a black move leads to a 7-men endgame table base draw, then it is good enough to achieve the game-theoretic value of a draw."
you need to prove it leads to a draw or not, your explanation doesnt do that. it just puts it off.
@7809
"black is not inherently winning, that hasnt been proven"
++ That is proven. See https://www.iccf.com/tables
Assume Chess is a black or white win and try to fit a Poisson distribution of the number of errors per game such that the probability of an odd number of errors corresponds. It is impossible.
Assume Chess a draw and try to fit a Poisson distribution of the errors per game such that the probability of an odd number of errors per game is. It is possible and leads to >1000 perfect drawn games with optimal play from both sides.
White has the advantage of the initiative, worth 1 tempo, but not enough to win.
Jan 27, 2023
0
#5760
tygxc wrote:
@7806
"for chess to be solved, an extremely strong engine needs to analyse an incredible 10^120 different games (estimation), a number higher than the number of atoms in the universe."
++ No. There are only 10^44 legal positions, of which 10^17 are relevant to weakly solving chess. There are between 10^29241 and 10^34082 different Chess games.
"you never know if the present engine's recommendations are truly the absolute 'best' moves"
++ We do know. If a black move leads to a 7-men endgame table base draw, then it is good enough to achieve the game-theoretic value of a draw.
As for white moves we need to explore all reasonable moves that oppose to the draw.
"We really should hope that chess never gets solved."
++ If it is desirable or not is a different question. Ask the Checkers, Nine Men's Morris, Connect Four and Losing Chess players, whose favorite games have been weakly solved.
10^29241?? 💀
It is solved for a 7-piece endgame but not the entire game. You can't determine if a move suggest by Stockfish in a complex middle game truly the 'best' possible.
Chess getting solved would be a disaster as it would end the beauty of this game. The game would turn from a competetive sport to something like Tic-Tac-Toe. The future generations won't be able to learn from the valuable life lessons learnt from the game. And also it won't be fun anymore.
@7786
"Then, the next question: Will it take a billion years?"
++ Weakly solving chess takes 5 years.
3 cloud engines of 10^9 nodes/s can exhaust the 10^17 relevant positions in 5 years.