one engame with 7 pieces remaining is solved provided you can memorize 549 exact moves that forces a checkmate, but there is a rule that you must capture a piece or move a pawn once every 50 moves or else it’s a draw.
the other issue is that to solve for a perfect game you need to calculate every possible decision tree. It is easier to map every single atom in every star system currently detectable by any means from the Earth than it is to map every chess move.
19 TB is not that big anymore. For a company that buys storage systems, the more standard amount of useable space is going to be closer to a PB per system.
I think standard openings will get more and more moves added to them, but even at the highest level of chess there are still many valid openings, and many valid responses to each of those openings. Then, even after playing those “known openings”, it very quickly ceases to be solved.
That’s why “solved” is in air quotes. I didn’t want to go into a long tangent about how while chess isn’t technically a solved game, the opening and endgame databases that computers use, along with pretty powerful chess engines, essentially change the nature of the game. There’s lots to memorize in the fairly rote early and late game, in the service of reaching an interesting and tactical middle game.
(I had a similar issue with starcraft…)
I think Kasperov has it right that hybrid chess is interesting because it let’s the computer do the memorizing and give you a hand with board analysis. But i don’t think of chess as particularly sacred so I just found other games that while not as popular or sometimes as deep, offer a more thrilling, unknown challenge.
It is true that there are theoretically good moves, and bad moves in the opening, but this is true of the entire game of chess. The existance of the opening database, and memorizing lines really just makes it a bit easier on the chess engine, and the player – processing each move to such a great extent isn’t necessary if one knows what move is safe beforehand – but that still doesn’t necessarily mean that the opening is solved. If by “solved” you are referring to the win rate for a particular opening, then I would caution against that, as it can be rather misleading, depending on how exactly you are looking at it.
As for the endgame, I mean, there’s a pretty small set of moves that are possible – if you see a mate in 2, is that a solved game to take it?
along with pretty powerful chess engines
This point is moot – are you using a chess engine when you play against a human player? I would certainly hope not.
There’s lots to memorize in the fairly rote early and late game, in the service of reaching an interesting and tactical middle game.
Why not skip the memorization, and, instead, put your faith in your ability to strategize in every phase of the game?
Chess is not a solved game.
deleted by creator
one of the issues with solving chess is that
one engame with 7 pieces remaining is solved provided you can memorize 549 exact moves that forces a checkmate, but there is a rule that you must capture a piece or move a pawn once every 50 moves or else it’s a draw.
the other issue is that to solve for a perfect game you need to calculate every possible decision tree. It is easier to map every single atom in every star system currently detectable by any means from the Earth than it is to map every chess move.
I read that supercomputers can solve any 7-pieces-remaining chess game, but they need a ~19TB database in order to do so.
19 TB is not that big anymore. For a company that buys storage systems, the more standard amount of useable space is going to be closer to a PB per system.
I think standard openings will get more and more moves added to them, but even at the highest level of chess there are still many valid openings, and many valid responses to each of those openings. Then, even after playing those “known openings”, it very quickly ceases to be solved.
That’s why “solved” is in air quotes. I didn’t want to go into a long tangent about how while chess isn’t technically a solved game, the opening and endgame databases that computers use, along with pretty powerful chess engines, essentially change the nature of the game. There’s lots to memorize in the fairly rote early and late game, in the service of reaching an interesting and tactical middle game.
(I had a similar issue with starcraft…)
I think Kasperov has it right that hybrid chess is interesting because it let’s the computer do the memorizing and give you a hand with board analysis. But i don’t think of chess as particularly sacred so I just found other games that while not as popular or sometimes as deep, offer a more thrilling, unknown challenge.
It is true that there are theoretically good moves, and bad moves in the opening, but this is true of the entire game of chess. The existance of the opening database, and memorizing lines really just makes it a bit easier on the chess engine, and the player – processing each move to such a great extent isn’t necessary if one knows what move is safe beforehand – but that still doesn’t necessarily mean that the opening is solved. If by “solved” you are referring to the win rate for a particular opening, then I would caution against that, as it can be rather misleading, depending on how exactly you are looking at it.
As for the endgame, I mean, there’s a pretty small set of moves that are possible – if you see a mate in 2, is that a solved game to take it?
This point is moot – are you using a chess engine when you play against a human player? I would certainly hope not.
Why not skip the memorization, and, instead, put your faith in your ability to strategize in every phase of the game?