How close are we to solving chess?
Sort:
also human being losing mmatch to present day programs is like human being losing 100 metre race against a scooter or car - it is not the end of anything!
normajeanyates wrote:
lol - CTC processors - when they dont win a game they go back in time and start at move 1. so in *some* universe of the multiverse they will solve and hence have solved chess - but in that universe they dont know it :) in other universes of course they cannot get info from that universe [by definition of universe]
Could anybody explain me what's a CTC processor?
polosportply wrote: ok, other question then: How close are we to making computers that are 99% unbeatable. GMs have limits and so do computers, but in the race between computers and grand masters, who's winning. With DeepBlue, something tells me the computers are. Am I right?
It depends on what type of chess you are talking about. If you mean OTB chess with the standar time controls then chess engines just kick butts all the time. But if you want to talk about correspondence chess then humans still kick butt and probabily will for a long time. The reason is that current day chess engines suck at strategical concepts and on CC (when propperly played) that's what matters.
Gabriel_dCF wrote: normajeanyates wrote:
lol - CTC processors - when they dont win a game they go back in time and start at move 1. so in *some* universe of the multiverse they will solve and hence have solved chess - but in that universe they dont know it :) in other universes of course they cannot get info from that universe [by definition of universe]
Could anybody explain me what's a CTC processor?
ctc - closed timelike curve. there are many solutions of the equations of general relativity [EFE ii einstien field equations] with CTCs - the first was the Godeldust universe - given bu Kurt Godel
In short, time machine :)
How nice, 
A computer like that, if possible, could literally foresee the opponent's move! But imagine what would happen if two CTC's played each other?
CTCs would logically necessitate either a multiverse or a total absence of sentient beings - in the previous case zillions of things would be happening in zillions of universes that keep splitting off. [by defination there cannot be any info-flow between two universes]
I once devised an idea for a chess tournament happening in the future (for a science fiction short) wherein the rules were that each side had X number of "splits" they could play. So they start with one game, and as soon as one of them plays a split they start another game simultaneously from that position only the opponent has to make a different move than the one they did in the new game. Both players must play all of their splits (how many?). I was just trying to come up with beautiful ways of complexifying the game so aliens and androids could enjoy it. Some interstellar tournament.
/Sorry, when you said multiverse and sentient beings it made me happy
//we ought to have a thread about chess and science fiction
:) good idea
About using programs in chess: I discovered this funny guy cheater_1 and his chalegne and game in progres not before today. LOL, what a boyish behaviour: "My dad (program) is stronger then your dad (program)!" 
So much noise around one chess game!
So much noise around one chess game!
:)
the answer is 42.
Ah yes, but nobody know what the question is.
The question is "How many roads must a man walk down?"
No :(
the benjy mouse and the other one made it up [ in one of the "hitchhiker guide to the galaxy" books]
Hi,
assuming both players never making mistakes (= avoid the move which can lead to the loss of the game) then finally we could have only one of following 3 possible recognitions:
a) White always wins
b) Black always wins
c) Each game always ends draw
Personally I think and guess that:
a) has probability 1%
b) has probability 0,1%
c) is most probably true 99%
Furthermore I think that only in case a) or b) we could really say "Chess is solved".
So finally even if we one day could calculate all 10^40 positions, then probably the only recognition will be: if both players are never making mistakes, then each game will end draw.
Günther
actually for two-person games of complete information (like chess) there are 3 basic degrees of solvability (with the obvious intermediate degrees):
1. weakly solved : from initial position game it is known whether the game is a win, draw, or loss for first player.
e.g. nxn Hex is easily seen to be a win for first player, but for large enough n no winning is known strategy (dont feel like describing Hex - pl look it up on wikipedia or academic sites [.ac or such] under combinatorial games / games with complete information). 100x100 Hex is only weakly solved. (even 10x10 iirc)
2. solved: weakly solved, plus method is known and works in reasonable time with present technology for best play by both sides from initial position. [e.g. complete tablebase to initial position has been made.] - Draughts [checkers in the USA] [8x8 board] - has recently been solved.
3 . strongly solved: solved [as in 2] for *any* legally arisible position. [eg if tablebases to *every* legal osition have been constructed].
No program will ever be able to solve chess. Even discarding the sheer capacity of the memory involved, it can't be done. Computers tend to be extremely materialistic. It will either be materialistic or positional with no care for material. There is little in between!
KillaBeez wrote: No program will ever be able to solve chess. Even discarding the sheer capacity of the memory involved, it can't be done. Computers tend to be extremely materialistic. It will either be materialistic or positional with no care for material. There is little in between!
That's just idiotic.
There is now way of telling what technology will be capable of in 50 years or so. So who knows, we'll just have to wait and see what new developments will bring.
However, if chess would ever be solved strongly, in such a way that you could say (as with checkers now for example) 1. e4 draws, 1.d4 loses (just naming a random example), I wouldn't really look forward to that day. That being besides the point.
Offcourse it wouldn't ruin the pleasure of over the board play, just you and another amateur playing eachother, however, it would take away a lot of the romanticism knowing that whatever position you're in, of every move it can be said it wins, draws or loses...
But we are still far away from such a day, so let's just keep on playing and improving ourselves 
I agree with you Sothilde
