Uh, yeah, I need some time to digest all that
I think your saying that this solution would only be aplicable if there was a way to force a queen exchange (which I doubt there is a way to do that)?
I like most of what you said though
Will computers ever solve chess?
Sort:
Just to be clear, I was not saying I had solved chess. My queen exchange was just to make it easier to understand how such a solution might happen, but queens don't need to exchange (maybe there is a forced repetition on certain paths or something else).
I was just trying to point out that solving it has nothing to do with huge number of different positions possible, and that posting that there are more possible positions than the envelope calculation of 10^80 particles in the universe does not mean it could never be solved. The information content of a solution can be very manageable in comparison to the size of the game space. That happens all the time in game theory. The solution will probably be pretty complex in terms of human calculation, simply because no one has found such a strategy yet and we have had some amazing players over it's history. But it's possible it could be understood once found and simplified some, and it is even possible people have missed some very simple strategy simply because people don't normally try for draws as white. It could be so complex only our largest computers today could play perfectly, but that is still universes of magnitude smaller than the full game space.
I like how a font change makes the spacing weird. I'll have to remember that.
If chess will ever be solved it will be by some trying. Here is a step in that direction. https://www.kickstarter.com/projects/1239228060/solving-chess-yes-solving-chess-step-1-of-3?ref=discovery
Apr 6, 2014
0
#66
bobbyDK wrote:
Hadron skrev:
Let me ask you this, even if someone announced tomorrow that chess was finally and completely solved....Would you seriously give chess up ?...If you answer yes, your too easily impressed...and if you answer no, ask yourself this: Why did you bother to leave a post?....and why are you tracking this?...and just what is the point of solving a game anyway? It is like pulling the wings of a dying butterfly you find...You do it because you can to make yourself feel omnipotent while everyone else thinks your just one sad tosser with way to much time on your hands.
I would still play chess and why because I am interested in it in a scientific matter: How much is computer able to do.
you don't have all the answers as it sounds like you think you have.
Oh dear. I don’t know where you came about the conclusion that I believe I have any sort of answers on the subject of computers eventually solving chess from one major open ended question. I really have no opinion on if computers will ever solve chess. The thread at hand as much to do about philosophy as it does the wanton application of (computer) science on a such an endeavour and my questions are in line with why exactly anyone would want to do it.
I am just at a loss as to the point of solving chess and I really at a loss why on earth anyone would carry on playing a game if the result is predetermined to those with the information at hand (and as far as anyone may know, if chess is ever solved in the future that well may be you).
Personal scientific curiosity? Really?
Thank you
Apr 6, 2014
0
#67
answer to question..yes, and who cares? Cars are faster than people. Move on. Enjoy the game.
Apr 6, 2014
0
#68
This one is easy to answer, and based on the current progress toward solving endgame tablebases, it's even possible to work out about how long it will be until chess is solved. This calculation takes into account the increasing difficulty each piece adds to the problem, and how long it has taken for each step from 3 piece endgames up to 7 piece endgames.
Based on that progression, I've seen calculations that suggest chess will be solved in about 220 years.
For now, we can enjoy the game ourselves and not worry about when it will be solved. It will be solved, but not in our lifetimes.
I have my computer working on it, I bet I will solve it before this thread ends.
Apr 6, 2014
0
#70
Thank you for your post. If I may:
>curiosity is mutual here...what prompted you to ask why? Or even more so, what precipitated your strange bizarre apropos take on the topic?<
Please refer to my post above to the other gentleman. I will reiterate it here the thread at hand as much to do about philosophy as it does the wanton application of (computer) science on a such an endeavour and my questions are in line with exactly why anyone would want to be bothered with trying to solve chess.
>Go back to pulling the wings off the butterflies you encounter my or admiring your insane avatar which only proves the point I'm making regarding you.<
What is your point? That you are a self-righteous judgmental pseudo-intellectual arse? Maybe? Maybe not? But you're content to try and judge me on an avatar based on a classic American movie monster when yours is some homo-erotic image of someone in leather chaps and a mask. Mind you, your avatar could just be a Roman Legionnaire but I guess it is how judgmental you are.
>Chess being solved is very intriguing to most with more than one or two neurons connecting in their brain unlike certain ignoramus souls out here or there<
How wonderfully myopic. I did not say it was not intriguing, I just questioned the point of doing so. With so many mysteries out there, notwithstanding scientific and medical you are happy to see time (and cost) spent on the solving of a game, intriguing or not? Sure the possibility is intriguing but is that really a point enough to actually do it or waste time doing so? I don’t think you should be looking to call anyone a ignoramus if one is so unwilling to look at the bigger picture.
>...the fact that there is more chess positions then there are atoms in our milky way galaxy is absolutely mindboggling to most thoughtful intellects<
I really do not what to say. It is fact that there are more chess position than in our galaxy? I would really like to know your source because it would have to know beyond any doubt just what exists beyond the limits of mankind’s abilities to investigate such as the number of planetary bodies and their exact dimensions and or whether or not life of any kind exists out on or between those planetary bodies. I suspect you are try to purport a mathematical supposition as fact
>but unfortunately our species is filled with liabilities to our continued survival like dogmatic simpletons with more time on their hands then they know what to do with so they self destruct and attempting to take all with them<
Finally, this comment is very pertinent don’t you think? It rather describes those to a Tee who place all their faith in the dogma of science without first ever considering the bigger picture and the real need to do what is being and or about to be done. As far as it appears as you are concerned, the means does really justify the end. If that is not dragging us towards destruction, I do not know what is.
>Go to sit on a couch with a professional who can shrink it down for you my friend<
Sit down while I tell you a tale……First the Flat Earth Society……then Scientology…..and now…You.
Only a computer that has completely solved chess would stand a chance of getting a draw against one such as I.
This often posted debate always makes me chuckle.
Its guaranteed to throw up the usual nuggets of wisdom such as more positional possibilities than atoms in a ham sandwich , storage capacities of computers i've never heard of and a whole fistful of psuedo scientific ramblings.
why would anyone want to see chess solved ?
because the minute that happens a certain amount of magic & mystery will be lost forever
its just a board game between 2 people , thats it really.
Humans will enter their historic games until eventually no one will come close to beating a computer. Theoretically if they play perfect humans can draw with a computer. Remember though, computers are only parroting moves and lines of play discovered by humans. It's a combined knowledge thing with a no errors, mistakes or blunders during a game! And as computers get more cores and faster any time limits will be in their favor as well. Use the computer to analize and teach chess wouldja please?
Apr 7, 2014
0
#74
I think chess computers will only store winning or drawing moves.
if a move is not winning or drawining it will not be in the database.
thus making it possible to store all moves needed cause you only record winning or drawing variations from whites point of view. if black should win with perfect play no moves will be recorded.
I think chess will be solved 100% if we do it this way as it is pointless to store losing variations as no computer will play a losing move.
if the computer during running through the variations finds out the variation is losing it will not store that move/variation.
if you want to analyze a position that isn't winning you will still have to bruteforce the position.
King's gambit accepted has been solved:
http://en.chessbase.com/post/rajlich-busting-the-king-s-gambit-this-time-for-sure
I don't know if any other openings have been. They didn't actually go through very possible variation, though. It's pretty neat.
Jan 3, 2015
0
#76
Did you notice the date of that article? It's rather critical to understanding it.
That was an april fools joke :\
Computers will never solve practical chess.
elephant32 wrote:
King's gambit accepted has been solved:
http://en.chessbase.com/post/rajlich-busting-the-king-s-gambit-this-time-for-sure
I don't know if any other openings have been. They didn't actually go through very possible variation, though. It's pretty neat.
Never say never but they haven't for the time being. I am not sure why that should remove all the fun out of the game, I think it would remain a very complex game nonethless and unless you are a super-human, that won't change a damn thing when you face your flesh and blood opponent.
Jan 4, 2015
0
#80
to be able to record all moves it should only contain winning or drawing moves for white. if a move leads to a forced losing x moves later.
no need to record the forcing moves. just mark the first move as bad. similiar if there is a win or draw x moves later no need to record the forcing moves just mark the first move as draw or win.
I wanted to add one more description to illustrate how chess could be solved, because I think it shows really well why the talk about the complexity of the game is completely misguided. This is a "what if" game, but that's what we're playing here.
What if we find out that there is a way to force the early exchange of queens with white in one of the 1.e4, ... 2. Qf3 openings (like the Napolean). These are oddballs, certainly, but lets say there is a strategy that can force an equal queen exchange pretty much whatever black plays. Black doesn't have to play best play, and neither does white (white can skip forced mates, for instance). But even if black plays best, the queens will come off equally. This isn't an outrageous scenario, clearly some openings like the Berlin have common queen exchanges.
And then say that the pawn structure after the exchange can be in one of 23 different patterns, and it is found how to force off at least 4 pawns from both sides in the course of the opening using a certain pawn advancement strategy. The strategy ensures that this is at least an equal exchange for white as well.
There are already some well-known ways to force piece exchanges (pins, forks, skewers, etc.). Let's say this is plugged in and we can always ensure at least 3 of the knights and bishops be taken off and at least one rook.
The hypothetical is that this can be pretty much forced equal, but if black doesn't want to take any pieces or put itself into forced mate, that is fine. It doesn't have to be equal. Even forced stalemates are fine here. But with the strategy, there is no way that black can checkmate white while these exchanges are happening.
Where this is leading to is what happens if we can prove a trade down to a very simplified scenario like maybe a rook, a piece, and 4 pawns, and we know the 4 pawns can be made to be on some collection of possible files and suddenly the position is calculable with a good computer. And now we show that all possible games are at worst a draw for white with best play.
And let's say sometime shortly after this last part is proved, someone shows that black can force the trade down of pieces to with his own strategy and can also force a draw at worst.
The point here is that all of this has now proven that chess is a forced draw. And at no point were all moves ever possible needed for the calculation. What was needed was only a particular move strategy that one could prove traded down pieces at least equally to the point where a brute force could prove forced draw. The move strategy determines one line for white that depends only on the black moves and doesn't need to figure out even every possible move black could play - it could be purely something like "if black moves into one of these pawn structures, counter with this pawn structure" and "in pawn structures X, Y, and Z, if one of blacks pieces protect f4, play the same type of piece to f4". General rules that someone proves are sufficient to force some trades or draw in 50 moves.
This kind of proof ignores nearly all possible lines of chess. What if white doesn't want to play 2. Qf3? Who cares! What if there are some forced wins in some lines for white? Who cares! What is shown is that white can force a draw (or an incidental win) if they make these specific kinds of moves and we'll ignore all the alternates. And so can black, with these other specific moves, even if white isn't playing this strategy themselves. So chess is a forced draw with what could now legitimately be called "best" play.
I think this is unsatisfactory for some because playing for a draw (at least as white) is usually seen as bad sportsmanship. Also, using a strategy and proving only whether that can force equality instead of looking at all the sidelines that might force a win seems to ignore the whole point of the game.
But if that kind of proof existed, for both sides, then chess would be solved. One could always force at worst a draw in every game, no matter which side you played.
I hope this hypothetical shows how the full complexity really is irrelevant to the solution problem. Proofs on strategies and the piece patterns they create do not need to calculate every possibility. Strategies might be shown to simplify the problem space to a handful of possible outcomes which then could be easily calculated or possibly proven independently. A universe of complexity may be ignored and only a manageable grain of sand remain.
You have an assumption that all the non-queen exchange paths are explored. In order for that to happen, chess would have to be completely explored. You cannot just assume best play up to the queen exchanges because best play is precisely whhat we mean when we say we have chess solved. How do you know that a move leading away from a queen exhange is not "best play"?
It is not necessary to calculate all paths in a strategy analysis. That was the whol point of my earlier comment on how a king and queen versus king endgame has been solved, despite there being a huge number (10^69) of possible play sequences where 50 moves are spent chasing the king around without mate. No matter which of the initial 960,000+ possible non-stalemated initial positions you start with, there is a forced mate by the side with the queen, which was proven (mathematically, rigorously) long before computers could brute force it. The proof doesn't involve knowing best play for the lone king side. It only speaks of restricting squares and forcing mating patterns.
Similarly, if white has a given strategy for the opening that forces a queen exchange or mates, that is all that needs to be proven in that hypothetical analysis. The whole point is that you don't need to find best play. Chess can be solved if there is a way for both sides to force at worst a draw, whether or not they pass up a million mates along the way. I gave one hypothetical path as an example, but the underlying point is not hypothetical:
The point is: it is completely irrelevant how complex the whole game of chess is to whether or not it will be solved. All that is need are proofs on given strategies.
I gave earlier literature references and an example on INFINITE games that can be completely solved with finite information.