Chess will never be solved, here's why

Sort:
tygxc

@9589

"'weakly solved' ... what is that like?"
++ Weakly solved means that for the initial position a strategy has been determined
to achieve the game-theoretic value against any opposition.
Games solved: Now and in the future

Likewise Schaeffer weakly solved Checkers: it is a draw.

Now for Chess for the initial position a strategy to achieve the game-theoretic value of the draw against any opposition is to follow an ICCF World Championship Finals draw for as long as possible and then proceed with an ICCF (grand)master and an engine at 5 days/move.

Elroch
tygxc wrote:

@9589

"'weakly solved' ... what is that like?"
++ Weakly solved means that for the initial position a strategy has been determined
to achieve the game-theoretic value against any opposition.
Games solved: Now and in the future

Likewise Schaeffer weakly solved Checkers: it is a draw.

Now for Chess for the initial position a strategy to achieve the game-theoretic value of the draw against any opposition is to follow an ICCF World Championship Finals draw for as long as possible and then proceed with an ICCF (grand)master and an engine at 5 days/move.

From many of your posts I have determined that you don't understand what weakly solved means.

It is necessary to deal with ALL LEGAL RESPONSES to a specific strategy, not just "several" responses.

Elroch
playerafar wrote:

That's pretty good sarcasm.
The forum could be used for 'sarcasm workout' ...
'weakly solved' ... what is that like?

That is providing a strategy for each player that is proven to achieve the optimum result against all legal responses. (One strategy is trivial to define if the other player has a win).

How about 'squaring the circle'?
Is it weakly solved?

The definition does not apply. It is proven that you can't do it with compass and straight edge constructions alone. Compass and straight edge constructions are a class of constructions of new points in the plane from a given set of points in the plane.

ok, I realise you were jesting...

Elroch

O: "T, you haven't proven the twin prime conjecture"

T: "You haven't disproved it, therefore I have proven it"

[Same reasoning as part of the discussion above].

playerafar

@Elroch
Yes - we're having a sarcasm-fest.
happy

playerafar

I'm posting this video about stalemate ...
yes it might not be relevant even at all.
If not - mea culpa.
https://www.youtube.com/watch?v=cqFykCZ4I34happy might put a 'wrensch' in things ...

playerafar

"The remainder descended further still into personalised dribble. Did you come here to show off your accomplishments? Is so it would have been better had you not behaved like a troll in your own threads for years by attacking those disagreeing with you. You are not an exceptionally intelligent person and no-one need take you the slightest bit seriously at all, except maybe those who feed off you. You are VERY much out of your league and I think sometimes that realisation dawns on you."

The 'O' guy talking about himself again while trying to pretend its about somebody else.
He winds up being 'below' everyone else's 'league'.
His unhappy masochism and 'little thunder'.
He's still foolishly furious that Elroch blocked him.
Dio also blocked him too. Long time ago.
He'll get his 'little anger' up about that too.
He's in his cycle ...

Elroch
Optimissed wrote:
Elroch wrote:
Optimissed wrote:
MEGACHE3SE wrote:

"++ Chess is known to be a draw." known, but not proven. ask any mathematician.

A scientist would give a different answer and it's scientists we have to trust.

To be fair, a plumber might provide a third answer and an orthopedic surgeon a fourth. If they didn't tell you to buzz off and stop asking silly questions.

Chess is NOT within the domain of science. If scientific methods are applied to chess (as they can sometimes be applied to topics in the mathematical and computational sciences) they never involve PROVING anything (except in the trivial case where an unambiguous example is exhibited - eg A: "prove tigers exist" ... B: "here is my pet tiger". With reasonable assumption, exhibition of an example is where the scientific method and the deductive method overlap in a rather trivial way.

You can emphasise all you like that chess is NOT within the realm of science. If it isn't, it must be in fairlyland, since chess isn't in the realm of mathematics. My son informs me that chess cannot be represented mathematically and that it probably never will be. He thinks it's impossible.

Seriously? I have previously got the impression that your son was bright but what you say undermines that. I believe you are grossly misrepresenting his position. I believe he would instead deny that chess can be dealt with mathematically in a way which allows results to be proven that would require a great deal of brute force otherwise. This is not the same thing at all!

It is not only possible to represent chess mathematically, it is easy and it is extremely well-known that chess is a well-defined example of a finite combinatorial game of perfect information, dealt with in general in game theory.
Here's a non-trivial example of game theory applied to chess.

playerafar

"It is not only possible to represent chess mathematically, it is easy and it is extremely well-known that chess is a well-defined example of a finite combinatorial game of perfect information"
Elroch is correct.
I was suggesting something similiar.
Chess is close enough to mathematics including mathematics referring to itself.
'Perfect' information.
Absolutely. Perfectly.
But 'solving' is something else again.
The game is Big. Mathematically Big.
Too Big. When you're talking about 'solving' in a total sense.
--------------------------
There are problems in Mechanics that are very difficult to solve perfectly.
Even with 'perfect' information to start.
Like for example four 'perfect' equal masses interacting gravitationally in three dimensions where you know their mass - radius - starting 3d coordinates - starting velocity (which is more than 'speed' - includes the 3d vector) - perfect vaccuum - no other influences .... can the future positions of the four masses be calculated at the end of time t?
if they were approaching each other in 2d at perfect right angles to each other - probably.
But not the case in the exercise where they're coming in obliquely to each other ... no neat angles and in three dimensions. No 'coplanar'.
Can you 'solve' for time t?
Can you calculate if any will collide with each other in the first 'pass'?
Or will they pass each other and then move away?
Or be drawn back in?
Point: even with 'perfect information' I think you'd find that many of these can't be solved.
Elroch probably knows a great deal about that exercise - maybe even computers making 'extrapolations'.
Even with just two masses approaching each other obliquely -'aiming' above and below each other - the lines defining their trajectories are not straight lines - they're 3d curves. Tough!
Each curve would not fit in any plane.
---------------------------------
'the guy' mainly personalizes. 'His son.' 'You'. 'IQ'. 'League'. His obvious narcissism. He doesn't really discuss it.

playerafar

There are many things that might never be 'solved'.
Some will say 'never say never'.
Some things can be proven to have no solution ...
like 'solve for the greatest prime number' ...
there isn't one.
But it is known how many seconds there are in a hundred year lifespan.
There is a first upper bound on the number of possible chess positions.
Its 13 to the 64th power.
That can be cut down some - but not enough ...
at the end - there's too many positions left and too much time to solve each of them.
And starting from 'the end' with two kings and adding - doesn't work either because the task gets too daunting even after adding a ninth piece to the board.
Maybe they have calculated or extrapolated how long it will take the computers to solve for all positions with nine pieces.
There's probably a graph - displaying how the needed total solution time increases with the addition of any of ten types of piece.
Say they can solve for nine pieces in ten years from now ...
does that mean it'll take 100 years when they add that tenth piece?
If there's a tenfold time increase with each addition of a piece ...
but maybe that goes up too.

playerafar

tygxc and Elroch do a lot better than 'O'.
But almost everybody does.
'O' is foolishly sensitive and insecure.
Imagines and craves authority he'll never have.
It means he always loses. Every time.

BigChessplayer665
playerafar wrote:

tygxc and Elroch do a lot better than 'O'.
But almost everybody does.
'O' is foolishly sensitive and insecure.
Imagines and craves authority he'll never have.
It means he always loses. Every time.

the best argument is a logical one not me bully opponent with a poem if your going to insult people being foolish back doesnt help

yes I'm fully aware that you probably will start insulting me to but I'd rather take the hit than @ Optimissed

DiogenesDue
tygxc wrote:

"Chess engines are not "god", they cannot play perfect chess."
++ ICCF (grand)masters with engines and at 5 days / move average have now reached perfection, they can now draw against any present or future superior being.

A false claim.

DiogenesDue
tygxc wrote:

"Optimal play is not determinable by either you, or engines."
++ The 105 ICCF World Championship games are optimal play, not by the engines used,
nor by me, but by the results themselves: 105 draws in 105 games.

105 or 10005 draws, makes no difference. It does not and *can not* prove perfect play when the players are demonstrably imperfect. You can have 100 GMs and engines confer for 5 days or 500 days...the result is still not perfect play.

"It will be determinable once chess is solved" ++ That is where we now about are.
Weakly solved means that for the initial position a strategy has been determined to achieve the game theoretic value against any opposition. The strategy to achieve the game theoretic value of the draw against any opposition is to follow an ICCF World Championship Finals drawn game for as long as possible and then proceed with an engine at 5 days / move until a 7-men endgame table base draw or a prior 3-fold repetition is reached.

I gave the actual definition of Weakly Solved. I don't really care what Herik says.

Your "game theoretic value" hedge is just there to allow you to pretend that a draw is the default result and that a forced win would be an aberration.

Elroch
Optimissed wrote:

chess isn't a game of perfect information

The only explanation of making a statement that is simply wrong is that you are using that technical phrase without having learnt what it means.

If you take the time to learn what the phrase means, you will discover that for chess it means that both players are fully aware of the moves that have been played so far (and which determine precisely what legal continuations are possible).

Elroch
Optimissed wrote:
Elroch wrote:
Optimissed wrote:

chess isn't a game of perfect information

The only explanation of making a statement that is simply wrong is that you are using that technical phrase without having learnt what it means.

If you take the time to learn what the phrase means, you will discover that for chess it means that both players are fully aware of the moves that have been played so far (and which determine precisely what legal continuations are possible).

I understand what it means and I know that there are those who agree with me and disagree with you, just as there are those who agree with you and disagree with me.

No.

Your statement is as wrong as saying "hydrogen is not an element" or "electrons are not fermions". Only possible by being ignorant of the subject (in our case game theory rather than chemistry and physics) and its terminology.

Search for the term and you aren't going to find support.

DiogenesDue
Optimissed wrote:

Thank heavens something I can disagree with. I don't rate Herik either so I don't care what he says. I don't like the hiding behind jargon like "game-theoretic value" because it seems deliberately misleading. We know it's a game and yes positions can be evaluated but not by theory so it's misleading. "Positional evaluation" is one of many much more descriptive tries. I just prefer "evaluation".

However, it seems that since there's no proof that chess is anything but drawn and since it seems to be the case that it's drawn, is it then so bad to allow "draw with good moves by each side" to be the default or probable value, subject to further confirmation?

There's a difference between default and probable value. It's fine to say that chess being a draw is more likely. It is incorrect to say that a draw is the "correct" result and that a forced win would be an unexpected and suspect surprise. If a set of forced wins exist, they will be necessarily narrow and probably dependent on some heretofore undiscovered truths about chess in terms of positional play, much like when Alpha Zero wowed all the super-GMs with some fancy bishop tactics, etc.

Tygxc's premise is that even though machine learning engines just hit the scene (and engines only just surpassed humans circa 2006), that the possibilities of chess have been exhausted, ergo all draws, ergo chess itself being a draw. This is no different than the best players in 1900 saying the same thing. Chess is not exhausted, and we've only played some infinitesimal percentage of possible games. What has happened is that engines have reached a plateau for the moment, and humans are no longer capable of helping out. We're obsolete.

If anyone besides Tygxc actually believes that engines 10, or 20, or 50+ years from now will not be capable enough to beat today's ICCF "centaurs" (more like hydras now, given the number of engines being used together), go ahead and speak up...

Elroch

It's an interesting question how badly they will be able to beat them!

DiogenesDue
Optimissed wrote:

Now, the coffee's been served but before I look at the given definition, I want to make a rhetorical arguments of "if you don't know what the information means, how can it be perfect or full".

If we imagine an analogy of a chess position in the form of a code which hasn't been broken. The earth's being attacked by Martians and we're able to pick up the coded message. But how have we got perfect information if we don't know what it means?

So similarly, a complex chess position. It may be a win for either side or perhaps a draw. We can see the position clearly but we are not capable of determining an evaluation. It's the same thing. We have the information but it's given in a code which we can't interpret. We can see the chess position but we can't tell who's winning. We can only giess, based on rules of thumb. How is that perfect information? If I had perfect information on something, I'd expect it to be understandable.

Your premise that is that you are capable of understanding any information given to you. Reams of evidence on these forums would show this to be demonstrably false. Don't feel bad, though...it's false for any and all human beings.

BigChessplayer665

I don't think that's entirely true I think people try to understand they just don't do a very good job of it