Chess Will Never Be Solved. Why?

Sort:
Avatar of Optimissed
playerafar wrote:

There was another forum on this - but it suddenly mysteriously disappeared.
Anyway - there are some scientific papers and studies on this.
For example - it appears that computers have solved and accounted for all legal positions with seven chess pieces or less on board.
But not for 8 pieces or more.

There have been some attempts to scientifically solve from the openings end starting with all 32 pieces on board and in their original positions.  
But with those - all kinds of shortcuts are ventured - some of them apparently not legitimate.  

The jump in difficulty with the addition of just one more piece to 7 to start - with ten piece types to be accounted for for the new piece - and all of up to 57 squares for it to start from - to each of whatever number of billions of positions for 7 pieces  ...  could be Enormous !  

Could perhaps be analagous to comparing the time needed to travel around our star system - with travelling to the nearest other star.

The addition of each piece - multiplying the time needed.
Time to explore this arm of our Galaxy ?  the center?  The other arms ...
Intergalactic travel?  Exploring all galaxies in this Cluster ?

It did disappear. Perhaps someone couldn't control what they were saying.

Chess will never be 100% analysed because it would go on forever and there's no need to do it, since we pretty much know it's a draw with best play.

If it weren't a draw, it would be necessary to find a line leading to a forced win and even that might well be impossible to achieve if we jetison the 50 move rule, which it seems reasonable to do. That's even if such a mysterious line exists, which seems unlikely in the extreme.

Avatar of tygxc

#70
Of course it is a draw.
Of course it does not go on forever: chess is a game with 64 squares and 32 men and has a finite number of positions 3*10^37 and hence needs a finite time 5 years to solve in a finite number of steps.
However, it would change if we had opening books without errors and with exact evaluations draw/win/loss instead of ambiguous +=.
It would also change if grandmasters had a book of 10,000 perfect games all draws with their opening repertoire that they could memorise. Grandmasters would then be sure they had a safe draw as long as they follow one of the perfect games. Grandmasters would then be hinted to look for a win as soon as their opponents deviate from the book of perfect games.

Avatar of Optimissed
tygxc wrote:

#70
Of course it is a draw.
Of course it does not go on forever: chess is a game with 64 squares and 32 men and has a finite number of positions 3*10^37 and hence needs a finite time 5 years to solve in a finite number of steps.
However, it would change if we had opening books without errors and with exact evaluations draw/win/loss instead of ambiguous +=.
It would also change if grandmasters had a book of 10,000 perfect games all draws with their opening repertoire that they could memorise. Grandmasters would then be sure they had a safe draw as long as they follow one of the perfect games. Grandmasters would then be hinted to look for a win as soon as their opponents deviate from the book of perfect games.

You and I agree that chess is definitely a draw with best play. It could be said that we know it's drawn.

I know you've mentioned this figure of five years several times and my reaction is to doubt it. I do not think there is even the necessary software yet.

Whatever the number of possible positions, that number is reduced by an enormous factor if we can throw out all the positions which are clearly not the result of best play. Probably to a very large root of the number given. But how can each position be analysed to a win or draw? Each position links to myriads of other positions and it is sequencing them which provides the difficulty, together with accurate positional assessments. I believe you are greatly over-estimating the speed at which that can be achieved.

If such a job turns out to be a thousand years rather than fiv years at present computer speeds, it would seem as good as impossible and the large, finite numbers would tend towards infinite ones, for practical purposes.

Avatar of Kowarenai

we wont ever have a clear answer and this thread has been done many times, i even made one questioning what will happen if chess actually gets solved, will it die out or not? that thread really became popular and lots of arguments ensued, in general it might get solved one day given our future technology but we might not even be alive by then still i love it

Avatar of tygxc

#72
"It could be said that we know it's drawn."
++ Formally "chess is a draw" is a conjecture believed to be true but not yet proven.
"I know you've mentioned this figure of five years several times and my reaction is to doubt it." ++ The figure of 5 years stems from the late GM Sveshnikov, but my calculation confirms it.
"I do not think there is even the necessary software yet." ++ The necessary software already exists: Stockfish can do it, just like checkers was solved with the checkers program Chinook.
"how can each position be analysed to a win or draw?" ++ By calculating all the way towards the endgame table base. That is also how checkers was solved.


The method I think of is this:
1) Human assistants with ChessBase prepare 26-men starting tabiya to speed up the process.
2) Stockfish plays the position against itself with ample time e.g. 60 h / move until it hits the table base, where it looks up the result and verifies it is a draw.
3) That is not proof yet, it is begin of proof. Stockfish might have made a mistake for white. So we retract the last white move and verify that other candidate moves lead to draws too. Then we retract the next to last white move etc. Ultimately we obtain full proof that chess is a draw.

Avatar of playerafar

Saying one 'knows' its a draw because one wants to know such -
is not 'knowledge'.
Its more like 'wishful thinking'.  A contrived 'default position'.  
But it has a flip side:
Existences of burden of proof (either way or whatever way) of whatever in this context - and its location are also 'in the mind'.  They exist or not according to the inclinations of whoever's mind is involved along with a 'default' position of a draw.

Regarding the disappeared thread - could have been the original poster deleting it.  Or the staff might have done something to it.  

As for the real task - and the algebra involved ...
I'm thinking perhaps it should not begin with the two Kings - nor the number of ways to have at least 32 squares empty.
Possibly - it should begin with the pawns and a 48-board instead of 64.

Avatar of tygxc

#76
We know chess is a draw from evidence: expert opinions by Steinitz, Lasker, Capablanca, Spassky, Fischer, Kramnik... classical game results: 12 draws in 12 classical games Caruana - Carlsen, more draws the higher the rating, more draws the longer the time, 97% draws in ICCF, TCEC needing to impose slightly unbalanced openings to prevent all draws... That is not yet proof, but it is objective evidence.

The paper started with the omnipresent kings in
6*6*(64 - 3*3) + 4*6*(64 - 3*2) + 4*(64 - 2*2) = 3612
possible arrangements and then added the other pieces on the 64 - 2 = 62 vacant squares and pawns on the vacant squares of the 8*6 = 48 available.

Avatar of Optimissed
playerafar wrote:

Saying one 'knows' its a draw because one wants to know such -
is not 'knowledge'.
Its more like 'wishful thinking'.  A contrived 'default position'.  
But it has a flip side:
Existences of burden of proof (either way or whatever way) of whatever in this context - and its location are also 'in the mind'.  They exist or not according to the inclinations of whoever's mind is involved along with a 'default' position of a draw.

Regarding the disappeared thread - could have been the original poster deleting it.  Or the staff might have done something to it.  

As for the real task - and the algebra involved ...
I'm thinking perhaps it should not begin with the two Kings - nor the number of ways to have at least 32 squares empty.
Possibly - it should begin with the pawns and a 48-board instead of 64.

You have no idea what knowledge is. You should have read my comment on it and learned something.

Avatar of Optimissed

You also seem to have no idea about the real nature and meaning of proof.

Avatar of mpaetz

     I see the usual suspects have joined in here. And I just put a comment on the old endless forum so maybe new people will see it and revive it.

Avatar of Optimissed

You think that people can't have known that draughts or checkers was a draw before it was proven. But what if the proof was incorrect and yet it was still really a draw?

Avatar of playerafar

False premise.  Checkers not referred to.
Nice 'red herring' I guess.
One could say 'you think' and then talk about the weather in Australia six years ago -
and then predicate and build and propound ...

Do people go to car races to see the race?
Or to see the accidents ?

Avatar of playerafar

Now ...  chess !!
A game that always has two Kings of opposite colors on the board but never on adjacent squares.
Exceptions?  Yes !   In live speed chess - its 'legal' to take the opponent's King off the board - if the other King is illegally next to it.
Ironic - in that if one player is desperate and needs a 'break' he might move his King adjacent (especially diagonally) which is illegal - hit his clock - and in the intensity of the time scramble the other player 'ignores' and 'hangs his King'  - also hitting his clock.  Fatally.  happy.png
King takes King !  Wins.

Avatar of Optimissed
playerafar wrote:

False premise.  Checkers not referred to.
Nice 'red herring' I guess.
One could say 'you think' and then talk about the weather in Australia six years ago -
and then predicate and build and propound ...

Do people go to car races to see the race?
Or to see the accidents ?

You know, you wouldn't know a false premise if one slapped you in the belly like a wet herring. (Slight adaptation of common but old-fashioned British saying). FYI, the game of draughts was used as a good example of something that was known before it was proven. A bit like the fact that the World is roughly spherical and you can sail right round it. The Phoenicians and Vikings obviously knew that but they probably couldn't prove it.

Actually, looking at your use of the word "premise", you don't know what one is and that isn't a premise. It's an observation.

Avatar of playerafar

Checkers.  False premise.  Stands.
Chess - is not checkers.

Regarding what happens in these forums about 'solving' chess -
if one person can get away with personalizing whoever disagrees with him in the forum - and isn't blocked by the opening poster -
and he posts enough and provokes enough response -
then various people in the forum might get the idea that the forum is about him.
And therefore that when he isn't posting - there's no 'drama' so therefore they don't post.  
The forum becomes 'conditioned'.

Issue:  when subjects are discussed on 'social media' - are the discussions about the subjects or ...  the people in them ?
Well there's no way to avoid the latter.  Its always there.
But how much?   
Depends on how the opening poster uses his blocking options.  Or not.

Avatar of Optimissed

And you accuse others of "projection"! happy.png
That was an interjection!

Avatar of playerafar
tygxc wrote:

#61
Yes, it takes 5 years on the Sesse computer for 1 ECO code.

That doesn't mean it has 'solved' that ECO code.
(is it an 'easy' code?  Did somebody play the Duras gambit?)  
Did the computer 'dismiss' various ensuing positions and options ?

Avatar of playerafar
tygxc wrote:

Can humans walk on Mars? - Yes.
Will humans walk on Mars? - Maybe, it depends on somebody authorizing some billions to build and launch a suitable spacecraft.

Can chess be 100% analysed? - Yes.
Will chess be 100% analysed? - Maybe, it depends on somebody paying some millions for 5 years of human assistants and modern computers.

Disagree - either they'll have to soup up the hardware/software/programming many millions of percent -
or the human race will have to survive millions of years ...
or both.
They could have solved chess on another planet in some other galaxy though.  Or maybe even in this Galaxy !
But their chess might have different rules though !   
Like - for castling - your rook can't castle out of or through attack - which it currently can in long castling.  
3d chess ?  Maybe that'll be on 'phones' in 100 years ...

Avatar of not_cl0ud
Optimissed wrote:

And you accuse others of "projection"!
That was an interjection!

lol

Avatar of not_cl0ud

I really dk about this forum

someone said in another forum to want to be more like magnets lol XD