Chess solved?

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
duncan
Posts: 12032
Joined: Mon Jul 07, 2008 8:50 pm

Re: Chess solved?

Post by duncan » Mon Aug 24, 2020 9:19 am

mwyoung wrote:
Sun Aug 23, 2020 11:05 pm

Yes, you can have strong evidence, and still be wrong. Lets go back to a physics allegory. Isaac Newton theory of gravity was thought to be correct. As it was proven right, it could predict planet motions, comets, and other large scale objects. Then Albert Einstein said but yes, but it can not explain the motion of the planet Mercury. Einstein found a hole, and it only takes one to prove you wrong. Now we have found holes, in Albert Einstein theory, as it only takes one hole in your theory to show you only have an approximation, not proof.
This is a bit off topic but I believe Einstein's special theory of relativity, was proven by the motion of the planet Mercury.
However Einstein's special theory of relativity, did not falsify newton's theory of gravity , it was only Einstein's general theory of relativity which falsified Newton's theory of gravity and this was to come later.

mwyoung
Posts: 2727
Joined: Wed May 12, 2010 8:00 pm

Re: Chess solved?

Post by mwyoung » Mon Aug 24, 2020 9:43 am

duncan wrote:
Mon Aug 24, 2020 9:19 am
mwyoung wrote:
Sun Aug 23, 2020 11:05 pm

Yes, you can have strong evidence, and still be wrong. Lets go back to a physics allegory. Isaac Newton theory of gravity was thought to be correct. As it was proven right, it could predict planet motions, comets, and other large scale objects. Then Albert Einstein said but yes, but it can not explain the motion of the planet Mercury. Einstein found a hole, and it only takes one to prove you wrong. Now we have found holes, in Albert Einstein theory, as it only takes one hole in your theory to show you only have an approximation, not proof.
This is a bit off topic but I believe Einstein's special theory of relativity, was proven by the motion of the planet Mercury.
However Einstein's special theory of relativity, did not falsify newton's theory of gravity , it was only Einstein's general theory of relativity which falsified Newton's theory of gravity and this was to come later.
No, General Relativity deals with curved space time. And was needed to explain Mercury's orbit. a 43'' per century precession of the perihelion of Mercury. Special Relativity applies to all physical phenomena with a flat space time. Mercury sits in the curved space time of the Sun.
"The worst thing that can happen to a forum is a running wild attacking moderator(HGM) who is not corrected by the community." - Ed Schröder
But my words like silent raindrops fell. And echoed in the wells of silence.

duncan
Posts: 12032
Joined: Mon Jul 07, 2008 8:50 pm

Re: Chess solved?

Post by duncan » Mon Aug 24, 2020 10:45 am

mwyoung wrote:
Mon Aug 24, 2020 9:43 am
duncan wrote:
Mon Aug 24, 2020 9:19 am
mwyoung wrote:
Sun Aug 23, 2020 11:05 pm

Yes, you can have strong evidence, and still be wrong. Lets go back to a physics allegory. Isaac Newton theory of gravity was thought to be correct. As it was proven right, it could predict planet motions, comets, and other large scale objects. Then Albert Einstein said but yes, but it can not explain the motion of the planet Mercury. Einstein found a hole, and it only takes one to prove you wrong. Now we have found holes, in Albert Einstein theory, as it only takes one hole in your theory to show you only have an approximation, not proof.
This is a bit off topic but I believe Einstein's special theory of relativity, was proven by the motion of the planet Mercury.
However Einstein's special theory of relativity, did not falsify newton's theory of gravity , it was only Einstein's general theory of relativity which falsified Newton's theory of gravity and this was to come later.
No, General Relativity deals with curved space time. And was needed to explain Mercury's orbit. a 43'' per century precession of the perihelion of Mercury. Special Relativity applies to all physical phenomena with a flat space time. Mercury sits in the curved space time of the Sun.
Thanks for your correction.

lkaufman
Posts: 4575
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: Chess solved?

Post by lkaufman » Mon Aug 24, 2020 5:35 pm

jmartus wrote:
Sun Aug 23, 2020 7:10 pm
How many more elo until chesd considered solved?
It is possible that it is already weakly solved: Suppose we put the latest SFNNUE on a very powerful computer, give it a serious time control (maybe 2 hours plus one min inc or so), and the best/safest opening book available (people will argue over what that is of course!). Then see if any opponent, either the same or different engine on any hardware and with any book or even human assistance, can win even one game in a thousand from it. Of course this wouldn't be a "proof" that it can't lose, but no one would be able to prove that it hadn't solved chess weakly. I'm not saying that we are already at that point, only that it is possible that we are at that point now. Based on the high draw rates even with variety opening books, I can imagine that with a book that always plays the Berlin when possible and other draw-seeking defenses to other openings, it might never lose, or at least not within any number of games that can be practically tested. It could play Black every game to cut the number of games needed in half, since it is far easier to draw with White.
Komodo rules!

syzygy
Posts: 5000
Joined: Tue Feb 28, 2012 10:56 pm

Re: Chess solved?

Post by syzygy » Mon Aug 24, 2020 5:52 pm

towforce wrote:
Sun Aug 23, 2020 9:03 pm
A mathematician might create an expression that quickly and correctly categorises a position as win/draw/lose. Failing that, it's a case of watching the elo v draw ratio chart.

It's surprising that a game designed so long ago has held out this long (but go will outlast chess).
A plain minimax algorithm is such an expression. Since chess is a finite game, the algorithm runs in O(1) and is therefore very quick. The same applies to Go.

syzygy
Posts: 5000
Joined: Tue Feb 28, 2012 10:56 pm

Re: Chess solved?

Post by syzygy » Mon Aug 24, 2020 5:56 pm

Vinvin wrote:
Sun Aug 23, 2020 10:43 pm
You make an error of logic.
A million (or a billion or a sextillion ...) of successive draw games can proof 2 different things :
1) the game is proof as draw.
2) the player didn't find the forced winning line(s).
Nothing less, nothing more.

Perfect plays is far away of tests and experiences.
A lot of people have trouble to understand that.
There are very good chances that the perfect game is 1.e4.
I am less certain about the outcome of this perfect game, i.e. whether black resigns or the players agree on a draw.

Dann Corbit
Posts: 12149
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Chess solved?

Post by Dann Corbit » Mon Aug 24, 2020 6:25 pm

1.e4 has the most mobility of any opening move.
1.d4 has been given a very good case by Berliner.
1.c4 was the result of a 70 ply SF search.
I am not as sure as you are about 1.e4
I also think that a black win as the outcome is not impossible, just very unlikely.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

nnnnnnnn
Posts: 26
Joined: Mon Nov 18, 2019 1:36 pm
Full name: Mark Thellen

Re: Chess solved?

Post by nnnnnnnn » Mon Aug 24, 2020 6:37 pm

Consider the initial position.

Remove black's queen.

Is there a proof that black does not have a forced win?

RogerC
Posts: 35
Joined: Tue Oct 29, 2019 7:33 pm
Location: French Polynesia
Full name: Roger C.

Re: Chess solved?

Post by RogerC » Mon Aug 24, 2020 6:47 pm

duncan wrote:
Sun Aug 23, 2020 9:51 pm
jmartus wrote:
Sun Aug 23, 2020 7:10 pm
How many more elo until chesd considered solved?
I think Kai thought 400 using some charts till you get all draws (but not solved) Best ask him though.
Yes it's an evidence ! As soon as we "Solve" chess that would mean you have evaluated/calculated every "best of the best" move for each position on each side. This would inevitably and ultimately lead to a draw and also the same game, again and again ! And it would mean you have 120-130 "best of the best" depth positions and each side playing the same game, again and again ! That would also mean that if you don't play "the best of the best" move on a specific position, you will lose.

Solving Chess = Boring games.

duncan
Posts: 12032
Joined: Mon Jul 07, 2008 8:50 pm

Re: Chess solved?

Post by duncan » Mon Aug 24, 2020 7:00 pm

RogerC wrote:
Mon Aug 24, 2020 6:47 pm
duncan wrote:
Sun Aug 23, 2020 9:51 pm
jmartus wrote:
Sun Aug 23, 2020 7:10 pm
How many more elo until chesd considered solved?
I think Kai thought 400 using some charts till you get all draws (but not solved) Best ask him though.
Yes it's an evidence ! As soon as we "Solve" chess that would mean you have evaluated/calculated every "best of the best" move for each position on each side. This would inevitably and ultimately lead to a draw and also the same game, again and again ! And it would mean you have 120-130 "best of the best" depth positions and each side playing the same game, again and again ! That would also mean that if you don't play "the best of the best" move on a specific position, you will lose.

Solving Chess = Boring games.
Imo it would be evidence but not proof.

Why do you have to play only one "best of the best" move on a specific position. There could be a number best of the best moves which draw?

Post Reply