Chess Will Never Be Solved. Why?

#243
In solving chess the position is looked up in the 7 men endgame table base once 7 men are reached so the 50-moves rule plays no role in solving chess.
Think a second before you mistakenly accuse me of being mistaken.

In solving chess the position is looked up in the 7 men endgame table base once 7 men are reached so the 50-moves rule plays no role in solving chess.
Doesn't follow.
It is you who needs to think.

#245
During solving chess the 50-moves rule is never invoked.
Once 7 men are reached the result is just looked up in the endgame table base.
You need to think if you can.

Here's another one of your recent posts -
"No, it plays no role at all. In no GM game or ICCF WC game is the 50 moves rule ever involved with >"
I am not mistaken.  You made that post.  You've been caught.
Again.  

#247
I posted:
"++ No, it plays no role at all. In no GM game or ICCF WC game is the 50 moves rule ever involved with > 7 men. ICCF WC games end on average at move 39. It is just a practical rule to avoid a game and thus a whole tournament to drag on for weeks e.g. as a weak player cannot checkmate KBN vs. K."
This is absolutely correct. Can you read? Can you think? 

And yet again you're Caught.
This time in a self-Contradiction.  Just now.
 "No, it plays no role at all. In no GM game or ICCF WC game is the 50 moves rule ever involved with > 7 men."

First - 'no role at all'.  Then - 'with > 7 men'
Contradiction.
I'm reasoning with you that you're making mistakes and being caught doing so 
but you're responding with personal attack.
You've admitted that nobody has been able to reason with you too.
Suggestion:  If you keep believing you're always right - that's not good.
You could try to follow your own idea for a change and try harder to understand that you're making mistakes.

#249
There is no contradiction. The 50 moves rule plays no role at all in solving chess.
In GM games or in ICCF correspondence the 50 moves rule is never invoked with > 7 men.
In GM games they play on and with <= 7 men the 50 moves rule occasionally is invoked.
In ICCF correspondence they can just claim a draw or win based on the 7-men endgame table base even if the win exceeds 50 moves without capture or pawn move.
In real ICCF games table base draw claims happen frequently, table base win claims do not occur.
In solving chess the position is just looked up in the table base.
"I'm reasoning with you" ++ No, you are not reasoning, you are just erroneously and thoughtlessly shouting that I am caught, wrong, mistaken, that I have to think, that it does not follow... Where do you see a personal attack from me?

"There is no contradiction. The 50 moves rule plays no role at all in solving chess."
I can and do understand  - that you're mistaken in what you say and keep saying there.
Its a bald generalized assertion that isn't accurate.
You do this over and over again.
It results from narrow perspective that is reinforced by 'belief' that you are always right.  The second makes the first worse and worse.

Its clear that the 50 move rule 'plays a role' in solving the game of chess because its part of the game of chess.  There's a 100-move rule too.
And its clear the world is not flat too.  

Saying it 'plays no role' qualitatively would be like saying the bishops 'play no role'.

For those who regard the change in the horizon as 'playing no role' that then reinforces their belief that the world is flat.
And that belief becomes an insistency.
Followed by more bald assertive over-generalizations that rationalize and build on each other.
There are ways to talk to and not talk to people who subscribe to a flat earth.
They reject all evidence and reasoning from others.
But this business of dispensing with horizons in views of chess and subverting math to do so too is probably quite a bit rarer.
Perhaps Sveshnikov is the 'guru' of such notions.
And parrotting Sveshnikov and den Herik to support unrealistic dogmatism  isn't 'pioneering' and isn't Galileo-like.
Reminds me of some 'pioneers' who claimed they had achieved 'cold fusion'.
Not that long ago. 
It was soon debunked.  Exposed.  Discredited.
Physics connects to industry - so there was a will to expose the false science quickly.
But chess connects to leisure time and entertainment more than to industry.

Repeat:
Its clear that the 50 move rule 'plays a role' in solving the game of chess because its part of the game of chess.  There's a 100-move rule too.
Would there now be a silly 'debate' about the semantics of the word 'solving' ?
Another 2000 posts on that ?
'Individuals' want to insist that their interpretation of the word 'solving' is to 'prevail' ?
There could be many opinions on the use of that word.
Whether rejecting the horizon while traveling - or rejecting the horizon in mathematics ... they're both extreme enough for others to realize there's no direct reasoning with those who try to so insist.
But that doesn't mean that such pseudoscience and pseudomathematics that is spammed so intensely in public forums shouldn't get some disagreement.
Chess will never be 100% analyzed - why?
'never' is kind of absolute.  Slightly provocative.   
Chess has a finite number of positions.
So in theory - it could be 100% analyzed eventually.
But what then happens is conversations about why this would take billions of years. 
The conversations then take at least two forms -
one with math and the other with pseudomath (and some pseudoscience and pseudochess too).

Pseudoscience: 
'chess can be weakly solved in five years with Cloud computers'
Maybe there are lots of variations on the pseudoscience.
Notions. 
First:
'Hey no - its three years not five.  You get the money from Paris Hilton.  To put in a lever on the sides of the Cloud computers.  
Press down Hard on the levers !! ...
and They then become Cloud 9 computers !
Sveshnikov would be Proud of you.

But then don't forget to 'take the square root'
drink real beer not 'root' beer before doing so.
And yes - only four options each move allowed ...
throw out the 50/100 move rules because they are just rules after all ...
but but but ...  put in Your Own Rule and limit chess to four options per side.
SInce they're just rules - why can't we add a New One ?
Hey its a pseudomath - pseudoscience - pseudochess Party !'

'(Bring your costume too.  Select in advance who will drive.)'

😁^✅ 👋 🎵🎸🏠🌹👍 😎🙂😡😂😈👻😿😹
 

A solution is required only to work for the positions it purports to solve. There is no requirement that it doesn't work for positions it doesn't purport to solve. 

@tygxc appears to think there is only one tablebase from his use of the definite article. Different tablebases may contain different illegal positions, but they all contain all the legal positions they purport to solve. (They don't purport to solve positions with castling rights.)

@MARattigan
 "(They don't purport to solve positions with castling rights.)"
which starts to underline just how difficult the task is -
supercomputers have been around for several decades but they struggle to have a 7-piece tablebase that can't even handle castling rights.
In simplified positions which would have a very low percentage of castling setups in them.
Where the castling option wouldn't even complicate the chess much in the rare instances its available.

20 years from now - maybe they'll have 8-piece completed - and with castling ...  but the scientists in those projects wouldn't scoff at the pseudoscience from one person here?  He's the only one pushing it here I believe.

@MARattigan if a position really is solved (forced win or forced draw (with details as to necessity) then it might be argued that all its descendant positions are therefore solved too?
Its Not True  !!  Diagrams would help -
but to try to put it concisely-verbally ...
Say a position has a single move option that forces a win. 
Then it might be Very Tempting to dismiss all the positions arising from some other move being played.
That is Misconceived Rationalization:
Because ...  such alternative positions might have been arrived at by other routes!
And then its no longer a 'forced win'.
So claiming that solving a position solves all its possible descendants is again pseudoscience/pseudomath/pseudochess.
I don't care nor know whether this was brought up before or not.  
Its still 'pseudo'.  But a lot of things are. 
Fiction - entertainment - movies - costume parties.  

To make this clearer - two illustrations might suffice.
White to move and win.  (obvious)

That one is an obvious win for white to move - with or without 'castling rights'. 
Simple 'skewer'.  (yes black wins instead if its his move.  Skewer again - but that's superfluous)
But in this next one - its a position that could 'arise' from the previous -
or by some other route !
White to move.

For example -
the White King could have just taken back a Queen at e2 and Black replied with Ke7. 
Rather than from white playing a silly non-capture Ke2 and black replying Ke7.
Point - you can't dismiss the position as 'won' as 'descendant' from the other one - nor dismiss because of 'inferior move' because it doesn't have to have got there by a silly Ke2 move with no capture the move before.
So now its a draw instead.  'Solved' - but you still - there too - still can't dismiss all the 'descendant' positions.
Draw !  Not win !   Meaning that the analysis had to start over !

So in the real project - its all about position by computer generation.
Not by game generation.
That could mean that in the tablebases - they have to look at every single position.  
Including if it was arrived at by blunder - as opposed to by alternate route
(there might not even be a possible alternative route sometimes)
- (but how could the computer know in advance).  
The project of knowing if positions could have been arrived at legally - 
is a 'game' project.  Its something for later stages.

Main Point: 
To do the project for real - the computers can't start to accurately conclude about positions that were arrived at - because another position had a winning move or moves ...
Positions have to be generated to have an accurate project -
Not 'game created'.

Unless pseudoscience is being done.  That's different.  

#262
Chess has a finite number of legal positions: 10^44.
Because of the 3-fold repetition rule the number of moves in a chess game is finite as well.
There are only a finite number of games possible with a finite number of positions and a finite number of moves.
The number of games is much much larger than the number of positions.
That is why solving chess has to be centered on positions, not games.
Solving chess is then exhausting all legal, sensible, reachable, and relevant positions.
Table bases contain some illegal and non-sensible positions as it is easier to leave these in than to weed these out.
The key question about the feasibility if weakly solving chess is the number of legal, sensible, reachable, and relevant positions.
That is also the time in nanoseconds it takes on 1 cloud engine of 10^9 nodes/s.
That is also the approximate minimal storage in bits to store the solution.
My estimation for this number is 10^17.
That makes it feasible to weakly solve chess in 5 years with 3 could engines of 10^9 nodes/s.
Sveshnikov was right.

The last seven lines repeating the same pseudoscience again.
'Argument by assertion'. 
"My estimation".
What was left out this time ?  the 'Galileo' pioneer claim.
and '50/100 move rule has no role' unreality.
Left out for some reason this time.  

A node cannot solve a position.  pseudomath.
But somebody can't seem to make up his mind between 'consider' and 'solve'.
"10^17" - more pseudomath.
'My estimation' leading to 'Sveshnikov was right' illogic.
That one's New.   Guru - disciple - back to Guru (returning the favor)  
And apparently nobody is a Convert yet.  
Nothing personal.  Just concerns the postings.

#265
"That's only an assertion. I asked for a proof."
That is proof.
There is only a finite number of legal positions (10^44).
Thus there is only a finite number of possible moves in a game as each of the legal positions can only be visited twice.
With a finite number of possible positions and a finite number of possible moves in a game, there is only a finite number of games.
q.e.d.

"How else are you going to assess the positions, rather than taking them forward, move by move?" ++ The method is indeed forward calculation from the opening towards the endgame table base, like Sveshnikov said.

"the positions all have to be assessed and that can only be done move by move. Therefore, analysis boils down to move by move assessment"
++ No, you can only assess at the end of a move sequence, i.e. on hitting the 7-men endgame table base or a 3-fold repetition. All evaluation functions are inherently flawed.

"I cannot think of any possible reason why analysis can be centred on positions and no-one here has managed a satisfactory explanation of how it can be achieved."
++ Just like has been done for Checkers and Losing Chess: calculation from the opening towards the table base with transposition tables to avoid double work on the same position.

"There is absolutely no reason why we should think Sveshnikov may have been right."
++ The reason why Sveshnikov was right is that there are only 10^17 legal, sensible, reachable, and relevant positions.

The idea of brute force solving is outdated. State of the art AI works on a more heuristic level. Does this mean that chess won't be 100% solved? I think the answer is a little more complicated that at first thought. If and when there is a 100% perfect chess AI, well have to wait very much longer to know that it actually is 100% perfect. I think we'll just give up trying to beat it with other AIs and it will be assumed to be perfect, but no one will know with absolute certainty.

"The reason why Sveshnikov was right is that there are only 10^17 legal, sensible, reachable, and relevant positions."
Its not reason.  Its pseudoscience.  And circular illogic.
Based on (quoted)  'my estimation' by a member here - not I.

One may as well say -
'the moon has no center' because whoever wants to say so.
Or even that the moon doesn't exist.  'Get rid of it'.
Like getting rid of 1000 trillion trillion trillion chess positions and then claiming 'five years'.  (and with lots of other illogic thrown in too)
To keep 'defining the narrative' ...  he'll also 'build on the narrative'.
A week from now there'll probably be at least one more repeat of the manyfold 'five years' illogic.  Maybe several.
Repeat(s).  But with slightly different word orders and phrasings.
But it'll still be the same illogic and attempted 'defining' of that narrative.

#267
"If and when there is a 100% perfect chess AI, well have to wait very much longer to know that it actually is 100% perfect. "
++ We already have over 1000 perfect chess games: draws from ICCF correspondence world championships.

The chess tree will never be entirely explored by bruteforce because of it size.

Exactly. There is absolutely no reason why we should think Sveshnikov may have been right.