Chess will never be solved, here's why

     I have previously posted here and in similar forums to point out the endless re-plowing of the same ground. The consensus seems to be that we are not going to see the breakthrough published in the immediate future. My own opinion is the five-year estimate is wildly overoptimistic and the never (or many millennia) option seriously underestimates human ingenuity.

Yes - exactly.  the 'solvability'.  Or the lack of same.
Nowadays - with computers and the internet and supercomputers and 'engines' available - totally 'solved' tactics problems and endgame problems and 'endings' are at the fingertips - beautifully isolating and demonstrating and illustrating perfect key points with an intense Efficiency never before possible.  

But openings continue to have 'mystery' and are 'not solved'.
Which means chess is not solved too.  
And a gigantic number of middlegames are 'not solved'.
In fact - anything with 8 pieces or more onboard including many endgame positions are generally 'not solved'.
That's right.  If its not 'all solved' - then its not solved.
But some might not want to see it that way. 
Choice.  Psychology of perception.

Regarding 'wildly overoptimistic' - even that looks like an understatement - just given the gigantic effort already just to 'solve' for 7 pieces or less.
(which it now turns out - skipped castling possiblities in those positions)
With a Herculean task ahead for 8 pieces.
Its not hard to extrapolate from there.  
Will we see 12 pieces 'solved' in our lifetimes? 

Regarding the 'many millenia' scenario at the other end -
- it could be more like 'billions of millenia'.
But that presupposes there won't be more vast increases in computer power.  But there will be.
But that might only cut off a few zeroes.  A few powers of ten.
The 'big jump' could be in the actual programming.
Not shortcuts that 'skip solving'.
Programming that's just much better in linking totally solved positions to unsolved ones.
Whether its to tablebased positions (but those only go up to 7 pieces onboard)
or to checkmates and stalemates or other certain draws - with more than 7 pieces onboard.
If the programming linkage occurs - then maybe many powers of ten can be knocked off the task - instead of just the few we see.  
But that's an unknown.  We don't know how much the programming will improve.
That might be the key.  The programming.  Not the computer power.

Can't be sure. I might have edited it in after without having read yours yet.

I've looked into all four.

None of them are draws. They're all wins for Stockfish.

A little less wine?

You couldn't even make it a single post without resorting to your old, tired "I win this argument because I say am smarter than all of you...and since I *am* smarter than all of you, you cannot even understand my superior arguments".  Congrats.

Once again:  I challenge you to make arguments going forward without elevating your own intellect or denigrating others'.  Can you?  If you *cannot* make valid arguments and make them stick (or delude yourself into deciding they have stuck) without using this crutch, then I propose that you are full of it in general and not nearly as accomplished in this arena as you think you are.  I am guessing this would not be a surprising verdict to those that know you offline.  

P.S. Calling yourself the best debater on Facebook is damning praise, I assume you realize?

Good post.  Yes.  Wins for Stockfish 14 or whatever.
As for chess being 'solved' before the year 2100 - 
could that be possible if they have breakthroughs in how to improve the computer programming of the task?
There's something called parallel processing.   And computers improved a lot when programmers got something going that's like a 'neural net'.  
Maybe they'll even find better ways for the computers themselves to generate better programs for the task.

Some side-tasks could become possible.
Like they might be able to find a way to more exactly compute the minimum number of zeroes (powers of ten) to be knocked off a time needed - to make the task feasible.
In other words the minimum number of divisions by ten needed. 
That's probably already known - in relation to the power of today's computers. 
But there are some peripheral issues.

Isn't the actual fact that there are more chess move combinations than the number of atoms in the observable universe? Regardless I agree with the point you're making.

There are various peripheral questions.
If the exact number of possible legal chess positions is ever computed - and is ever then set out as an actual number -
would that mean in turn that chess has then therefore also been 'solved'?
After looking at this for a bit -
I came up with a kind of proof of No.

If all position-counting "could be done" by generating all positions by reversed legal capture then in that impossible hypothetical case then 'hypothetical yes'.  
But that Can't be done.  Not that way.
Because anytime a new position is generated from a 'solved' position by reverse legal capture - then other new positions could be generated from that reverse capture-generated position -
by reversed non-capture moves ! 
And then those thus-generated positions would themselves have other option non-capture forward moves other than the generating reversed non-capture move.
And it wouldn't follow that those are 'solved'.  That would still remain to be done.
So No.  No #1.

Another side issue - but the No just arrived at doesn't depend on this next one.
You could then generate all legal non capture move possibilities from the previously stated - but that would take time too.  And then you've still got to 'Solve' the ensuing too.  Prove they're forced draws or wins or lead to same in all possibilities.
No again.  No #2.  
  
Another concept:   If you have a position with four minor pieces each - and no pawns - do we 'ordain' that as a draw?  
No.  Can't.   Depends on the position.
And perhaps computers would compute for years on some - looking for a forced win or draw millions of moves deep.  
You need the 50 moves rule there.
Could that be then called 'weak solving' by '50'?  Its not 'solved'.
No #3.

But for each new No - the previous No doesn't depend on the subsequent series of No. 

Anyway - No.  Knowing the exact number of legal possible chess positions - wouldn't 'solve'.
Has that been done already? The count?  I don't know.  
Do I care?  No. We're just talking.  Care free.   
(except for those who do)
That's another issue.   The 'count' of all possible legal positions.
Is it possible its not only been done but all such positions are itemized and stored?  (with most of them Not solved) ?
Again - I don't know.  I don't care.   
When I watch a movie - do I care who the author has written the 'bad guy' to turn out to be ?
No.  But that doesn't mean its not entertaining or not interesting or not an exercise.
This whole website is mostly people using spare time and not getting paid for it.  Relaxed.   

Putting it in terms of games or move combinations makes it harder.
But even if you move down to just positions - instead of move sequences - its still going to be far too big a number.
The number of atoms in a gram-molecular weight of a substance is called the Avogadro number.  Its less than 10 to the 24th power - which is one trillion squared.   Within 'range' of supercomputers?
But the number of possible chess positions is much bigger than that.
Too much bigger.  

Much of the discussion has been about 'weak solving'.
Which is unfortunate terminology. 
With too much depending on its semantics.  And unnecessarily.
and without much gain - even if optimum semantics are realized on that.

Much better: 
Computer tasks relevant to the forum topic that can be done thoroughly and perfectly. 
Presenting 'upper bounds' on various types of related computer task - with no opinion-based premises - and with derivations of those upper bounds included.
Upper bounds are by their very definition - somewhat imperfect.
But that doesn't mean perfect and simple derivations can't be presented.
Following that - time estimates presented for each task.  
This series of tasks doesn't require computers.  Nor even websites.

In Lasker's manual of chess - 
he gives various estimates of how long it takes to present and train a student regarding various aspects of chess.
Training a chess student regarding positional play is much harder and takes much longer than training a student in tactics.  
If I remember his estimate correctly -
It was 120 hours of training for positional play.  Maybe a quarter of that for tactics.  That's master training though - with various basics already in place.

Is there something analagous to this in computer analysis?
Computers can tactics-check a position quickly or very quickly.
Or even solve - and maybe report something like 'no decisive tactics here'.
But comparing positional moves - finding one better than another ...
computers here on the site will mostly display little differences in values of positional play.   But they're not 'solving' there.  Its much much harder.

Common mistakes in chess: 
1) Putting positional play and tactics in 'separate boxes'.  
Other mistakes:
2) Pre-assigning themes before doing 'calculations'.  Often the 'theme' is only useful as an 'afterwards' thing.  Much better:  observations first.
3) Failing to distinguish observations in chess from calculations.  Includes judging that's 'just semantics'.  Its not.  They're very different.
4) Failing to appreciate 'misdirection' from chess positions.  Causes many if not most of blunders and inadequate play.  The usual blockage to communication on this:
"hey - only living things can cause misdirections."  Wrong.
Or:  "I don't know what you mean !  You mean misdirection from the opponent?"
Experience - if the term 'misdirection' is not accurately interpreted right away - then its probably not going to be.  

Nonliving entities can and do cause all kinds of things. 
Again - the principle barrier there is concern with semantics rather than making the jump to realizing what is involved.  The realities of misdirection instead of its semantics. 
Ever hear this one?  "A falling tree in the forest makes no sound unless somebody is around to hear it." 
Of course it makes sound ! 

Chess positions send us signals.  Those signals cause chess move errors.
Millions of instances of this happen every single day !

Which is nonsense. They are just different things.  I explained that one hypothetical weak solution of chess would be the combination of a 16 piece tablebase,  a complete deterministic opening book for white (i.e. including all black options at every move, but just one white move) with all leaf nodes having 16 pieces or a decisive result, and a similar opening book for black.  (This description presumes the optimal result is a draw - if the result is a win, an opening book for just the winning side suffices).

An example (provably) less than a 32-piece tablebase of all legal positions (which is what a strong solution amounts to).

The truth of the preceding statement is not compromised by my lack of such a weak solution of chess.

"Which is nonsense. They are just different things."
Entirely reasonable and legitimate in addressing the particular context concerned.
As for 'hard guy' - I happened to glance at the bottom of a previous post.
He has now switched from 'fools' to 'trolls' for now.  Going through his 'cycle'. 

Perspective on the topic:
Weak/strong 'solving' (unfortunate terminology) can be replaced.
Tablebase solving by computers of positions with seven pieces onboard or less - 
can be compared with how human players go at positions in general.
Whether in attempts to find best move or moves - to 'solve' - or in methods of selecting moves during games.

This discussion equates to administering medicine to the dead !

Yeah I don't think it was that. It had something to do with n, or  a certain number of queens. I think it said the n number was solved up to 27, (which has 200 quadrillion solutions) but not higher than that. Computers are not that powerful yet. 

So apparently this guy figured out a way to solve it for above 27. They said in one example an answer had 5 million digits. That's a big number. 

The new mathematical result was about the behaviour of the solutions as n tends to infinity.