Chess will never be solved, here's why

Sort:
playerafar

@MARattigan 
I tried the button and White is lost.

But regarding as to when we can dismiss a move for 'weak solving' purposes' ...  there's some 'big problems'  happy.png
Say one move was +2 and another -2 at the same point in the game with the same player to move.
The minus 2 can't be discarded - because it is not 'losing' ....
Lol !    happy.pngevil.pnggrin.pngblitz.pnggold.pngdiamond.png

playerafar

And - running that analysis button - which is Stockfish I believe (I don't know what version) 
then bxa6 isn't rated that far below Nxa6. 
(after giving it 30 seconds or so)
An advantage of over four points for black. 
But less than a point between the two takings of the bishop.

Next point:  Is a difference in evaluation of 4 points more than enough to dismiss a move (or to dismiss a candidate move for the opponent?)  ?
Could be.
Especially if the 'spread' is close to zero?
Say +2 versus -2.
But if one move was +12 and the other was +8 ... should the +8 be dismissed?  I'm thinking no.
Some would say "At +8 its 'over' anyway - no need to 'solve' further - so no worries about the +12 instead"
Which brings on another idea - anytime the evaluation for a move hits a certain plus number - then look for 'final time' on that move.
Same where all choices are too negative.  Its considered lost.
I would say even +6  (or minus 6) is Crushing. 
If that's the best move available.
But then there's a Nasty hitch ...  somebody has +6 available but plays a plus 3 instead.  Could we 'dismiss' the plus 3 ?  How?  It doesn't lose.

MARattigan

@playerafar

Try the analysis button here.

White to play and draw

 

playerafar


Hi Martin !
I answered your post but my reply refused to post !
I'll try this one now - and if that works - edit it maybe.

@MARattigan - the analysis button in your diagram indicated that white is Lost.

But now - there's a big 'problem' with 'dismissing moves'.
Say at any point in a game - with the same player to move - (computer vs computer) the top choice is +2 and the next one down is -2 ...
then can the minus 2 and below be 'dismissed' ?
It starts to look like No.  Because minus 2 isn't necessarily Losing !  
happy.pnggrin.pngevil.pngblitz.pnggold.pngdiamond.png

MARattigan

@playerafar

Point is it indicated wrong.

playerafar
MARattigan wrote:

@playerafar

Point is it indicated wrong.

You mean white can draw ??  happy.png
I guess I better look at your diagram some more.

playerafar
MARattigan wrote:

@playerafar

Point is it indicated wrong.

I think I get it - White's King just runs down to a1 and 'hangs out' there.
Doesn't shave.  Shirtless.  And ...
Great one Martin !
Several points arise from that.  
One could be that Nxa6 could be Much better than bxa6 in the Ba6 thing ?

playerafar
MARattigan wrote:

@playerafar

Well, I can't see any way he can win.

I tried some moves. 
Black can try to cut off b2 but then white just steps over and wins that pawn.
But it looks like there's a really Nasty Flaw in the engine there.
Because when I click on lines that have the White King shuttling between a1 and b2 - the engine Refuses to assign 0.00  !!
I wonder if chess.com is aware of the Bug ....
Am I certain that white draws this ?
Not 100%.
I don't see how black stops white from bouncing a1 to b2.

MARattigan

Main point is SF's numbers prove nothing.

playerafar
MARattigan wrote:

Main point is SF's numbers prove nothing.

I used chess.com SF just for example.
The project engines would be better ...
If you're saying 'weakly solving' requires 'proof' ...  well that's relevant.
But 'weakly solving' could have multiple definitions ...

MARattigan

If you look back at the posts it does, but a proof is always required.

But you just pointed out a flaw in the definition I posted earlier. It should read:

A (timely) weak solution means that for the initial position either a proven (timely) strategy has been determined for one player that achieves a win for that player against any opposition, or a proven (timely) strategy has been determined for each player that avoids a loss for that player against any opposition. 

Edit: Modified the wording to make it explicit that the result of a strategy refers to the result for the player to whom the strategy applies.

playerafar
MARattigan wrote:

If you look back at the posts it does, but a proof is always required.

But you just pointed out a flaw in the definition I posted earlier. It should read:

A (timely) weak solution means that for the initial position either a proven (timely) strategy has been determined for one player that achieves a win against any opposition, or a proven (timely) strategy has been determined for both players that avoids a loss against any opposition. 

That's a nice definition.  I appreciate that.
Care to qualify how it would differ from 'strongly solving'?
Yes I realize that's probably been done multiple times in the forum -
but this would be a good place.
And our posts crossed - so I'll bump my other post to a bit later.

playerafar

The overall discussion doesn't have to be 'airtight solving' -
it seems to be well established that's gazillions of millenia away ...
unless there's big improvements to the software and hardware.

But then there's other related things.
'Weak solving' has been looked at for 2000 posts or so.
What about -
Will all possible legal chess positions ever be Categorized ?
Some could reply:
"Already done" - every position is either 'solved' or it isn't.
But that's always been true.
Maybe the percentage of 'truly solved' positions so far - is known.
So maybe there'll be estimates of how long to improve that percentage by how much.
In other words - a graph.  

MARattigan
playerafar wrote:

The overall discussion doesn't have to be 'airtight solving' -
it seems to be well established that's gazillions of millenia away ...
unless there's big improvements to the software and hardware.
...

Well established, yes, but not to all.

The overall discussion is clouded by people contributing without realising what solving means.  If it's not airtight it's not a solution.

@tygxc is planning to spend the next quintillion years getting several supercomputers to not solve chess.

playerafar
MARattigan wrote:
playerafar wrote:

The overall discussion doesn't have to be 'airtight solving' -
it seems to be well established that's gazillions of millenia away ...
unless there's big improvements to the software and hardware.
...

Well established, yes, but not to all.

The overall discussion is clouded by people contributing without realising what solving means.  If it's not airtight it's not a solution.

@tygxc is planning to spend the next quintillion years getting several supercomputers to not solve chess.

Lol !   Hahahahhahahaha
Regarding your great position with white drawing by running his lone King to a1 ... even though SF keeps stubbornly assigning it as lost for white even after advancing the moves ...
Well if you don't mind the pun ...
that position you've posted ... is a1 !!! happy.png

playerafar

There's people in various clubs who might like seeing that one.
And - I wonder if the staff know about it.

MARattigan
playerafar wrote:
MARattigan wrote:

If you look back at the posts it does, but a proof is always required.

But you just pointed out a flaw in the definition I posted earlier. It should read:

A (timely) weak solution means that for the initial position either a proven (timely) strategy has been determined for one player that achieves a win against any opposition, or a proven (timely) strategy has been determined for both players that avoids a loss against any opposition. 

That's a nice definition.  I appreciate that.
Care to qualify how it would differ from 'strongly solving'?
Yes I realize that's probably been done multiple times in the forum -
but this would be a good place.
And our posts crossed - so I'll bump my other post to a bit later.

A (timely) strong solution means that for all legal positions either a proven (timely) strategy has been determined for one player that achieves a win for that player against any opposition, or a proven (timely) strategy has been determined for each player that avoids a loss for that player against any opposition. 

Edit: Modified the wording to make it explicit that the result of a strategy refers to the result for the player to whom the strategy applies.

playerafar

So you're contrasting weak and strong by whether its for the opening position or all positions?   That's the difference I saw.

MARattigan

Yes that's the difference.

DreamscapeHorizons

Chess will never be solved because all u have to do is tinker with the rules ever now and then.