Chess will never be solved, here's why

Sort:
playerafar
OctopusOnSteroids wrote:

You guys debating (with each other and taking it up with AI as well) what is positional and what is tactical are just talking about different things in different context. Its a useless debate. Nicely demonstrates how vague definitions are. The concept of positional vs tactical is useful for human understanding and evaluation, not useful for a context of solving chess. For robots its just pieces interacting with each other in different ways, complexity arising from amount of possibilities available. For humans theres the visual illusions, and danger of mistake from immediate traps etc that play a role.

Discussion of what is positional and what is tactical is one of many basics to chess discussion.
And again we have a particular member complaining about others discussing something.
With that same member wanting to claim that a 'game mechanics' terminology knocks out the existence of luck in chess .... that's quite a need for a conversation piece.
Discussions of positional versus tactical clearly have more merit regarding chess.
----------------
'game mechanics' semantical determinism ... is it akin to flat-earthism?
Wrong forum? 'Luck in chess' would be relevant to solving. So no. Not wrong forum.
How would a denier that Australia exists push a denial of luck existing in chess?
Could be tough.
Regarding Australia he would have to make too many claims of too many things being 'faked'.
Your internet opponent who was beating you in a chess game has an electrical power failure and you win ....
Would the denier of Australia want to claim that the power failure was faked and that the game was 'thrown'?

OctopusOnSteroids
playerafar wrote:
OctopusOnSteroids wrote:

You guys debating (with each other and taking it up with AI as well) what is positional and what is tactical are just talking about different things in different context. Its a useless debate. Nicely demonstrates how vague definitions are. The concept of positional vs tactical is useful for human understanding and evaluation, not useful for a context of solving chess. For robots its just pieces interacting with each other in different ways, complexity arising from amount of possibilities available. For humans theres the visual illusions, and danger of mistake from immediate traps etc that play a role.

Discussion of what is positional and what is tactical is one of many basics to chess discussion.
And again we have a particular member complaining about others discussing something.
With that same member wanting to claim that a 'game mechanics' terminology knocks out the existence of luck in chess .... that's quite a need for a conversation piece.
Discussions of positional versus tactical clearly have more merit regarding chess.
----------------
'game mechanics' semantical determinism ... is it akin to flat-earthism?
Wrong forum? 'Luck in chess' would be relevant to solving. So no. Not wrong forum.
How would a denier that Australia exists push a denial of luck existing in chess?
Could be tough.
Regarding Australia he would have to make too many claims of too many things being 'faked'.
Your internet opponent who was beating you in a chess game has an electrical power failure and you win ....
Would the denier of Australia want to claim that the power failure was faked and the game was 'thrown'?

Im genuinely impressed by your post somehow ending up from positional vs tactical to "denier of Australia claiming that a power failure was faked"

playerafar

If somebody were to just assert there's no luck in chess - as a policy or position -
that could be considered both sincere and respectable. Regardless of the inaccuracy.
Why not? A thing doesn't have to be perfect to be respectable.
Notice that chess not being solved further promotes luck in chess.
They're connected.
-------------------
Positional situations and tactical situations in chess are 'useless debate' in a forum about chess? while claims that winning a losing game because the winning opponent had an internet power and thus lost - 'isn't luck' - is considered 'useful debate'?
Would seem to relate to flat-earthism and denial of australia existing - whether to those sincerely or insincerely so claiming.
------------------
Idea: since somebody is claiming that positional versus tactical is 'useless' then how about that means it should be discussed further.
happy

playerafar

I asked Grok if it acknowledges that both positional situations and tacitcal situations exist in chess games and also whether there are many positions with both.
And of course it acknowledged Yes to all.
And we further discussed further whether game theory actually 'dismisses' positional situations and play.
It summarized like this:
-------------------------
Grok AI. Its 'take':

"While their point has a kernel of theoretical truth, it misses the practical and conceptual reality of chess, even within a game-theoretic framework. Here’s why:

Tactics and Position Aren’t Mutually Exclusive in Game Theory: In game theory, the "optimal strategy" for chess (if we could compute it) wouldn’t just be a string of checks and captures. It would include moves that improve your position—say, centralizing a knight or opening a file—because those moves increase your future tactical opportunities or restrict your opponent’s. Positional play isn’t separate from tactics; it’s the groundwork that makes tactics possible. Even in a solved game, many moves in the winning (or drawing) line would look "positional" to us—quiet moves with no immediate threats—yet still be optimal.
Perfect Play Doesn’t Erase Human Distinctions: Chess being "solved" in theory doesn’t mean humans (or even current engines) play perfectly. In real games, positional play exists because we operate under limited calculation. A position with no immediate tactics—like those closed setups we discussed—requires strategic decisions: “Do I reposition my bishop or push a pawn?” Game theory’s perfect strategy might assign exact values to these choices, but for players, they’re experienced as positional, not tactical, because no forcing line is visible.
Scale of Tactics Matters: If we define "tactical" as any move contributing to the forced outcome, then yes, every move in the perfect game-theoretic solution is "tactical" in a trivial sense. But that stretches the term beyond its chess meaning. Tactics in chess are short-term, calculable sequences with concrete gains (e.g., winning a piece). Positional play operates on longer horizons, often without forcing moves. Saying chess is 100% tactical because of game theory is like saying a novel is 100% punctuation—it’s technically part of the structure, but it ignores the broader narrative (strategy).
Evidence from Play: Grandmasters like Capablanca or Karpov excelled in "positional" games, winning through slow strangulation rather than flashy combinations. Engines like Stockfish often make quiet, "positional" moves in equal positions, only unleashing tactics when the opponent errs. If chess were 100% tactical, we’d see constant fireworks, but we don’t—quiet phases dominate many games.
Conclusion:
Game theory doesn’t negate positional play; it encompasses it. The forum claim conflates the theoretical endgame (a perfect strategy exists) with the practical experience of chess (where positional and tactical phases feel distinct). Positional play "exists" as a real phenomenon in how chess is played and understood, even if, in some abstract sense, every move serves the ultimate tactical goal of the perfect line. Their argument is clever but reductionist—it’s like saying music is just math because it has rhythm."
------------------
So Grok does not agree that 'positional' in chess is fiction ...
Nor does Grok agree that game theory 'dismisses' positional situations and play from chess games.
Even reminding how Capablanca and Karpov often won their games.

playerafar

The idea that 'positional chess' doesn't exist in 'solved chess' does not hold up.
In many book draws the defending player 'plays positional' in order to maintain the draw.
Including because he has no tactics to play anyway.
But in book wins there are many positional moves too.
Like for example N + B versus lone King. Most of the moves are positional.
Want a more common one? K + R versus King.
Conclusion: In all three phases of game theory 'white has a strategy to force a win - black has one - or either side has a strategy to force a draw' - there are many positional situations and moves.
------------------
Regarding the 'trichotomy' of game theory itself as applied to chess -
can it be challenged?
Well if the idea is to insist that for any given chess position there Can be only One of the three applying in each position - that can be challenged in a practical way.
Which is that in most positions nobody has the foggiest which of the three it is.
So it could be all three in that case. As in any of the three.
But are there others?
How about 'at least a draw'?
In other words say one side has a known strategy to force a draw - but decides to play on and look for a win and it isn't known if there's also a strategy to force a win there or not ... Now what?
That would imply more than three. Imply.
Implications. Real implications as opposed to something being implied.
I'm confident that Zermelo was aware of such things. Haven't checked though.

Elroch

Some of the following will be very familiar to most people, but probably worth saying anyhow.

All engines have an analogue of positional understanding, It is the evaluation function. This is particularly important for neural networks (and Stockfish uses a smaller neural network for its evaluations).

This function provides a first guess at what result will be achieved.

In fact the main neural network engines provide at least one number for every legal move in a position (or the result if there are no legal moves). One version is to provide probabilities of a win, draw or loss after each legal move. This can be converted to an evaluation function, for example by picking the move with the best expected score, and then using that score as the evaluation.

So I am asserting that something like "this position is worth +0.5 pawns to white" (a Stockfish-style evaluation) or "this position gives WDL probabilities of 30%/50%/20% for white" (Leelachess) are positional evaluations.

Such routines are not generally very good at tactics. While they are often in the right direction, calculation of explicit lines can reach a different evaluation that is quite different to the initial one.

The main purpose of the initial evaluation is to indicate how worthwhile it is to explore a line, but it is also used as all you have at a leaf node, by definition.

playerafar
Dubrovnik-1950 wrote:

That is because you have learned chess from other human chess players.

Positional play is just a teaching concept for humans or a way to explain very deep tatics. In a simpler way. But it does not really exist.

For example to the best computer chess programs. GM Carlsen's great positional insight, and understanding.

Is no different then the random play of a monkey playing the Strongest chess engines. And the results are exactly the same for GM Carlsen, and the Monkey playing Stockfish for example.

The notion that 'positional play doesn't really exist' looks like quite a hard sell.
Grok seems to agree that positional situations and positional play definitely exist.
An idea that they don't could be like saying that chess doesn't exist - or that its 'just a concept'.
Its not about me.
Who actually subscribes (among top players of the present and past) to the notion that 'positional play and positional situations don't exist in chess'?
Would the idea that they don't - be of any use?
If a chess instructor wanted to beat the idea into his student that 'positional play doesn't exist' because he wanted to make sure the student concentrates on tactics - then what does the instructor do when his student asks 'what about positions where there are no tactics - in other words no tactics available - no checks - no forks - no captures - no threat-moves to do so?'
And then the student adds: 'And what about positions where there's no good checks? No good captures? And the threat-moves available aren't good moves?'
I think most instructors would reply - 'Then you're going to be making a positional move. WIthout adding 'and they really do exist' - because that's obvious.
-------------------
Student: What if there are some good tactical moves available but I decide that at least one positional move option is better than all the tactical moves available at the time?'
Instructor: 'Then you're going to be making a positional move in that case too.'
Student: 'What if the best move is both tactical and positional and I've found it?'
Instructor: 'Then you're going to be making that move that is both positional and tactical.'
-----------------------------
Example: You take a pawn. Its a capture so its tactical. But the significance of the particular move happens to be that your piece is then on that square with the pawn kind of superfluous.

playerafar
Dubrovnik-1950 wrote:

It is not a hard sell at all. To those who understand chess theory, and how computers pick their final move to be played.

"Positional play" is a statistical understanding of something that could be true. And it could be true, or it could be false.

Computers use this knowlede to make search move ordering choices. As computers can not search to the end of the game. In most positions.

But all final moves are decided with a deep tactical search.

That is why the exact same Stockfish is stronger on modern hardware, then a 10 year old laptop.

Not because of more positional knoledge, but because of a deeper faster tactical search.

Who told you this idea of 'positional situations and play not really existing'?
Where did you get it from?
Did you think of it yourself?
You don't have to answer. Obviously.

playerafar

From Bobby Fischer:
"Quote: 'Tactics flow from a superior position.'
Fischer, known for his incredible tactical skills, understood that without a sound positional foundation, tactics become ineffective. Tactics were his sharp sword, but positional play was his shield. This quote directly counters the idea that chess is 100% tactical, highlighting the fact that tactics are most powerful when they emerge from a position of strength."
---------------
Tal - arguably the player most famous for tactics.
"Tal's Famous Quote on Position:
'The most important thing in chess is to know what kind of position you need to get.
The other parts of the game will come from there.'
This quote shows that Tal was very much aware of the role of position in chess. Even though he was known for his daring sacrifices, Tal's understanding of what kind of position would allow his tactics to flourish was crucial. He was always looking for imbalances, open lines, and weaknesses in his opponent's position to exploit tactically. But these opportunities often stemmed from positional decisions he made early on."
-------------------

mpaetz
VerifiedChessYarshe wrote:

"Chess will never be solved, here's why" the statement is true if humans keep fighting each other for dominance.

Chess IS two contestants " fighting each other for dominance" over the board. If humans no longer had any interest in such matters, chess would disappear.

Elroch
Optimissed wrote:
MARattigan wrote:

No. The game theoretic value was a win and remained a win. The game state changed with every move.

So far you've confused game state with both game theoretic value and nodes reported by an engine.

Since you're apparently totally confused by all the terminology, hadn't you better define terms or jargon that you're using so that we know what we're discussing?

I sense that by "game state" you mean "position". Since "position" is a perfectly good word and we know that it refers to the position at any given time and which changes with every move, hadn't you better stick to that (that is, "position") so you understand what you're talking about at any given time, without it changing its meaning, sentence by sentence, in your mind? It would make it all much easier to follow.

Your sense is unfortunately incorrect, even if, as invisible parts of the "position", you include castling rights and e.p. possibility, as in FEN format.

MARattigan understands that in chess as played on chess.com, the set of possible future positions that have been reached twice before is part of the state, for example, since these determine the value of nodes that repeat that position. In addition the ply count since the last irreversible move is an essential part of the state.

Elroch
Dubrovnik-1950 wrote:
Optimissed wrote:
Elroch wrote:
Optimissed wrote:
MARattigan wrote:

No. The game theoretic value was a win and remained a win. The game state changed with every move.

So far you've confused game state with both game theoretic value and nodes reported by an engine.

Since you're apparently totally confused by all the terminology, hadn't you better define terms or jargon that you're using so that we know what we're discussing?

I sense that by "game state" you mean "position". Since "position" is a perfectly good word and we know that it refers to the position at any given time and which changes with every move, hadn't you better stick to that (that is, "position") so you understand what you're talking about at any given time, without it changing its meaning, sentence by sentence, in your mind? It would make it all much easier to follow.

Your sense is unfortunately incorrect, even if, as invisible parts of the "position", you include castling rights and . possibility, as in FEN format.

MARattigan understands that in chess as played on chess.com, the set of possible future positions that have been reached twice before is part of the state, for example, since these determine the value of nodes that repeat that position. In addition the ply count since the last irreversible move is an essential part of the state.

Do us all a favour. Go with the sense of the comment instead of making a fool out of yourself with the pedantry.

If you weren't totally biassed towards aways arguing in support of people who support you, because you constantly need support, I would have some respect for your opinions, even though they are often incorrect. And don't try to pretend that since you two have faked some arguments, you aren't in love.

He has got to twist himself into knots to try and be right here. But it just does not work.....

I AM right here. And if you think otherwise it is likely due to poor understanding or (or wild guessing about) some position of mine.

Do try to justify your comment or acknowledge it was inappropriate.

Positions are fine for basic chess, and although basic chess is an infinite game, there are technical workarounds that make it solvable in principle. But most people consider chess to be a game with additional drawing rules. For any version with a repetition rule, positions don't cut the mustard. States with more information are needed. (Even for versions with an n-move drawing rule, the state has that ply count that is not part of the position in the sense understood by most people. It's in FEN because the state is what matters).

playerafar

Of course Elroch is right.
(But 'bored' whoever (hs name begins with O) will worry pathetically about who supports who and troll and spam-troll predictably to that effect)
Elroch and MARattigan are rightly pointing out that just a diagram of pieces on the board doesn't properly describe the state of the game - more information is needed to do so - like the very obvious 'whose move it is' and castling and en passant information and 3-fold and 50-move rule information and in game situations clock-information too.
Obviously.
Disputing the obvious is 'cute'? I guess so. Sometimes.
Flat-earthism 'adds colour' to behaviour. Sometimes.
Whether the 'flatter' believes his own nonsense or not.
Pretending that variables like whose move it is 'don't matter' to the position would be like a form of 'chess flat-earthism'. But who lets themselves get that bored they would need to claim such things?

playerafar
Elroch wrote:
Dubrovnik-1950 wrote:
Optimissed wrote:
Elroch wrote:
Optimissed wrote:
MARattigan wrote:

No. The game theoretic value was a win and remained a win. The game state changed with every move.

So far you've confused game state with both game theoretic value and nodes reported by an engine.

Since you're apparently totally confused by all the terminology, hadn't you better define terms or jargon that you're using so that we know what we're discussing?

I sense that by "game state" you mean "position". Since "position" is a perfectly good word and we know that it refers to the position at any given time and which changes with every move, hadn't you better stick to that (that is, "position") so you understand what you're talking about at any given time, without it changing its meaning, sentence by sentence, in your mind? It would make it all much easier to follow.

Your sense is unfortunately incorrect, even if, as invisible parts of the "position", you include castling rights and . possibility, as in FEN format.

MARattigan understands that in chess as played on chess.com, the set of possible future positions that have been reached twice before is part of the state, for example, since these determine the value of nodes that repeat that position. In addition the ply count since the last irreversible move is an essential part of the state.

Do us all a favour. Go with the sense of the comment instead of making a fool out of yourself with the pedantry.

If you weren't totally biassed towards aways arguing in support of people who support you, because you constantly need support, I would have some respect for your opinions, even though they are often incorrect. And don't try to pretend that since you two have faked some arguments, you aren't in love.

He has got to twist himself into knots to try and be right here. But it just does not work.....

I AM right here. And if you think otherwise it is likely due to poor understanding or (or wild guessing about) some position of mine.

Do try to justify your comment or acknowledge it was inappropriate.

Positions are fine for basic chess, and although basic chess is an infinite game, there are technical workarounds that make it solvable in principle. But most people consider chess to be a game with additional drawing rules. For any version with a repetition rule, positions don't cut the mustard. States with more information are needed. (Even for versions with an n-move drawing rule, the state has that ply count that is not part of the position in the sense understood by most people. It's in FEN because the state is what matters).

Yes. Elroch is right.
And now a poster who's mostly capable - giving in to 'O-nuts' and doing what O-nuts does which is to try to intimidate other posters from posting. And also pretending to miss a point he didn't and then projecting that trolling behaviour.
Its too bad really. 'D' was doing fine - talking about Grok and making valid comments about the nature of solving ... but in addition to falling into such desperation -
also tripped up on game theory in a somewhat minor way and then in a major way with a bad stumble about 'positional situations and play not really existing' ....
he stopped pursuing that one further and now is back to pushing the three game theory situations in a way I think Zermelo himself would not have pushed. Kind of like Stephen Hawking followers being less objective than Hawking about alternatives to his theories.
------------------------
Regarding the three situations - would that have been basic to chess computer programming and pioneering? Would Botvinnik have used Zermelo game theory in his contribution to chess computing? I don't know. Yet.
Obviously the computers have to mostly deal with situations where any of the three basic situations might apply. Without knowing which one.
But that could even be expanded to five situations.
Black - or white - is in a chess situation where a known strategy exists to force a draw - but doesn't know whether there's also a strategy to force a win - which would be present in many cases without it being known at the time 'the player or computer had a win there with that other move' (but that doesn't come to light until after the game with supercomputers and players going over the game without the presence of the chess clocks and 'the game ongoing'.)
Zermelo would have been aware of such ideas - except for the chess supercomputers part - which didn't exist in his time.

playerafar

"What chess engine would play perfect chess, and would be able to play the stronger "positional" chess game,?"
There's no such computer.
Manoeuvering AI into giving the answers you want is well known.
-------------
Could be related to flat earthism.
'I want you to talk to me from the perspective of a flat-earther - not the real perspective.'
AI : 'Okay - what have you got?'
'The earth is flat. Right?''
AI: 'Yes. Of course it is'

playerafar

A month from now - will there be a 'raging debate continuing' in this forum that positional situations and play 'don't really exist' in chess?
For two years in this forum - a member perpetuated spam that 'chess can be solved in five years with Cloud-computers if the money's there'. With another member trying for those two years to deter opposition to that spammed disinformation. Both of them continued to fail and then finally gave up completely.
In another forum - a member is trying to pretend that just mentioning the phrase 'game mechanics' would eliminate any chance of 'luck existing in chess'. He's still pushing it months later. WIth the same other member trying to deter interference with that disinformation - but again failing in his usual failed ways.
Reactions of other members to such things vary.
But many just use the false claims as 'placeholders' to have the real conversations about the forum subjects - whether posting around the misinformation and disinformation - or using it as direct opportunities to post the real information including by refutation.

AbhayB2025

Bro I just clicked on @TheChessIntellectReturns profile it says you closed your own account in 2022?

VerifiedChessYarshe
Dubrovnik-1950 wrote:
playerafar wrote:

Of course Elroch is right.
(But 'bored' whoever (hs name begins with O) will worry pathetically about who supports who and troll and spam-troll predictably to that effect)
Elroch and MARattigan are rightly pointing out that just a diagram of pieces on the board doesn't properly describe the state of the game - more information is needed to do so - like the very obvious 'whose move it is' and castling and en passant information and 3-fold and 50-move rule information and in game situations clock-information too.
Obviously.
Disputing the obvious is 'cute'? I guess so. Sometimes.
Flat-earthism 'adds colour' to behaviour. Sometimes.
Whether the 'flatter' believes his own nonsense or not.

Except I defined the 3 game states in my example. As being a Win, Loss, or Draw by force.

These morons will say and twist anything to try and win!

Is the picture depicting you or your fellow morons?

playerafar
Dubrovnik-1950 wrote:

Right! And this is a thought experiment to show positional play is a construct of the tactical game tree of chess. Showing chess is a 100% tactical game. And no positional understanding is needed to understand chess, or to solve chess.

Positional play is just a way to describe deep tactics in the game tree of chess. They are just teaching concepts invented by human chess players.

And that are correct sometimes, but not all the time.

You've chosen to not answer the question -
is the notion (false) that chess is '100% tactical' your idea?
Or you got it from somebody? (indoctrination).
Is that a sign of wavering - 'are correct sometimes' ?
Uh oh.
Any concession is going to invalidate '100%' ....
------------------
what's really going on?
Lasker suggested that positional play is a tougher part of chess than tactics is.
Neither positional play nor tactics exist in a vaccuum.
But improving players usually prioritize tactics in their studies.
Its a consequence of a general principle of taking care of what's less difficult first.
Also - trying to maintain that chess instruction and chess games are in different universes - or in 'separate boxes' ... another mistake.
Parellel: Some try to maintain that 'infinity' doesn't exist. That its 'just a mathematical concept.'
In other words they are 'finite universe' subscribers.
'infinity doesn't exist'. That's dogmatic. Lacking in objectivity.

playerafar
AbhayB2025 wrote:

Bro I just clicked on @TheChessIntellectReturns profile it says you closed your own account in 2022?

Correct. He closed his account the day after opening this forum.
'Bro' isn't around to answer your question.
I'm thinking he closed his account in frustrration when tygxc served notice as to what he tygxc intended to do to this forum in the very next post after the opening post and subsequent posts