Chess will never be solved, here's why

Sort:
MARattigan
Dubrovnik-1950 wrote:
MARattigan wrote:
EndgameEnthusiast2357 wrote:
Dubrovnik-1950 wrote:
crazedrat1001 wrote:
Dubrovnik-1950 wrote:

Yes, as it was posted before. If you are going to continue to post nonsense.

I am just going to have Grok pull the facts about Chess, and Game theory, and repost them as you are doing.

It is clear you know nothing about chess, computer chess, or game theory.

No more time needs to be wasted, then just reposting the FACTS.

Christ you are dense.

I know you can not fix your kind of stupidity, but just stop your nonsense!

Conclusion
Chess is a perfect fit for game theory because it encapsulates the core elements of strategic interaction: two rational players, a finite set of choices, perfect information, zero-sum outcomes, and a structure amenable to equilibrium analysis. Its complexity ensures it remains a rich testing ground for game-theoretic concepts, while its clarity makes it a textbook example. Whether viewed through the lens of minimax, equilibrium, or extensive-form games, chess is a living embodiment of game theory’s principles—a battle of minds where every move is a calculated step in a grand strategic dance.

And a finite length of games. Former world champion Max Euve incorrectly claimed in his Mathematics PHD dissertation that an infinitely long game of chess is theoretically possible. Finite board, finite number of pieces, repetition rules..etc, not sure what I am missing but simple logic, all chess games ended eventually no matter how long you try and extend them out.

What you're missing is that the 50M and 3R draws do not have to be claimed and FIDE Removed the 50M and 3R rules from the basic game in any case in 2017.

Under competition rules game length is necessarily finite by the 75M/5R rules.

And game theory assumes that chess players are not MORONS. And will play for their best outcome at all times.

Er, no it doesn't. Not if it's calculating the longest possible game.

In fact it doesn't really assume anything anywhere about chess players except they play according to the rules of the version they're playing.

MARattigan
Dubrovnik-1950 wrote:

"Totally clueless. Under FIDE basic rules the number of possible games is infinite. Under competition rules a lower bound of 10^29241 is estimated here."

You think this because you are just ignorant to the facts of the Shannon number.

The Shannon number made decades ago. And was cutting the number of possible game to 10^120.

By taking the average length of humans games ending at 40 moves. And is a lower bound of possible games. And making the assumption of 35( if I remember correctly) possible legal move average, and a game length of of 40 moves, or 80 plies.

And you're calling us morons.

MARattigan

Have you read the basic rules? Can you read the basic rules? If some of the words are too long we can maybe help you.

MARattigan

You need help alright.

Elroch
MARattigan wrote:

Have you read the basic rules? Can you read the basic rules? If some of the words are too long we can maybe help you.

I am not sure about his reading skills but he has mastered both copying and pasting.

Wxrenski
Ok
Elroch
Dubrovnik-1950 wrote:
MARattigan wrote:
Dubrovnik-1950 wrote:
MARattigan wrote:
EndgameEnthusiast2357 wrote:
Dubrovnik-1950 wrote:
crazedrat1001 wrote:
Dubrovnik-1950 wrote:

Yes, as it was posted before. If you are going to continue to post nonsense.

I am just going to have Grok pull the facts about Chess, and Game theory, and repost them as you are doing.

It is clear you know nothing about chess, computer chess, or game theory.

No more time needs to be wasted, then just reposting the FACTS.

Christ you are dense.

I know you can not fix your kind of stupidity, but just stop your nonsense!

Conclusion
Chess is a perfect fit for game theory because it encapsulates the core elements of strategic interaction: two rational players, a finite set of choices, perfect information, zero-sum outcomes, and a structure amenable to equilibrium analysis. Its complexity ensures it remains a rich testing ground for game-theoretic concepts, while its clarity makes it a textbook example. Whether viewed through the lens of minimax, equilibrium, or extensive-form games, chess is a living embodiment of game theory’s principles—a battle of minds where every move is a calculated step in a grand strategic dance.

And a finite length of games. Former world champion Max Euve incorrectly claimed in his Mathematics PHD dissertation that an infinitely long game of chess is theoretically possible. Finite board, finite number of pieces, repetition rules..etc, not sure what I am missing but simple logic, all chess games ended eventually no matter how long you try and extend them out.

What you're missing is that the 50M and 3R draws do not have to be claimed and FIDE Removed the 50M and 3R rules from the basic game in any case in 2017.

Under competition rules game length is necessarily finite by the 75M/5R rules.

And game theory assumes that chess players are not MORONS. And will play for their best outcome at all times.

Er, no it doesn't. Not if it's calculating the longest possible game.

In fact it doesn't really assume anything anywhere about chess players except they play according to the rules of the version they're playing.

Under FIDE basic rules the number of possible games is infinite - MARattigan

And it is not infinite moron.

You have advertised your own lack of intellectual credentials by disagreeing with him.

Here is a proof that the number of possible games in FIDE basic rules is infinite.

Use 0 to represent any one of the moves Nf3, Nf6, Ng1 and Ng8 and 1 to represent the moves Nc3, Nb1, Nc6 and Nb8.

Then for every infinite binary number between 0 and 1, there is a legal FIDE basic rules chess game defined by picking the (unique) moves corresponding to the digits of that binary number.

If not already clear this will be clear with an example. Take the binary number 0.1101... Ignore the 0 before the decimal point (it never varies). Then the first move is 1. Nc3 (see the definition above. Then black plays 1...Nc6 (again see the definition). Then 2. Nf3 and finally 2...Nb8.

A simpler way to express the definition is that each player moves their king's knight if the digit is 0 and their queen's knight if the digit is a 1, and they only moves the knights between two specified squares (in either direction).

I assert that every distinct sequence of binary digits gives a different legal basic rules game and that the number of the distinct games is uncountably infinite.

Of course this is only a subset of all legal basic rules games, but if a set has an infinite subset it is infinite.

It's worth emphasising that there is a crucial reliance in this reasoning that a draw is not automatic when there is a threefold repetition or 50 moves without a pawn being moved or a piece being taken. While the players playing one of the games described COULD claim a draw at some point, they are not obliged to, and that is what matters to the count.

Perhaps @Dubrovnik-1950 can ask an AI to help him out if this is not obvious?

Elroch

The history of the rules of chess is interesting.

Before 1561 there was no repetition draw rule and no 50 move rule. Ruy Lopez introduced the latter (there was some disagreement about how many moves it should be). So my "basic rules" chess was the version that predates that. Chess with a 50 move rule and no repetition rule was the norm until 1871, when the first repetition rule was introduced (a five-fold repetition rule, later modified to three-fold). This version from 1561 to 1871 has the nice properties that a Forsyth FEN position is the game state, and the number of states is only 50 times bigger than the version for the version before 1561.

It's only the post 1871 versions where Forsyth notation is no longer adequate to describe a game state, because of the repetition rule.

MARattigan

In fact the basic rules since 2017 no longer contain 50M/3R rules, so not only are the players not obliged to claim under the current basic rules, they're not allowed to.

Elroch

But they are allowed to agree a draw, I presume?

Note that from a mathematical point of view, in basic chess (pre-1561) with its infinite games, we can define any infinite game to be a draw. This is convenient and works fine - it does the job for tablebases, for example, as once you have stopped adding positions with a value, you can be sure all the others lead to infinite games with perfect play, so are draws - but it is a tad impractical for real world use! wink.png

Elroch
Dubrovnik-1950 wrote:

"draw is not automatic when there is a threefold repetition or 50 moves without a pawn being moved or a piece being taken. While the players playing one of the games described COULD claim a draw at some point, they are not obliged to, and that is what matters to the count."- Elroch

Why can not a chess game last forever under fide rules?

Left to the players it can, when claims are required. What can stop it is tournament directors stepping in.

As I said, this is not true of chess on chess.com, where players are assisted by automated claims. (Just noticed an interesting point that in a situation where both players want a win - eg to win a tournament - the automated claim might be unwelcome to BOTH of them! But it is still there)

MARattigan
Elroch wrote:

But they are allowed to agree a draw, I presume?

...

Yes - art.5.2.3 the last art. in the basic rules.

Elroch
crazedrat1001 wrote:

This Dubrovnik character cannot come up with an original thought to save his life.

AI does require some intelligence to interpret and use correctly. I'm afraid this new mindless spamming of AI answers (in this case not asking the right questions) is probably going to become commonplace.

Yes, it is odd how @Dubrovnik-1950, asks a simple question of the AI, copies and pastes the answer (possibly without even reading it), adds nothing and then believes he is being clever by doing so.

Might I contrast this with say my original contribution in #19451

MARattigan
Dubrovnik-1950 wrote:

It has been said!

But that game is over!

Your nonsense will be shown to be false with Grok!

Rather Grok has been shown to be false with our nonsense.

(You almost got it right. Keep trying.)

playerafar
MARattigan wrote:
Dubrovnik-1950 wrote:

It has been said!

But that game is over!

Your nonsense will be shown to be false with Grok!

Rather Grok has been shown to be false with our nonsense.

(You almost got it right, but it makes a difference. Keep trying.)

Grok seems to have a definite edge over chatgpt which has an edge over copilot.
But Grok makes some mistakes too.
Last night I spent some time with Grok discussing the position following:
1) Nf3 Nf6 d4 c5 dxc e6 ....
And it made several mistakes but did carry on some good conversation.
Example of mistake it made later - suggesting Bxc5 was an option for black instead of e6 - which of course is impossible without e6 first.

AGC-Gambit_YT

what happened when I was gone lol

playerafar
Elroch wrote:
Dubrovnik-1950 wrote:

"draw is not automatic when there is a threefold repetition or 50 moves without a pawn being moved or a piece being taken. While the players playing one of the games described COULD claim a draw at some point, they are not obliged to, and that is what matters to the count."- Elroch

Why can not a chess game last forever under fide rules?

Left to the players it can, when claims are required. What can stop it is tournament directors stepping in.

As I said, this is not true of chess on chess.com, where players are assisted by automated claims. (Just noticed an interesting point that in a situation where both players want a win - eg to win a tournament - the automated claim might be unwelcome to BOTH of them! But it is still there)

Implication of draw not being automatic on 3 fold repetition - that means there could be up to 25-fold repetition apparently. Or 16 repetitions of 3-fold.
So if neither player claims a draw or a draw is not agreed - one of the players could fall on time.
There's probably been many arguments in live tournaments about whether there was 3-fold - or 50 - in time scrambles where players aren't noting the moves on paper.
I haven't seen one though.
On chess.com maybe its automated? Or you have to hit the draw button perchance ...

MARattigan
Dubrovnik-1950 wrote:

The bottom line is if the players are morons, the game will still end under fide rules by force.

Fide Rule...

If 75 consecutive moves have been made by each player without a capture or a pawn move, the arbiter is required to declare the game drawn, regardless of whether either player claims it.

Art.9.6.2 in fact.

If you could read you'ld be able to see that the FIDE rules define several different games and that the basic rules game stops at art.5. The 75M rule and also the 50M and 3/5R rules appy only to the various competition rules games.

MARattigan

I was convinced that never in my lifetime would an AI reach the intelligence of a human. You have changed my mind.

Elroch
MARattigan wrote:

I was convinced that never in my lifetime would an AI reach the intelligence of a human. You have changed my mind.

Picking the right human is key.