Chess will never be solved, here's why

I didn't say me.
tygxc claimed he knew more than anybody here.
You didn't read properly.
But that's okay. I can ignore.

I think tygxc's claim that he knows more than everybody here should not be ignored.
It reveals motivation.
Plus its a false claim.
Proven by others more knowledgable constantly. Or more logical. Or both.
I knew instantly years ago his claims were false as soon as he tried to claim that the fundamental speeds of computers involved 'don't matter'.
And again when he 'took the square root'.
--------------------------------------
My point is what his posts are really 'about'.
And when he falsely claimed 'knows more' - a big chunk of 'what its about' suddenly revealed. Very reminiscent of the 'absent person'.
He's still a 'foil' (in a positive sense) to the discussion though and unlike that absent person - keeps the forum mostly 'on topic'.
Its just that when anybody asks 'why is tygxc doing this' it seems he has now provided a lot of the answer.
As you say @Elroch it can be 'less personal' now - because that part has been partly 'solved' now. No pun intended.

I don't get it either. I believe there is probably no practical calculation. But I am not sure.

Nobody has been able to figure that one.
In the meantime - a lot of progress with engines 'playing better' though.
Not perfectly. Just better. Better than humans can. Including faster.

AIs kind of try to solve this. They learn the probabilities of each outcome for a position. This can then be applied to the first position. But their probabilities are more like practical outcomes than theoretical ones. Eg (click to enlarge)

Yes, that's pretty well no uncertainty.

I was also thinking - how do computers do against humans at poker?
Poker AI is quite strong. And the more a known human opponent plays a poker engine the better the computer plays against that human player.
But apparently AI doesn't have quite the edge in poker that it does in chess - although the poker engines continue to improve.
An article on it here:
https://www.reddit.com/r/askscience/comments/2d6my1/can_computers_beat_champions_in_poker/

la verdad solo se español

quien se une

comenten los que saben español

Maybe he's saying 'the truth can only be known in Spanish'.

Yes, it's like a coin flip. It can only be a head or a tail. No uncertainty. Except the uncertainty whether it is a head or a tail.

I know this sounds inane - but the coin could stand on edge.
Or be vaporized on the way down.
Or land in quicksand.

Always such relevant points...

It is so easy to make money using this

Nice little passive aggressive insult there.

@11323

"distinction between a weak solve and strong solve but it isn't relevant"
++ The difference is relevant. Checkers for example has been weakly solved, not strongly.

"a weak solve means only the games derived from perfect play are known but it still isn't known the exact result of every possible position reached through non-perfect play (ex computers solve that Ne4 in whatever position will result in a win, but doesn't know whether all other positions after not playing ne4 are losses or draws, but is nonetheless irrelevant)"

++ No, that is not true.
Weakly solved means that for the initial position a strategy has been determined to achieve the game-theoretic value against any opposition.

@11317

"it's like a coin flip. It can only be a head or a tail. No uncertainty."
++ Indeed. Chess either is or is not a draw.

"Except the uncertainty whether it is a head or a tail" ++ There is no uncertainty: chess is a draw. You cannot say it is 99% sure to be a draw, because by definition there exists no real or thought experiment that yields 99 draws and 1 decisive game with optimal play by both sides.

Likewise you cannot say the Riemann Hypothesis is 99% true. It is either true or false.
Schrödinger's cat can be 50% dead, 50% alive, because you can open 100 boxes with Schrödinger cats and find 50 dead and 50 alive.

Truth and proof are distinct. Fermat's last theorem was true all the time since Fermat formulated it, and did not suddenly become true after Wiles proved it. Besides Fermat might have had a proof as he asserted in his marginal note.

@tygxc, Bayesian probability - the only type relevant here, and not the Frequentist paradigm to which you refer - is about quantifying belief. Read that again and try to understand it.

You need to read the following carefully (and critically, if you wish, but I suggest not banging your head against a brick wall, so to speak) because it is clear to everyone that it is a gap in your understanding. You are unaware of Bayesian reasoning, and familiar only with the separate Frequentist paradigm, which is irrelevant to this discussion. While both paradigms refer to numbers called probabilities, these numbers refer to different concepts in the two paradigms - it's only the word (and some of the maths) that is the same.

At every point in time, you have a state of belief about the truth of a proposition. This is a probability ascribed to the truth of a proposition. This probability is a function of all of the relevant information you have (combined with a very uncertain prior belief from a position of zero or little knowledge). A good way to conceptualise the value is that it is the one which you would consider fair for having a bet on the truth of the proposition, the value for which you would be no happier on one side of the bet than the other.

It is inarguable that your state of justifiable belief changes in steps based on the information you receive. This means that the probability you ascribe to an event changes. (Bayes rule is a systematic way of changing your state of belief, the only consistent one with general conditions, but we do not need to delve into the technical details).

A key point is that it is very hard to justify moving from a state of uncertain belief to a state of certainty. It can only occur when the belief becomes a logically proven consequence of known facts (Boolean logic is an extreme special case of Bayesian reasoning).

This can occur for example at the end of a mathematical theorem when the result is proven. The same occurs when a weak solution of a game is completed. This can be viewed as the completion of a logical deduction - it is the point when every single possible counter has been covered and there is none left to be concerned about. To be very precise, you can think of the step from uncertainty to certainty as occurring when the number of nodes left to analyse falls from one to zero.

You claim certainty about the result of chess. Initially you had no reason for certainty about the result - all you knew were the rules of chess and the results that could occur. You could only have had a guess at the result. So of course, the justifiable state about, say, whether the game was a draw was some probability p, with 0 < p < 1. Those inequalities are strict.

The above, with no assumptions about Bayesian reasoning, means that in order to reach a state of certainty, you would have had a state of uncertainty and then received a single piece of relevant information and became certain.

What was that step? If you cannot identify it and justify that this single piece of information had the pivotal significance necessary to cause a radical change in belief state you have confirmed that you are fooling yourself and confusing strong belief with certainty. Given the facts about the problem, this is the only conclusion available. If not, tell us what that magic step was.

You can choose to learn a powerful concept and principle from this. I am sure you will not - you will remain obstinately and arrogantly wrong.