Chess will never be solved, here's why

Id like to point out that we could just sit here and continuously repeat the core fallacies that you make tygxc in your mental delusion of what logic is, but we are playing along for the sake of your convenience and those around you.

you could at least do us the favor of actually interacting with the questions posed. because elroch has asked this question to you before, multiple times, with many different forms, and you completely ignored it. you didnt even recognize it's existence.

MEGA its multiple choice.
T types the false choice or choices.
Others type the True choices.
Side-feature - there's nobody 'marking' the exam for pass and fail.
Can you imagine T picking his own answers in a real Exam?

Thank you for giving the specific name. I chose it as an example of zugzwang, of course.

So the fact that every single came in the ICCF finals ends in a draw increases the Bayesian probability that it's a draw with perfect play - not to 100% but to somewhere close.

You don't seem qualified to comment either way.

Fixed.

Seems the usual suspects have treked over here to start with their adhominen attacks (not including Elroch in this pool), and one of them could not resist mentioning the climate. The nit picky math analogies aren't even necessary as solving chess via a table base would be numbercrunching and retrograde analysis (working backwards from checkmate positions although I don't understand how that helps calculates the best forward moves), it's not a mathematical proof. Chess has mathematical patterns but the solution isn't mathematical.

Would also like to note that top computer programs drawing each other is the weakest most invalid evidence possible for chess being a draw. First of all they are programmed by humans, and ironically have shown in top engine matches that many of our "generalized" concepts about open files, doubled pawns, central control, tempo, "solid" positions don't even end up holding up in these 3500 to 4,000+ rated games. They run into the horizon issue, which causes them to fail to understand even some "easy" puzzles that take alot of moves to win, and even fails to understand some semi-realistic complicated endgames. Most chess programs have a limited amount of AI to actually "understanding" chess positions vs numbercrunching. Even numbercrunching, some algorithm has to understand whether the final position at the calculation horizon is "good" or not and thus whether to discount the line or not. And that can only be best estimated with human programming. If the next decisive move choice is one move pass that, it won't be accurate.

Well, not in our lifetimes, but in the future, with highly developed computers probably one day!

@11341

"1 tempo is certainly enough to win, in given situations."
++ Yes, and 1 pawn is sometimes not enough to win, in given situations.
I mean +1 tempo near the initial position is not enough to win and +1 pawn near the initial position is required to win.

White should still be able to draw this

And white should be able to win this

With the current information we are dealing with in this thread, it's very hard to estimate probability for those games being "error free". It's certainly impossible to come up with a % that is near 100 at the moment.

We've discussed this here before, but for better estimation we would need more information on the relationship between double error games and single error games played by engines of current strength. Tygxc estimation of >99% doesn't account for this. That'd bring us closer to a valid estimation.

Even then - If we determine these games are likely error free, that of course doesn't lead us to the conclusion that chess is certainly a draw.

Engines of similar level tend to draw despite being imperfect. An engine making no errors against a similar level of play does not mean they'd make no errors against an engine that has access to perfect information.

That's how far these games are from a proof of chess being a draw.

@11358

"If we determine these games are likely error free, that of course doesn't lead us to the conclusion that chess is a draw." ++ If only 1 drawn game is error-free, then chess is a draw.

"engines of similar level tend to draw" ++ ICCF is not engine vs. engine.
It is human ICCF (grand)master + engines vs. human ICCF (grand)master + engines
and at average 5 days/move.
Engines of similar level tend to draw more at more time/move as they approach perfection.
ICCF WC Finals games tend to draw more over the years and have now reached 0 decisive games, perfection.

"That's how far these games are from a proof of chess being a draw."
++ So in your opinion the 17 ICCF (grand)masters and their engines all collude to make exactly 1 error, never 0, never 2.

Here's a totally random chess game I constructed, to put it in perspective. Only 67 moves, now imagine all the different variations that could have occurred within those 67 moves. Now imagine 1,000 moves. 10,000 moves, how many random chess games exist. It's unimaginable. The number of total possible chess games must be in the powers of a googol. Maybe 10^1,000,000, who knows, impossible to even estimate. I know computers optimize the more realistic moves and positions but the point is I don't even know if quantum computers would be enough to solve every possible game and position.

@11360

"The number of total possible chess games must be in the powers of a googol"
++ Much more: between 10^29241 and 10^34082.
However, the number of legal chess positions is 10^44.
The number of legal chess positions from a box of 34 chess men including spare queens of both colors is 10^38.
Of these 10^17 are relevant to weakly solving chess.

A side question, it would be interesting to set up a random move generator that plays out totally random moves for both sides until the game accidentally ends. No repeating positions (unless it's the opposite side to move in the same position) so no draws by repetition, no repeating allowed, but other than that it just selects a random move out of all the legal options until the game end up in a checkmate, stalemate, or insufficient mating material position..etc, how many moves do you think it would continue?

How would one "solve" chess lol. It's an arms race with development, development will never end.

@11360

"No repeating positions (unless it's the opposite side to move in the same position) so no draws by repetition, no repeating allowed, but other than that it just selects a random move out of all the legal options until the game end up in a checkmate, stalemate, or insufficient mating material position..etc, how many moves do you think it would continue?"
++ It is unlikely to accidently stumble into checkmate or stalemate by random moves, and if you remove the draw by 3-fold repetition, then most games are likely to draw by the 75-moves rule somewhere between 75 and 8848.5 moves, e.g. around 815 moves.

@11361

"How would one "solve" chess" ++ Just like Schaeffer solved Checkers.
It is now ongoing.

@11364

Chess is for nerds.
Mathematics is for nerds.
The intersection of both is for nerds.

"So in your opinion the 17 ICCF (grand)masters and their engines all collude to make exactly 1 error, never 0, never 2."

They don't have to collude. Again, the problem in your thinking is considering all errors equal. If you combine two important factors - Very few available winning lines from the starting position and engines of similar strength playing each other - The first error is extremely likely to occur and lead to a draw.