jefk wrote: ↑Sat Nov 11, 2023 4:19 pm
yes syzygy, for practical purposes (eg. in physics) we can regard pi as a rational number; in physics numbers are not infinitely long.
And I am not a physicist but a mathematician. In mathematics, a proof has to be fully rigorous. The term "ultraweakly solved" refers to mathematically proving that the starting position is a win, draw or loss for white. Overwhelming numerical evidence is meaningless.
As for a proof that pi is irrational, such proofs have been stated with reduction ad absurdum arguments (eg. Chudnovsky); if you think that is insufficient proof, and for a real 'proof' you want to calculate all decimals of pi to infinity, well good luck.
The Chudnosvky brothers have calculated very many digits of pi, which has nothing to do with a proof. That pi is irrational was proven by Lambert in 1761 by giving a rigorous argument.
As for chess, there is a reductio ad absurdum argument, namely, that if White despite have a little first move advantage) cannot win, then it must be a draw. Think of balanced games where the first moving side can only win if the second side (Black) makes a mistake. It's becoming more obvious every year, that chess is such a game.
Nothing to do with ultraweakly solving.
Thus, i doubt that for ('ultraweakly') solving a game, in all situations you have to calculate all (or most) variations from opening to endgame, like was done for checkers. I don't exclude the possibility that a brilliant game theorist in a few years will claim there's a proof that chess is in some way is a solved game and it's a draw. Then there's not need for an 'ultraweak' (calculated) 'proof'.
There are games that can be proved to be a win or a draw by a clever argument. I am very confident that chess is not such a game.
In chess, it is far more likely that white has a very ingenious winning strategy that is completely overlooked by humans and engines than that chess allows for some very ingenious "short" mathematical argument that proves it to be a draw. But I cannot mathematically rule out either possibility, so I cannot prove that chess is a draw, and I also cannot prove that chess will never be proved to be a draw. But I can claim that chess will not be ultraweakly solved just as I can claim that chess is a draw.
Besides the reductio ad absurdum argument (according to AI):
1)Topological proof: Chess has been proven to be a drawn game topologically based on properties of the graph of legal positions/moves under the no-repetition rule.
(Note checked the above, as it was unknown to me &unlikely;
see this for further info, it's above my current knowledge
https://poe.com/s/BTWfJWAl5SsvNv8CTeuq)
This is just nonsense. By the same non-argument, any chess position is a draw.
2)Strategy-stealing argument: If White could force a win, Black could emulate White's optimal moves and obtain at least a draw, a contradiction.
(note this is stated a bit too simple, imo, but it's clear that there's a strategic equilibrium in the game of chess)
Does not work for chess.
If chess allowed null moves, then you can prove that white cannot lose by a strategy-stealing arugment. Because if you assume black has a win, then white can steal black's winning strategy by playing a null move first (contradiction, so black has no win).
But obviously chess does not allow null moves.
2) Minimal winning margin: No line of perfect play has yet proven White can build an advantage beyond the 0.5 pawn margin indicative of a draw at highest levels.
(note this is true, and again, this makes it very likely there's not winning strategy for White, thus, applying Zermelo, it's a draw)
Anything that relies on "0.5 pawn margin" is not a mathematical proof.
PS it's easy to make strong statements like 'chess will never be solved
Sure, but I know what I am talking about and I am very confident that I will not have to eat my hat. So I am not talking about anything that convinces you that chess is draw, I am talking about a mathemtical proof. I am confident that the only way to prove chess is a win or a draw or a loss is by such huge amounts of computation that we simply lack the resources (including time).
I agree with you that chess very likely is a draw, and this only strengthens my confidence. Computationally it is much easier to prove a win than it is to prove a draw.
Don't make claims about things you don't know, rather say you suspect something (will turn out to be such and such in future, or not).
I see no problem with making claims. If in a few years or decades someone comes with a mathematically sound proof (an actual proof, not a bogus crackpot proof), then I will happily admit that I was wrong.
I suspect that in a few years it will be acknowledged by most game theorists that chess is a draw, and that 0.33333333333333 ad inifitum is 1/3
and that 1/2 + 1/4+ 1/8 + 1/16 etc is one aka 1.0000
No chance. Game theorists are not interested in numerical evidence. Game theory is not physics. "Ultraweakly solving" means mathematically ultraweakly solving.