Chess will never be solved, here's why

Yup optimissed lol

@11476

"But there could be an odd number of errors in every draw."
++ If there were an odd number of errors in every draw, then Chess would be a white win or a black win. There are many other arguments against that, but leave those for now.
Say the error distribution for the 107 draws were 0 - 70 - 0 - 37 - 0. Then the same argument applies: it is not plausible to have 70 games with 1 error and 37 games with 3 errors and none with 2 errors.

"The simplest alternative possibility is that all the first moves played are errors."
++ This is clearly nonsense. So you think 1 e4, 1 d4, 1 Nf3 only draw and 1 a4 wins for white.
That goes against all human knowledge acquired during centuries and independently corroborated by AlphaZero with no other input but the Laws of Chess. Figure 31

"You are at this point thinking..." ++ No, I am not thinking as you think I think.
I am not thinking in terms of probability, but in terms of logic, combinatorial game theory.
I observe the 107 draws and conclude from those 107 sequences of legal moves from the initial position to a draw.

"This is clearly nonsense"

ah yes, more claims without proof

" So you think 1 e4, 1 d4, 1 Nf3 only draw and 1 a4 wins for white.
That goes against all human knowledge acquired during centuries and independently corroborated by AlphaZero with no other input but the Laws of Chess. Figure 31"

ah yes, more strawman arguments. the fact that you cannot prove otherwise to their statement makes your argument invalid

""You are at this point thinking..." ++ No, I am not thinking as you think I think."

no we know exactly what you are thinking. you just are deluded in thinking that you are arguing on any sort of logical plane. you are presenting a sheep and calling it a moose. everyone else is telling you that you are arguing a sheep, and you thionk that they are misunderstanding you because they dont agree that you have a moose.

"I observe the 107 draws and conclude from those 107 sequences of legal moves from the initial position to a draw."

this is objectively a probabilistic/heuristic argument. we are calling it that because that is objectively what it is. you argue that the whole is derived of the structure of a part of it, without proof that the rest of chess has to follow by the determinations of the engines. thats objectively an heuristic/probabilistic argument

isnt it funny how tygxc still refuses to look at the four points I listed?

Don't know about funny, I think it's just getting tedious.

I have invited him to look at several counterexamples to the first point well over a dozen times in big bold yellow letters. Zilch.

maybe we should just keep repeating the points that tygxc ignores.

funny was sarcasm

With O (that's Optimissed) back from his six days of being muted by chess.com staff - tygxc will look good and watch O get the negative attention that O loves to hate.
While tygxc continues to do his thing.
And what is that thing?
He knows something about mathematics and something about computers. ----------------------------------------
So where are tygxc's disinformation and denials located exactly?
Apparently its essentially denial of logic by tygxc.
He uses math out of context or inappropriately or invalidly to deny logic and otherwise denies logic in order to push disinformation about computers and to distort and crassly falsify their abilities as to chess for yet another purpose which is to attract disagreement and refutation with the idea of crassly continuing with his denial of logic for a purpose of self-assertion apparently.
-------------------------------------
Would it be like walking along with a Tshirt message proclaiming 'the earth is flat you know'. ,,,??
That would get attention too. Simpler though.

'gets solved' ...
trillions of trillions of years from now. Humanity would need to survive that long and also to care enough about such a thing?
-------------------------------------------
e and pi are transcendental numbers.
They transcend.
There are infinite series to approximate them ...
but they aren't like other numbers - you never quite get there.
Their numerical 'value' goes on forever.

@11531

"nobody can memorize every possible win"
++ Chess is a draw. Indeed, nobody can memorize every possible draw.
However, it helps to know some of the ways to draw.

FOREWORD

Irrespective of the tortuously inept output of some poor souls here, of which more may be said later, the real situation is given here.

There are two camps. One camp is led by someone whose ego is the size of Mt Everest, without an intellect to match. It consists of the belief that chess theory, expressed mathematically, will solve all and is indeed the only way forward. This rests on the hare-brained idea that chess will be solved mathematically, therefore we can guess what the result will be. This must be kept secret from those approaching the subject with any degree of honesty, which is why I'm pointing it out. All we are likely to see is a very dense smokescreen.

The leader of the cult of mathematics has been known to claim that there is no way that solving chess can be regarded as a scientific project. This, of course, may not be part of the doctrine, except inasmuch as it may be pretended as true, from time to time, in the cause of expedience: for instance, when he's losing or has lost an argument. He loses arguments so frequently that "special methods" have been built into the received doctrine of the cult, in order to deflect outsiders from discerning that he is fairly useless at debate.

Since chess exists as an entity, albeit one constructed by humans; but since it hasn't been solved; the only way forward is to approach it scientifically. That means it's necessary to devise tests to try to discern what should be the result of best play (or good play). At first, there are no "shoulds" in science. It is only when scientific investigation proceeds sufficiently that definite correlations are discerned, which may be interpreted as "cause and effect on the chessboard", that a good scientist will adopt the "should" approach, in that s/he is testing theory, which in the first case is created as hypothesis. If and only if (iff) the results are such that they support the theory, can we say that they have occurred as they "should" have done. It other words, the results support the emergent theory in that they conform to those results that the theory would predict.

EPILOGUE

I wrote the above for a bit of fun. It goes without saying that there are several people here who won't be able to understand it because they don't have the intellectual equipment to do so. Therefore they will attempt to portray it as nonsense. That is for two reasons. They are obviously unintelligent, since if that were not true, they would be happy to concur with a pretty good bit of writing which points out some truths rather accurately. But also, they don't want other people to understand it because then, those people would change their minds about who is right in this little discussion.

They would realise that, despite the occasional mistake, tygxc is correct in the large picture and his detractors are completely wrong and are merely ego-driven. Thankyou.

Well said.

@11534

"the sheer number of possible positions" ++ There are 10^44 legal chess positions,
of which 10^37 without promotions to pieces not previously captured.
Of these 10^17 are relevant to weakly solving chess, as Schaeffer did for Checkers.

"calculate every potential outcome" ++ It is not necessary to calculate everything:
weakly solving chess only needs 1 black reply to all reasonable white moves.

The fact that games have been solved shows that the scientific/empirical approach espoused by @tygxc with its lack of certainty is not the only approach. By definition, all solutions are mathematical in character, not scientific, and fall into the classes of:

1. ultra-weak solution - proof of the value of the result with optimal play
2. weak solutions - (generally computer assisted) proofs of the value of the result with optimal play by proving specific strategies are optimal
3. strong solutions - proofs of the value of the result with optimal play from every legal state.

Mathematical solution includes rigorous proofs incorporating computer checking of examples - famously an early example of these outside of game theory was the 1976 solution of the 4 colour problem, showing (in intuitive terms) that regardless of the nature of the boundaries, any map can be coloured with 4 colours. The proof incorporated computer checking of a large (but finite) list of examples too numerous for a human to check.

The reason mathematical solution is the type that is of interest is that a rigorous solution of a game is an entirely specific endpoint. It is either achieved or it is not. It is not yet achieved for chess, and the size of the problem makes it beyond what is currently feasible.

By contrast, there is not such thing as "the" scientific solution of chess - there is a continuum of uncertainty regardless of what (empirical) methods are used. The uncertainty can go down but it is impossible for empirical, non-rigorous means to achieve the magic step where uncertainty goes to zero - that distinction is exclusive to rigorous mathematical solution. I understand that not everyone here understands that small numbers (probabilities) are not zero, but that is undeniable.

Even @tygxc vascillates between assertions that chess is definitely a draw and describing a multi-year program of empirical exploration for the purpose of convincing himself that chess is a draw (comically achieving this while ignoring 17 of the first moves for white). Neither is final. Neither is certain.

I think that if you would pause and think for yourself for a moment, you would realise that in order to find that one black reply to each move by white, it is necessary to investigate very thoroughly quite a large number of black replies to every white move. I mentioned five candidate moves earlier (a couple of weeks since) but I also mentioned that to do the job properly, maybe nine candidate moves would suffice. That would be your "weak solution".

If your interlocutors of the "mathematical school" were more capable, they would raise salient points, such as I have raised here, rather than waffling on with some kind of completely subjective idea, as in post #11539. I am convinced you are right in the big picture and of course, I share your belief that we know that chess is a draw with good or best play by both sides. I think you greatly weaken your case, though, and I also think that the people in the mathematical school are not really capable of picking up on errors, since they are rather too obsessive regarding the "pi-in-the-sky" solution.

Yes, that is a good working definition.

reminder that tygxc makes false calculations here.

first, the 10^34/10^17 are relatively unsubstantiated heuristics, but thats not the main issue.
the main issue is that he's misinterpreting his own 10^17 number as all relevant positions, when thats actually just the number of positions in the solution.
tygxc calculated the 10^17 as the solution set for weakly solving chess. far more positions than that are actually relevant, as WHICH positions are the solution set needs to be calculated.

@11539

"empirical exploration for the purpose of convincing himself that chess is a draw"
++ No, chess is a draw, I am fully convinced.
The purpose of weakly solving is to establish how to draw.
The 107 ICCF WC Finals draws indicate black has not 1, but several ways to draw against whatever white tries.

"ignoring 17 of the first moves for white"
If black can draw against the 3 best moves,
then a fortiori black can draw or even win ( 1 g4?) against the 17 lesser moves.
That is Best first or Alpha-beta pruning, also used by Schaeffer in weakly solving Checkers.
That 1 e4, 1 d4, 1 Nf3, 1 c4 are the 4 best moves has been established long ago.

Steinitz and Lasker were mathematicians, Capablanca dropped out of engineering study.
'From the outset two moves, 1.e4 or 1.d4, open up lines for the Queen and a Bishop.
Therefore, theoretically one of these two moves must be the best,
as no other first move accomplishes so much.' - Capablanca

'Chess is a very logical game' - Capablanca

This was independently corroborated by AlphaZero, with no human input but the Laws of Chess.
AlphaZero used 4 training seeds of each 1,000,000 training steps and arrived at
1 d4 > 1 e4 > 1 Nf3 > 1 c4 > 1 e3 > 1 g3 > 1 Nc3 > 1 c3 > 1 b3 > 1 a3 >
1 h3 > 1 d3 > 1 f4 > 1 b4 > 1 Nh3 > 1 h4 > 1 Na3 > 1 f3 > 1 g4

That is game knowledge. It is no heuristic, but logically follows from the Laws of Chess.

"++ No, chess is a draw, I am fully convinced."

Never said it wasn't we said you don't have prove just because "no mistakes " the number of errors is still unknown  what makes you think they thinking longer will cause some to have no mistakes all that does is that the mistakes they do make are really obscure and hard to counter 

And also don't forget that engines always go for a draw so using engines +humans as prove isn't very good cause most of the time stockfish likes drawing moves 

Or did you just suddenly forget how stockfish thinks

Like other people have said it is well known that chess is most likely a draw but we don't have evidence 

We have scientific evidence but not mathematical. What is important in the world is scientific evidence. Maths just crosses the t, dots the i and in Victorian times they built great bridges by over-engineering. Maths allows us to be more precise with our scientifically based estimates but maths alters no realities.