Will computers ever solve chess?

Btickler: Prove to me that grandmasters do not assume a draw because it is simply expedient for prep/analysis to do so, given that their goal is to win a game/match/tournament, not to solve chess.

 Grandmasters cannot solve chess, try or not try, consciously or not consciously. So the point of what grandmasters do or not do, hidden agenda or not, is not related to the question of solving chess.

 Remove this whole argument completely, what they assume, what they think, what they know, what they believe, how confident they are—it is irrelevant. 

 Scratch it out.

Ok, that's reasonable.

But at the same time, seems you have a severe case of SIWOTI disease

  The rest is conceit, as Btickler previously noted.

[a whole heap of BS]

I have annotated all my games and was able to point out errors and good moves.

By and for your own limited understanding, yes.

You are correct that not all grandmasters assume chess is a draw. I know of one who thought chess might be a win for White.  I could have said "more than 99%"

No, you couldn't.  You have zero basis for making such a claim.  You are assuming that all GMs that have not spoken must agree with you.  I only know 1 person that doesn't think you are crazy...does that means over 99% of people think you are crazy?

I cannot prove this to you as you do not have enough chess knowledge to understand my statement is true

Nor do you have enough knowledge to claim it is true.  What you have is just enough (but fading further every day) knowledge to be an a$$hat about it.  There's not a damn thing special about your opinion on this thread, Ponz.  You are just a schmuck like the rest of us in terms of this particular debate.  You don't understand 99% of GMs, and you don't have any secret proof that they lack...what you seem to have are fading memories of better days and not enough Omega 3 fatty acids in your diet.

high very high--used in Law. Used in science.

Wow.  You really channeled Trump there.

I did not say they thought chess was solved.

If you claim is that 99% of GMs know that chess is a draw, and that this is part of conclusive circumstantial proof, then you are claiming chess is solved and worse, claiming on GMs' behalf for them that chess is solved.  I don't get how you still cannot fathom that these 2 things are tied together.  If chess is actually proved to be a forced draw conclusively, then chess is solved.

oh you think me making a spelling mistake makes me wrong?

That's not a spelling mistake.  It's a lack of education.  Specifically, a lack of understanding of the difference between nouns and adjectives.  Typing "teh" instead of "the" is a spelling mistake.

A lack of education colors all your arguments.

I used his quote because he put it in words better than I did.

What does this say about you, Ponz?  You're a much better chess player.  You have "researched" all the GMs games and know what they think!  You have been working on this proof for a long time...then some guy shows up in the thread and says it better than you can immediately upon arrival?

Perhaps you should not be appointing yourself a spokesman for your position.  You don't seem to be very good at it.

To me being 99.9999% sure of something is to know something.

Fine.  They are the same thing, then.  Stamp it.  Mark this spot.  Next time you say "I have never claimed to know that chess is a draw 100%", people are free to ridicule you for flip-flopping yet again, right?

Really my opponents have authored chess books?

I don't know, but then, neither do you.  You have this assumption, though, that your opinion here on this topic is just operating/existing on a different level than everyone else.  It's not...you're right here with the rest of us...and I hate to break this to you, but there is zero barrier to entry to publishing a chess book.  Any dumbass can write a chess book (and many. many dumbasses do).  Chess books are moving the way of the dinosaur for this very reason.  Most players get more out of a single YouTube video these days.  I'm almost tempted to write a book for Kindle that debunks all your BS and sell it for 99 cents, but...damn, that is just way more work than anything you are saying could/would ever be worth.  It would be funny, though.

There is a very clear progression here. You may not see the progression but it is there.

Progression towards what?  TCEC and WCC matches have gone as high as the mid-80%, so your 66.7% is proving what, again?  You said that A0/Stockfish was 75% draws.  If Alpha Zero/Stockfish is the highest possible level of engine play (it's not, but you are claiming it is as evidence here), then why wasn't it 85%-90% draws?  If you are going to pretend to establish your trends based on a small number of data points, then you need to stay consistent.

However if you have a whole lot of circumstantial evidence it is very good proof.

You need to add the word "decent" in front of "circumstantial evidence".  There's a big difference between DNA-level evidence and "there were Coke cans in the trash can nearby and the defendant has been known to drink Coke".  Let's just say you have not produced anything anywhere even close to a parking lot a mile away from the ballpark of the reliability of DNA evidence.

Might or might not be able to prove it.

Nope.  Cannot prove it.

one more small piece of evidence

Nope.  Saying that if a forced win were possible someone would have already done it and become a multimillionaire is not *evidence*, Ponz.  It's not even a credible *opinion*.

I'll leave the rest of the post for others to pick over if they want to, but honestly, the most damning evidence of your published proof is that nobody...not a single person...is stepping forth to proclaim your triumph now that you have finally tried to explain yourself in detail, Ponz..and before you even try to delude yourself further, it's not because everyone is beneath you and cannot understand what you posted, which is pretty basic.

It's relevant in pointing out Ponz's lack of logical thought.  Even his *bogus* evidence does not hold up to the most basic scrutiny.

I'm not sure I even care enough to find out what that acronym might be.  I do this here, in this thread...nowhere else.  Consider it a public service...just plucking the weeds.

  Oh, I see. It’s like:

 Let’s just look at his logic, not relevant for the question of solving chess, but let’s examine his logic. And irrelevant as it might be, it does not hold water.

  Quite right, sir.

someone is wrong on the internet

Wow! let's see who gets 10,000!

Re #1:

How the game ends isn't part of the criteria. It's only necessary that both players play perfectly and aren't aware of their opponent's infallibility. Since the question is about perfect play, the outcome of the game would be known if/when the question is answered (unless of course there are no perfect games with 16 moves).

In such a game White plays to win, but if he cannot win he plays for a draw. If he can neither win nor draw he plays to survive for as long as possible. Black also plays to win, but if he cannot win he plays for a draw. If he can neither win nor draw then he plays to survive for as long as possible.

(fyi): I forgot where I first posed this question. I believe the discussion started where it was suggested that chess might be easier to solve if some perfect games exist that are shorter than what Shannon assumed in his paper. I consider either 31 or 32 plies to be short compared to what Shannon assumed. Even 20 or 25 moves might be considered short. There's no strict basis for a cutoff as far as I'm aware.

Re #2:

A little but not much. I was curious if any expertly-played games ended in a short number of moves. One is Carlsen vs. Anand:

1.Nf3 d5 2.g3 g6 3.Bg2 Bg7 4.d4 c6 5.O-O Nf6 6.b3 O-O 7.Bb2 Bf5 8.c4 Nbd7 9.Nc3 dxc4 10.bxc4 Nb6 11.c5 Nc4 12.Bc1 Nd5 13.Qb3 Na5 14.Qa3 Nc4 15.Qb3 Na5 16.Qa3 Nc4 1/2-1/2.

Although it ended by agreement, I did find it interesting that it simplified enough that both sides believed there was little chance to win so they called it a draw.

Re #3:

If I see an intriguing comment by anyone I may ask if it's opinion or proven knowledge. I mentioned it because you asked why I pointed out that there's a distinction between speculation and provable statements.

Congrats on being an xkcd fan.  I don't find it applicable in this case.  People are wrong constantly...couldn't care less.  It's more about pervasive annoyance.   It's kind of like trolls, purveying a form of lowbrow malice that the world could/should do without.  Feed them into a woodchipper or just preempt them entirely by testing for a snivelling weasel gene...everyone's better off.  Less annoyance and more resources for people that contribute something worthwhile to the world.

For threads like this, I only really engage with a non-troll when someone tries to spread misinformation around obnoxiously.  Physically, I don't believe in eye for an eye, but in the world of internet douchebaggery, I think it's fine to give people a mirror of what they like to toss out themselves.  If chess.com actually moderated the forums, there'd be a lot less of it around...but they don't.  Erik must have been a forum troll in his teens or something...he's got a soft spot for them.

Meh, Erik is just a businessman. Don't expect him to think too deeply.

Internet douchbaggery mirrors? I like it.

I agree Ponz isn't a troll... maybe that's why I have a soft spot for him?

In any case, I'll continue to be an infrequent contributor to this topic.

You'll need to provide some actual numbers, then, in order to properly clarify your position. How much of an improvement do you expect self-trained engines to provide, in terms of approximate Elo rating differences? How confident are you, in terms of estimated probabilities, that such an improvement will be irrefutably demonstrated within the next 5 years? Without these details, your argument becomes nothing more than an exercise in bet-hedging; it sounds like you're just trying to ensure your prediction will be right regardless of what ends up happening in the near future.

As for the rest of your post, the majority of it consists of unsupported and purely deflective arguments in which no attempt is made to provide legitimate evidence. Because of this, I have no reason to respond to all of it.

For your own credibility's sake, I'd strongly recommend that you avoid further references to the AlphaZero-Stockfish match in your arguments.

Worth watching: https://www.youtube.com/watch?v=ZGypfNUXM2U

This is as close as you came to supporting any of your arguments in post #7980. There are still two problems, though:

• Given the amount of publicity generated by DeepMind's published results, there were obviously compelling business reasons to develop and market products with features similar to those of AlphaZero.
• .Assuming a new technique will probably yield marginal benefits is quite different from assuming it will quickly lead to 4000-rated engines and spark a radical paradigm shift in modern chess theory in which all previous assessments (and all previous methods of assessment) are overturned.

To adequately support your position, estimates of expected Elo rating increases must be cited from a reliable source, such as direct quotes from Stockfish and/or Komodo developers.

No. I was pointing out that ML-based systems are programmed entirely by human developers, which means you're ultimately comparing human-derived valuations with alternative human-derived valuations, and arguing that the former (or the latter ) is fundamentally inferior because... well, humans are categorically inferior to machines. 

You're claiming that neither the practical tendency nor the theoretical tendency of a game has any bearing whatsoever on its likely game-theoretic value. You support this assertion by pointing out that any game with enduring popularity would necessarily have the same drawish tendencies -- except... hey, wait a second...

Seriously, just stop. At this point, we're no longer having a discussion involving objective facts. I've been following this thread for months (despite not posting anything until recently) and you were clearly among the 2 or 3 most intelligent people posting here. But now that I've dared to challenge a few of your lazier, more laxly researched comments with better research, your latest response has devolved into hasty deflections and horrifically flawed arguments like the one quoted directly above. You'd do better to put your ego aside and just admit there is probably no forced win from the starting position (requiring countless sound defenses to be refuted), and that self-trained engines probably won't be notching 100% scores versus elite correspondence players anytime soon.

Perhaps your perceptions have altered less because I have actually changed how I discuss/support things and more because now you are not on the sidelines and find yourself on the receiving end of someone that is not as duly impressed by you as you felt I would/should be?  You are looking for more rigor from my arguments, but I don't see that this warrants any more rigor as things stand.

Thanks for "you were clearly among the 2 or 3 most intelligent people posting here".  As for being "lazy", maybe you should try holding the line with the Axis of Error for a year or two and see how you feel after trying to support reality for the Nth time.

And, how exactly would self-trained engines ever score 100% against elite correspondence players?  The ICCF didn't/doesn't allow engine use because it makes correspondence chess better, that's a press release for the neophytes to lap up.  They were forced to allow engine use because it was and still is useless to try and stop engine use in this format.  The same will be true for self-trained engines.  So, you won't get pure centaur vs. self trained engine matchups, regardless.  The self-trained engines will be playing themselves, ultimately.

Wooooow btickler tries to change the world trying to fight fake news wooow impressive! clap clap clap haaaaaa how about stop spreading BS yourself. You should wipe your mouth because it is brown from the sheet you're talking about.

Correct. It’s nothing more than a bald assertion. IF a perfect game was ever played nobody can possibly know it happened.

Fixed that for you. If you can’t identify every mistake in a game, you can’t identify if it is perfect.

Smositional here is a perfect example of the kind of lowbrow malice I am talking about.  

“We” in this context is shorthand for mankind. When I say “we” I’m including everyone, including you. That you don’t like it doesn’t make it a mistake.

😮🎉