True or False Chess is a Draw with Best Play from Both Sides

Sort:
MARattigan
Alika05 wrote:
What about any game between two top level computers (e.g, Stockfish, Komodo,...) where one side gradually grinds the other into a winning position? We can’t really say that the losing side made a “mistake” if it’s an engine with a 3000+ rating, right? Engines are designed not to make even the slightest of inaccuracies, but they also need some variation, because otherwise every game played among engines would be the exact same.
Maybe I’m overthinking this, but to me this is a very interesting concept...

Engines don't play with perfect accuracy unless they're attached to an EGTB and the material on the board is covered by the EGTB. (The EGTB in this case must take account of the 50 move rule if it's in force.) The fact that you get different results in a series of games between engines necessarily means they make mistakes.

EndgameEnthusiast2357
Alika05 wrote:
What about any game between two top level computers (e.g, Stockfish, Komodo,...) where one side gradually grinds the other into a winning position? We can’t really say that the losing side made a “mistake” if it’s an engine with a 3000+ rating, right? Engines are designed not to make even the slightest of inaccuracies, but they also need some variation, because otherwise every game played among engines would be the exact same.
Maybe I’m overthinking this, but to me this is a very interesting concept...

Is the difference between a 4000 rating and 3000 rating even significant? ...5000 and 6000...?

Prometheus_Fuschs
EndgameStudier escribió:
Alika05 wrote:
What about any game between two top level computers (e.g, Stockfish, Komodo,...) where one side gradually grinds the other into a winning position? We can’t really say that the losing side made a “mistake” if it’s an engine with a 3000+ rating, right? Engines are designed not to make even the slightest of inaccuracies, but they also need some variation, because otherwise every game played among engines would be the exact same.
Maybe I’m overthinking this, but to me this is a very interesting concept...

Is the difference between a 4000 rating and 3000 rating even significant? ...5000 and 6000...?

Yes, you'd hardly notice it just as I can't notice much difference between a good expert and a world champ because they'd both maul me OTB.

MARattigan
Prometheus_Fuschs wrote:
EndgameStudier escribió:
Alika05 wrote:
What about any game between two top level computers (e.g, Stockfish, Komodo,...) where one side gradually grinds the other into a winning position? We can’t really say that the losing side made a “mistake” if it’s an engine with a 3000+ rating, right? Engines are designed not to make even the slightest of inaccuracies, but they also need some variation, because otherwise every game played among engines would be the exact same.
Maybe I’m overthinking this, but to me this is a very interesting concept...

Is the difference between a 4000 rating and 3000 rating even significant? ...5000 and 6000...?

Yes, you'd hardly notice it just as I can't notice much difference between a good expert and a world champ because they'd both maul me OTB.

I imagine @EndgameStudier meant significant compared with perfect play. E.g. Carlsen would probably find the difference between players with ratings of 1000 and 1500 insignificant. (The players would not.) Similarly a perfect player might find the difference between a 3000 and 4000 rating insignificant.

ponz111

It is an observation that quality of play improves with the higher the rating. 

I do not agree that players rated 1000 and 1500 would not find the rating differences significant.

Actually it is possible to show that a losing engine with a 3000 rating made a mistake.

There are some chess problems that the best chess computers cannot solve but a human can solve quickly.

 

lfPatriotGames
ponz111 wrote:

It is an observation that quality of play improves with the higher the rating. 

I do not agree that players rated 1000 and 1500 would not find the rating differences significant.

Actually it is possible to show that a losing engine with a 3000 rating made a mistake.

There are some chess problems that the best chess computers cannot solve but a human can solve quickly.

 

That's probably one of the best reasons yet why there is a question of whether or not chess is a draw with best play.

ponz111

Patriot the fact that computers are not yet perfect has little bearing as to the main question of this forum.

But of course I gave 4 statements and am not sure which one you meant?

There is no question from the top players that chess is a draw with optimum or correct play.

lfPatriotGames
ponz111 wrote:

Patriot the fact that computers are not yet perfect has little bearing as to the main question of this forum.

But of course I gave 4 statements and am not sure which one you meant?

There is no question from the top players that chess is a draw with optimum or correct play.

Of course there is a question. Not all top players believe as you do. But the fact people, and especially computers are still improving leaves a lot of room for doubt when taking an absolute stand. 

Prometheus_Fuschs
ponz111 escribió:

It is an observation that quality of play improves with the higher the rating. 

I do not agree that players rated 1000 and 1500 would not find the rating differences significant.

Actually it is possible to show that a losing engine with a 3000 rating made a mistake.

There are some chess problems that the best chess computers cannot solve but a human can solve quickly.

 

That is indeed true, however, the amount of positions on which they can evaluate better than us is vastly larger than the amount of positions on which we can evaluate better than them.

Prometheus_Fuschs
MARattigan escribió:
Prometheus_Fuschs wrote:
EndgameStudier escribió:
Alika05 wrote:
What about any game between two top level computers (e.g, Stockfish, Komodo,...) where one side gradually grinds the other into a winning position? We can’t really say that the losing side made a “mistake” if it’s an engine with a 3000+ rating, right? Engines are designed not to make even the slightest of inaccuracies, but they also need some variation, because otherwise every game played among engines would be the exact same.
Maybe I’m overthinking this, but to me this is a very interesting concept...

Is the difference between a 4000 rating and 3000 rating even significant? ...5000 and 6000...?

Yes, you'd hardly notice it just as I can't notice much difference between a good expert and a world champ because they'd both maul me OTB.

I imagine @EndgameStudier meant significant compared with perfect play. E.g. Carlsen would probably find the difference between players with ratings of 1000 and 1500 insignificant. (The players would not.) Similarly a perfect player might find the difference between a 3000 and 4000 rating insignificant.

By significant you mean noticiable in the quality of play?

Billkingplayschess

AlphaZero says white wins...

DiogenesDue
ponz111 wrote:

There is no question from the top players that chess is a draw with optimum or correct play.

You already lost this line of argument on the "Will computers ever solve chess?" thread.  Please speak factually.  If you have a verified poll of super-GMs in your back pocket, please produce it.  Til then, your claim of what top players believe across the board is meaningless.

MARattigan
ponz111 wrote:

It is an observation that quality of play improves with the higher the rating. 

Undoubtedly.

I do not agree that players rated 1000 and 1500 would not find the rating differences significant.

Neither do I, that's what I said.

Actually it is possible to show that a losing engine with a 3000 rating made a mistake.

SF8 (sans EGTB) can't play KNNKP. I've studied the endgame and can see its mistakes while I'm playing it.

Neither SF8 nor I (sans EGTB) can play KBNKPR but Lomonosov can see its mistakes instantly. So -agreed.

I think it's almost certain that the winning engine also did.

On the other hand is it generally possible in a game from the starting position before the game reaches a position covered by an EGTB? (I'm sure it would be possible in some positions, but doubt it would be in the main.) Looking at the evaluation from the winning engine is no guarantee. Do you have examples?

There are some chess problems that the best chess computers cannot solve but a human can solve quickly.

Oodles. E.g. this one https://www.chess.com/forum/view/more-puzzles/white-to-play-and-mate-in-24 and this nice Otto Blathy one posted by @9135I in a different thread.

White to play and win

They haven't solved this one, which is the subject of the thread, but neither have humans. In this case the machines have a strong tendency to do better than the humans in actual play.

All of which gets us no closer to an answer to OP's question.
 

 

jazzjune

Chess is armwrestling the Hindu Kali

Prometheus_Fuschs
Optimissed escribió:

None of your comments are to the point, MAR. You seem to be obsessed with the idea that no-one plays perfectly or that mistakes will always happen, without realising that it's these mistakes which cause games to be won and lost.

Are you trolling?

The odds are stacked against you on that one.

MARattigan
Excalibr4 wrote:

AlphaZero says white wins...

My big brother says Black wins.

Jeppesen960
One would think that with the absurd skill and development of chess engines, which play eachother and end plenty games decisively, would show that while best play comes from each ‘player,’ it is not always a draw.
lfPatriotGames
Optimissed wrote:
lfPatriotGames wrote:
ponz111 wrote:

It is an observation that quality of play improves with the higher the rating. 

I do not agree that players rated 1000 and 1500 would not find the rating differences significant.

Actually it is possible to show that a losing engine with a 3000 rating made a mistake.

There are some chess problems that the best chess computers cannot solve but a human can solve quickly.

 

That's probably one of the best reasons yet why there is a question of whether or not chess is a draw with best play.>>

There isn't any question about it. Here we have a forum of chess players. What's the common factor with chess players? It's that they are trained to consider specific variations, concrete variations. Time and again I've been told off by higher rated players for daring to use general principles in looking at a game situation. It is the advent of computers that has got people thinking in this inefficient way, stuck in a mindset where only concrete variations count.

The problem is as though we're trying to work out if there's any commonality between dropping a massive object and watching it fall towards the most massive centre of mass ... the Earth. This is why it actually took people with brains to work out the effects of gravity. People who can see common factors and identify and isolate them. Same with any scientific experimentation. Yet here, so many people are stuck in the mindset of thinking they have to look at every possible chess variation, to see if there's one that is not a draw with best play by both sides. They've just never learned to use their minds as they are capable of being used. There's a reason why the strongest players have no doubt about this issue.

 

If there isn't any question about it, why is it questioned? I agree many players, strong players, have no doubt about the issue. But they dont know. They believe, they may not have any doubts, they have faith, but they dont know. I just think the abilities of people and computers isn't nearly advanced enough to say for certain yet. 

I understand what you are saying about gravity. But with gravity there are no exceptions. Even weird things like black holes or something can be explained. In chess there are exceptions to the belief that its a draw. Sometimes one side wins. Even at the very, very highest level, one side sometimes wins.

najdorf96

Indeed. It's like grandpa's old joke he keeps telling to us, then of course to the next gen. It sucks us in. It's soo good I find myself telling that very same joke to kids but indubitably "adding" in my own nuances & set up. You know how it is. I've fought the good fight with ponz: punched & counterpunched with him in this very thread for many years on & off (going on 8ish years now I guess heh) and yet he remains staunched, unyielding. Like grandpa's ol' joke, man. Gotta respect that. I rather like his other proclamation; that human players will undoubtedly join (collaborate) with engines, like we people call, "centauries" and play World Championship tournaments in the future (formats like this has been envisioned as currently it's been going on for a number of years however it's popularity hasn't been realized YET methinks). I can dig that stuff. "Battlebots" for Chess! Yeah, man. Anyways back to the topic. Ponz has often been quoted (paraphrasing) saying that engines will eventually show that Chess is a draw by virtue of them, themselves drawing every game between them. Now, in 2019 we can safely say that, his prediction has not come to fruition YET. Throw in AlphaZero in the mix, and it somehow becomes more opaque rather than clearer. He's also referred to future WC games, GM games (paraphrasing again obviously) too, will become more and more drawish. Again, after almost a decade from the time he started this forum til now, it hasn't been the case for the most part. For me leastways, anyway🤔. Maybe Carlson & the Gang didn't get the memo. Like how when the Great Capablanca said this very same idea many many moons ago. Guess it got lost in the trees. The proverbial dog ate it. Because the late 60's, 70's, 80's, 90's produced fine players, eventual Hall of Fame players like ponz111 and cascaded into the new millennium producing chess players of all ranging walks of life & diversity. Each infusing new vigor, overturning years and years of old thoughts, concepts, perceptions and bringing in year after year younger n younger GMs. No Got Mail, perhaps? The one that says, "Chess is a Draw"? As long as Humans play this game, making different formats to update the old; each ascending generation playing out innumerable variations, lines thought dead (20-30 moves deep, 30-40 ?) or capable of going farther out. Indeed, what of the advent of AI, an autonomous neural net per se, as opposed to simply chess playing engines? Who can honestly say in totality if Chess is a Draw in the end? Is there an End to our deep, blue ocean's mysteries? We have some proof of Galaxies; theories and such, but we really don't know. They say Chess complexity is almost infinitesimal as the Universe as how we know it. AFTER ALLL THAT? WHAT DO I THINK? I believe White wins. Heh. Can anyone disprove that? Let's play😎

najdorf96

(p.s.~I play White, of course of course!)