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

Sort:
ZouDynasty

Would two sides with equal accuracy draw a game every time?

pfren
ZouDynasty έγραψε:

Would two sides with equal accuracy draw a game every time?

 

No, this is mathematically wrong.

Ziryab
ponz111 wrote:

In over-the-board chess at levels below GM there are many wins as players make mistakes. 

 

 

 

 Those are bad stats. A typical game might have a dozen errors by each player, but only one win. So the number of wins is vastly less than the number of errors.

Even the draws usually proceed from a battery of errors.

ponz111

Ziryab  I said nothing to indicate there are as many wins at there are errors. Read my statement again?

Of course in the average game there are often many errors and only 1 win or draw.

JimDiesel22

I ran through some data bases with human and non-human play... very interesting. It's almost as though correspondence chess follows human bias to draw... and objective agents don't...

ponz111

Jim  What is an "objective agent"?  Players in general do not want to draw--they want to win!!!!!

When I played correspondence chess I played to win every game. 

There are problems with your chart.  Correspondence with engine can be a very poor game at the lower levels.  What do the numbers on the bottom mean?  You do not explain your numbers.  You do not state the strength of the players or the chess engines?  How do you know the green correspondence is with or without chess engines?  You would need to answer many questions for your presentation to make much sense?

JimDiesel22
ponz111 wrote:

Jim  What is an "objective agent"?  Players in general do not want to draw--they want to win!!!!!

When I played correspondence chess I played to win every game. 

There are problems with your chart.  Correspondence with engine can be a very poor game at the lower levels.  What do the numbers on the bottom mean?  You do not explain your numbers.  You do not state the strength of the players or the chess engines?  How do you know the green correspondence is with or without chess engines?  You would need to answer many questions for your presentation to make much sense?

"Correspondence with engine can be a very poor game at the lower levels." Oh, well this the highest level play on Lechenicher SchachServer. It's pretty self explanatory, I didn't feel like labeling the axis but it's rating vs draw rate.

Henry-Wood

What is your opinion on the boom gate?

ponz111

Jim  not so explanatory at all  you ducked my questions

ZouDynasty
JimDiesel22 wrote:

I ran through some data bases with human and non-human play... very interesting. It's almost as though correspondence chess follows human bias to draw... and objective agents don't...

U made this graph yourself???
I doubt it

JimDiesel22
ZouDynasty wrote:
JimDiesel22 wrote:

I ran through some data bases with human and non-human play... very interesting. It's almost as though correspondence chess follows human bias to draw... and objective agents don't...

U made this graph yourself???
I doubt it

I can provide references for the data. I used python with matplotlib. Took like 2-3 hours. I can answer any question about it.

ponz111

first question what was the strength of the players correspondence and engines and correspondence with engines and over the board???  so 4 questions?

JimDiesel22

Oh, so a stupid question that I already answered. Then FIDE rankings are on the bottom and were from Lechenicher SchachServer for the correspondence games, CCRL for the engine games, and actually the "human" games are CCLA servers, but I don't think anyone cares about otb.

ponz111

JimDiesel  If I understand your posting it shows that with increasing strength of players and chess engines--there are more and more draws.  That is okay but something we already knew and is an indication that chess is a draw with best play.

A lot of people care about over-the-board games and play--I would say the vast majority of chess players care about otb.

What is an "objective agent"??

 

RookeryOfCows

There is a bit of a paradox in the structure of the problem -- a win or loss is assumed to indicate a mistake was made. The question is better phrased as "Is there a line of play which will always results in a win no matter how the opposing player responds."  

It is a serious question all chess players eventually consider and mathematicians contemplate, and programmers seek as a goal.

I don't think we a capable of fully addressing this question, but machines seem to be the bet way it can be addressed. To actually mathematically prove such a solution had been found sounds difficult.

I will add that if I had one question I could ask the character Date (or Spock) from Star Trek it would be whether chess is purely deterministic with a flawless line of play. 

The same question can perhaps be asked of the game Go.

JimDiesel22

[1] - FICS standard time controls - https://www.ficsgames.org/download.html

[2] - Lechenicher SchachServer - https://www.chess-server.net/en/

[3] - Engine (40 moves in 15 minutes on an Intel i7-4770k) - https://www.computerchess.org.uk/ccrl/4040/

It is clear that human games follow a very specific trend toward 100% draws that engines don't follow due to human biases, whatever they may be. The only way to address the solution to a perfect information game is with agents that don't succumb to biases. I would lean to a draw based on this data, but it's really interesting how slowly engines are getting there. Even the best engines win a very significant amount like 15% in stockfish 10 vs alphazero (https://www.chess.com/news/view/updated-alphazero-crushes-stockfish-in-new-1-000-game-match). That being said, ponz111 has never given a logical reason for why chess is a draw, and I'd debate him on twitch.

ponz111

JiomDiesel I have given many reasons why chess is a draw.  Just read the postings. 

And you have been debating me and I have answered all. 

It is not clear at all that human games follow a very specific trend towards 100% draws AND engines don't.  Yes, at the highest levels human games are trending towards draws but they are not nearly there yet. [just look at the games by Magnus and other great players]  But engines even more so are trending towards 100% draws. But in the strongest chess which is the highest levels of correspondence chess--they are already there.!!  This is very obvious from posts already given here. There are quite a few correspondence players who have not lost a game in a bunch of years.  You are ignoring the facts already presented to you.

You are also quite wrong to think that Stockfish 10 is one of the best engines. This is why Alpha Zero got somewhat the best of Stockfish 10. 

But as mentioned--current correspondence chess at the highest levels is the strongest form of chess and this results in no errors. 

Also I do not understand how you get the absolutely crazy idea that and I quote "Even the best engines draw a very significant amount like 15% in stockfish 10 vs alphazero."  15% only?? really??

lfPatriotGames
ponz111 wrote:

JiomDiesel I have given many reasons why chess is a draw.  Just read the postings. 

And you have been debating me and I have answered all. 

It is not clear at all that human games follow a very specific trend towards 100% draws AND engines don't.  Yes, at the highest levels human games are trending towards draws but they are not nearly there yet. [just look at the games by Magnus and other great players]  But engines even more so are trending towards 100% draws. But in the strongest chess which is the highest levels of correspondence chess--they are already there.!!  This is very obvious from posts already given here. There are quite a few correspondence players who have not lost a game in a bunch of years.  You are ignoring the facts already presented to you.

You are also quite wrong to think that Stockfish 10 is one of the best engines. This is why Alpha Zero got somewhat the best of Stockfish 10. 

But as mentioned--current correspondence chess at the highest levels is the strongest form of chess and this results in no errors. 

Also I do not understand how you get the absolutely crazy idea that and I quote "Even the best engines draw a very significant amount like 15% in stockfish 10 vs alphazero."  15% only?? really??

Most people here have cut you a lot of slack because you are getting up in years, and often say things that dont make much sense. If you are going to say "and I quote" then quote it. Don't say "and I quote" then say something that the person you are quoting did not say. I know it's a simple mistake, but if you are going to be that specific, then BE that specific. 

ponz111

PATRIOT  Well one thing--I can admit when I made a mistake. You are correct that I got a word wrong when I was quoting JimDiesel and this completely changed what I thought was his quote and I apologize for my mistake.

Now hopefully all the people who have misquoted me will apologize!?

ponz111

JimDiesel  I have a question out of curiosity?  In your statement that starts "Even the best engines--was the next word you originally posted "draw" [and maybe you  changed it later??] Or was that next word "win"???