@4508
"the advantage of the first shot of the whites can hardly be missed by the blacks and gives at best a draw for the blacks."
++ The advantage of 1 tempo is not enough to win,
so chess is a draw with best play from both sides.
You cannot queen a tempo.
Chess will never be solved, here's why
Sort:
Sep 2, 2022
0
#4443
tygxc wrote:
@4505
"But there's no reason to suppose that it's a Poisson distribution."
++ There is reason, a Poisson distribution is derived from the Binomial distribution and applies to many similar stochastic processes.
Actually, we can be sure that this is a rather crude model that can be improved upon, and that the improved model would give different results (how much is unclear). For example, we could certainly use errors Poisson by move or errors Poisson by game, with the former being more natural to me. In the former case, there is surely a better model that takes into account empirical dependence of errors on game length. In the latter case there is surely a better model that takes into account the variation in error rates depending on the stage of the game (maybe errors are most common in the middle game). Both could be improved by incorporating any information on the strength of the two players etc.
Sep 2, 2022
0
#4444
tygxc a écrit :
@4508
"l'avantage du premier coup des blancs peut difficilement être raté par les noirs et donne au mieux un nul aux noirs."
++ L'avantage d'un tempo n'est pas suffisant pour gagner,
donc les échecs sont un match nul avec le meilleur jeu des deux côtés.
Vous ne pouvez pas ajuster un tempo.
Yes, of course, but this one-shot advantage is decisive for machine-to-machine matches, which is why blacks do best with a draw. No?
@4511
"this is a rather crude model"
++ Yes, I first calculated with an even cruder model with just high school math.
The result is about the same.
A more refined model may yield slightly different results, but no drastic changes.
"errors Poisson by move or errors Poisson by game"
++ Yes, you could use Binomial Distribution by move, but it would not make much difference.
"the variation in error rates depending on the stage of the game"
++ Yes, as chess is most complicated around 26 men, it is plausible that most errors occur around 26 men. It would not make much difference in the result: chess being a draw and the number of games with 0 errors.
"information on the strength of the two players"
++ No, that is not relevant. I take a sufficiently large tournament with a sufficient number of players of sufficient quality and then apply statistics to that. The last ICCF WC is suitable as well as Zürich 1953. From the statistics follows that only chess is a draw is consistent with the observed data and follows the number of games with no errors: 127 for the ICCF WC and 74 for Zürich 1953.
@4512
"this one-shot advantage is decisive for machine-to-machine matches"
++ No, the first move advantage of 1 tempo is not decisive in machine vs. machine matches.
On the contrary: in the TCEC top chess engine competition they had to impose slightly unbalanced openings to avoid all draws.
The more time you give the engines or the humans, the more they draw.
Sep 2, 2022
0
#4447
Optimissed wrote:
I'll try to explain it. You're a statistician and therefore accustomed to dealing with sets of events in the "macro". Considering a unique event, you will tend to see its outcome "a" or outcome "b" probabalistically, since as a unique event but still an event which is viewed in the context of the set, superficially outcomes a or b follow probability patterns. Within the context of the set, viewing all the events as a whole, each unique event is simply part of this or that subset as part of a statistical pattern. That tells us nothing about the event itself.
Bayesian probability is about the state of belief in a proposition that is uncertain (cannot be deduced by Boolean logic from the known information). It is well-known to be fully suited to dealing with samples of one, using all the inductive power available and no more.
An outcome on the chess board may be considered to have a definite result with best play: but it is one that may be unknown.
Correct. That is a situation where one wishes to quantify belief, and the only consistent way to quantify belief is Bayesian probability.
All probabilities that are assigned to it are therefore the result of guesswork based on inductive reasoning.
This is an inconsistent sentence.
The prior is the only thing that could be described as "guesswork" (though you would not use this word if you were familiar with the work in the subject, which often gives an answer to "what is the appropriate prior?". And of course, inductive reasoning has no element of guesswork. Fundamentally it is the correct application of Bayes rule (difficult as this may be to formalise in a specific example).
That is effectively equivalent to basing them on error limits.
This is an incorrect comparison to something you are familiar with from frequentist statistics.
Anyway, that's my thinking.
Sep 2, 2022
0
#4448
tygxc wrote:
"information on the strength of the two players"
++ No, that is not relevant. I take a sufficiently large tournament with a sufficient number of players of sufficient quality and then apply statistics to that. The last ICCF WC is suitable as well as Zürich 1953. From the statistics follows that only chess is a draw is consistent with the observed data and follows the number of games with no errors: 127 for the ICCF WC and 74 for Zürich 1953.
It is relevant because it is available information that definitely affects the rate of errors. Weaker players make more errors. Players playing against stronger players make more errors (because stronger players are so because they provide more opportunities for the opponent to make errors). The only question is how relevant it is. Undoubtedly, like most things, you would be keen to make an absolute proclamation about this without any quantitative reasoning.
Elroch wrote:
Optimissed wrote:
I'll try to explain it. You're a statistician and therefore accustomed to dealing with sets of events in the "macro". Considering a unique event, you will tend to see its outcome "a" or outcome "b" probabalistically, since as a unique event but still an event which is viewed in the context of the set, superficially outcomes a or b follow probability patterns. Within the context of the set, viewing all the events as a whole, each unique event is simply part of this or that subset as part of a statistical pattern. That tells us nothing about the event itself.
Bayesian probability is about the state of belief in a proposition that is uncertain (cannot be deduced by Boolean logic from the known information). It is well-known to be fully suited to dealing with samples of one, using all the inductive power available and no more.
An outcome on the chess board may be considered to have a definite result with best play: but it is one that may be unknown.
Correct. That is a situation where one wishes to quantify belief, and the only consistent way to quantify belief is Bayesian probability.
All probabilities that are assigned to it are therefore the result of guesswork based on inductive reasoning.
This is an inconsistent sentence.
The fact that it's inconsistent with your own thoughts on the matter is no proper argument against.
Sep 2, 2022
0
#4450
Thank you for all your explanations.
I am also paying attention to the other discussion where I learn lots of interesting things
Have a good evening
Nicoquelicots wrote:
Thank you for all your explanations.
I am also paying attention to the other discussion where I learn lots of interesting things
Have a good evening
Although we may not have quite hit it off immediately, because you don't like my terrible ego, nevertheless I'm glad that you are enjoying these interesting things. I've had a difficult day, including a business trip, by car, with a person who was driving and shouting at all the traffic for 150 miles. It wasn't easy and so thankyou for your calmness and friendliness.
NervesofButter wrote:
Has the debate been settled as to whether chess will ever be solved?
Well, I have spoken, if that's a help.
Sep 2, 2022
0
#4453
Optimissed a écrit :
Nicoquelicots wrote:
Thank you for all your explanations.
I am also paying attention to the other discussion where I learn lots of interesting things
Have a good evening
Although we may not have quite hit it off immediately, because you don't like my terrible ego, nevertheless I'm glad that you are enjoying these interesting things. I've had a difficult day, including a business trip, by car, with a person who was driving and shouting at all the traffic for 150 miles. It wasn't easy and so thankyou for your calmness and friendliness.
Thank you. You know, egos are well shared on this page. It’s normal and it doesn’t bother me at all. The important thing is to turn the pages and that, like you, I also know quite well how to do it.
tygxc wrote:
@4502
"why you think that ICCF games can be proven to be of sufficient quality"
++ A World Championship Finals, 17 ICCF (grand)masters with engines, 50 days / 10 moves.
"why ICCF games have reached some "quality" threshold that is absolute"
++ There is no such threshold.
Take the 1953 Zürich Candidates' Tournament:
210 games = 118 draws + 49 white wins + 43 black wins
Assume chess a draw.
Fit a Poisson distribution so the probability of an odd number of errors is (49 + 43) / 210.
Result: mean value = 1.044 error / game.
Games with 0 errors: 74
Games with 1 error: 77
Games with 2 errors: 40
Games with 3 errors: 14
Games with 4 errors: 4
Games with 5 errors: 1
Now assume chess is a white or black win.
Fit a Poisson distribution so the probability of an odd number of errors is 118 / 210.
Result: impossible fit
Conclusion: chess is a draw with best play from both sides.
Except for the part where you cannot determine what is actually an "error" with certainty. You are using engine evaluations of errors to evaluate the absolute accuracy of engines.
"Hey Joe, are you best the carpenter in the world?"
"Yep."
"....and do you ever build things that aren't 100.00% straight?"
"Nope, measure everything myself."
"Ok thanks, case closed."
Surely you can see the issue...you just choose to ignore it.
Elroch wrote:
tygxc wrote:
"information on the strength of the two players"
++ No, that is not relevant. I take a sufficiently large tournament with a sufficient number of players of sufficient quality and then apply statistics to that. The last ICCF WC is suitable as well as Zürich 1953. From the statistics follows that only chess is a draw is consistent with the observed data and follows the number of games with no errors: 127 for the ICCF WC and 74 for Zürich 1953.
It is relevant because it is available information that definitely affects the rate of errors. Weaker players make more errors. Players playing against stronger players make more errors (because stronger players are so because they provide more opportunities for the opponent to make errors). The only question is how relevant it is. Undoubtedly, like most things, you would be keen to make an absolute proclamation about this without any quantitative reasoning.
If this same reasoning/method were used on a planet that was populated with slightly intelligent church mice, it would be exactly the same...
"Everybody draws now, even the best players, ergo chess must be a forced draw".
Nicoquelicots wrote:
Thank you. You know, egos are well shared on this page. It’s normal and it doesn’t bother me at all. The important thing is to turn the pages and that, like you, I also know quite well how to do it.
A false positive result, sadly. You'll find that out eventually if you disagree on other occasions.
Sep 2, 2022
0
#4457
btickler a écrit :
Nicoquelicots wrote:
Thank you. You know, egos are well shared on this page. It’s normal and it doesn’t bother me at all. The important thing is to turn the pages and that, like you, I also know quite well how to do it.
A false positive result, sadly. You'll find that out eventually if you disagree on other occasions.
You know, you can’t just look at the negative:
If I disagree I simply express it. But, as in chess, everyone can be mistaken. And in the next game, we fix to improve.
This is what interests me in Life, as in chess.
Yours truly
Nicoquelicots wrote:
You know, you can’t just look at the negative:
If I disagree I simply express it. But, as in chess, everyone can be mistaken. And in the next game, we fix to improve.
This is what interests me in Life, as in chess.
Yours truly
I do not just look at things negatively, it took almost decade of observation to reach my conclusions.
Where's my negative take on, say, some Batgirl post?
Your response might be "who's Batgirl?", which would illustrate my point. I have bumped heads with the likes of Ziryab, Elroch, various Bacon incarnations over time, etc. but that's all good in the end. It takes a consistent pattern of negative behavior to influence my thinking in that direction.
Yours truly.
Sep 2, 2022
0
#4459
Optimissed wrote:
All probabilities that are assigned to it are therefore the result of guesswork based on inductive reasoning.
This is an inconsistent sentence.
The fact that it's inconsistent with your own thoughts on the matter is no proper argument against.
No, it is inconsistent because you refer to inductive reasoning as being "guesswork".
On the contrary, inductive reasoning is the correct way to modify beliefs based on evidence.
Elroch wrote:
Optimissed wrote:
All probabilities that are assigned to it are therefore the result of guesswork based on inductive reasoning.
This is an inconsistent sentence.
The fact that it's inconsistent with your own thoughts on the matter is no proper argument against.
No, it is inconsistent because you refer to inductive reasoning as being "guesswork".
On the contrary, inductive reasoning is the correct way to modify beliefs based on evidence.
I'm afraid that your arguments are mutually contradictory.
It's for the following reasons. You often remark that inductive reasoning doesn't give a guarantee of accuracy and therefore you would not assign a probability of unity to its product.
However, guesswork isn't pure chance. Guesses can, of course, be random but regarding something like the solution of 1. e4 e5 2. Ba6, the idea that white should lose isn't the product of random guesswork but of indirect evidence. Neverthless, the fact that YOU won't assign it a probability of 1 means that in some part, it's also guesswork, to you but not to me. It cannot be anything else, unless you wish to play semantic games.
So evidence or not, it involves guesswork, because the indirect evidence isn't deductively conclusive. Another word for guesswork might be uncertainty.
u-n-c-e-r-t-a-i-n-t-y. In other words, guesswork. It may be because the way some ideas are stated is unfamiliar to you. Nevertheless, please be sure that you understand the meanings of things you delight in arguing against. I've just had a difficult, annoying and ultimately successful day, dealing with business transactions. I'm tired but when you make an obviously bad argument, as you just did, I can still find it easy to show you where you have gone wrong. I only hope you understand that and don't try to press your incorrect arguments. I'm sure you and I would be much stronger as a team, instead of you persistently battling the inevitable.
@4507
" there has never been a game with no errors"
++ They lied to you.
99.7% of ICCF WC draws are games with no errors.
Even Zürich 1953 had 74 games with no errors.