I spotted an example of llama being a tad reckless. Understandable, though. He will be back!
Chess will never be solved, here's why
Sort:
@12240
"you can't prune any legal moves" ++ I sure can. I prune 1 e4 e5 2 Ba6? as it loses.
I prune 1 a4 as it logically cannot be better than 1 e4.
"the evaluation of every such move is unreliable"
++ The provisional, heuristic engine evaluations like +0.33 are unreliable.
That 1 e4 e5 2 Ba6? loses is reliable. That 1 a4 cannot be better than 1 e4 is reliable.
"we really mean evaluations based on analysis to a substantial depth" ++ Provisional heuristic evaluation +0.33 at depth 25 is more reliable than provisional heuristic evaluation +0.33 at depth 15, but still unreliable. However, calculation until the 7-men endgame table base or a prior 3-fold repetition or an otherwise known drawn endgame is reliable.
The 109 ICCF WC Finals games link the initial position to a known reliable draw by average 39 moves. That is reliable.
"that take a significant amount of computing time" ++ Average 5 days / move.
"For pruning moves all you have is the basic evaluation with no analysis"
++ Also logic and game knowledge.
Elroch wrote:
I spotted an example of llama being a tad reckless. Understandable, though. He will be back!
Of course. Intelligent guy.
tygxc wrote:
@12240
"you can't prune any legal moves" ++ I sure can. I prune 1 e4 e5 2 Ba6? as it loses.
I prune 1 a4 as it logically cannot be better than 1 e4.
"the evaluation of every such move is unreliable"
++ The provisional, heuristic engine evaluations like +0.33 are unreliable.
That 1 e4 e5 2 Ba6? loses is reliable. That 1 a4 cannot be better than 1 e4 is reliable.
"we really mean evaluations based on analysis to a substantial depth" ++ Provisional heuristic evaluation +0.33 at depth 25 is more reliable than provisional heuristic evaluation +0.33 at depth 15, but still unreliable. However, calculation until the 7-men endgame table base or a prior 3-fold repetition or an otherwise known drawn endgame is reliable.
The 109 ICCF WC Finals games link the initial position to a known reliable draw by average 39 moves. That is reliable.
"that take a significant amount of computing time" ++ Average 5 days / move.
"For pruning moves all you have is the basic evaluation with no analysis"
++ Also logic and game knowledge.
Given that you can prune lines, and I agree with that, where does it stop? We can't prune everything that probably loses.
Jun 16, 2024
0
#12006
Elroch wrote:
You don't know the difference between logic and heuristics. Botvinnik was neither a scientist nor a logician. He was a chess player (and also an electrical engineer, interestingly)
tygxc perhaps bases his arguments on his chess training.
Jun 16, 2024
0
#12007
Elroch wrote:
llama_l wrote:
Interestingly, Botvinnik never won a world championship match as the world champion. He only ever lost or drew.
An electrical engineer who went to school in (I'm assuming) the 1930s... that would be interesting. Post Maxwell, but pre transistors, and the math they could handle would have been very limited.
I have a lot of respect for people who tackled Newton, and Laplace, and Fourier, and abstract ideas like electrical fields... all of this they had to understand without modern technology. Visualizing such ideas without the help of modern diagrams and videos. It's pretty impressive.
Living back then would have been fun. There were enormous leaps in the understanding of physics, and also pre-chess-engine times meant adjournments and correspondence chess were alive. I think I would have enjoyed it.
Yes, 'electrical' meant literally electrical. Power grids, substations, three phase supplies, that sort of stuff. That being said, it is worth remarking that Botvinnik was very interested in the possibility of chess computers at a time when they were scarcely feasible. Turing and Champernowne wrote the first computer chess program in 1948, the year that Botvinnik became world chess champion. A few years later Botvinnik got interested in them. But he was also interested in the possible application of AI for the benefit of the Soviet economy! Way ahead of the curve (and all practical technology in both cases - even Turing's early program was too complex to run on any computer of the era. And AI has taken a while to meet the hype).
Turing. Alan Turing. Again.
Broke the german code in World War II apparently.
But died young from cyanide poisoning at age 41.
It was apparently never determined whether it was suicide or murder or other.
Reminds me of the death of Alekhine.
"Turing has an extensive legacy with statues and many things named after him, including an annual award for computer science innovations. He appears on the current Bank of England £50 note, which was released on 23 June 2021 to coincide with his birthday. A 2019 BBC series, as voted by the audience, named him the greatest person of the 20th century."
Heuristics are based on logic but don't consist of purely deductive logic. Hence heuristical evaluations aren't reliable. The aim may be to produce, say, a strong engine by means of heuristics that have a strong statistical accuracy rather than any absolute accuracy. In other words, heuristics that generally perform well.
Regarding Turing, his contribution, although very great, has been romanticised. There were other great contributors who are forgotten.
Jun 16, 2024
0
#12010
I don't know the ones who have been forgotten, but Goedel, von Neumann and co. are well remembered.
Jun 16, 2024
0
#12011
tygxc wrote:
@12240
"you can't prune any legal moves"
++ I sure can. I prune 1 e4 e5 2 Ba6? as it loses.
I stand corrected. What I meant was that a competent person cannot prune any legal moves for the opponent of a strategy (all need to be resolved).
I prune 1 a4 as it logically cannot be better than 1 e4.
Similarly, the likes of Schaeffer cannot prune such moves, because they know what a weak solution is. What you do is, frankly, irrelevant.
"the evaluation of every such move is unreliable"
++ The provisional, heuristic engine evaluations like +0.33 are unreliable.
That 1 e4 e5 2 Ba6? loses is reliable. That 1 a4 cannot be better than 1 e4 is reliable.
You really think proclamation is a valid proof technique?
"we really mean evaluations based on analysis to a substantial depth" ++ Provisional heuristic evaluation +0.33 at depth 25 is more reliable than provisional heuristic evaluation +0.33 at depth 15, but still unreliable.
Same with -3.9. Just a LOWER PROBABILITY of being disastrously wrong.
But, let's face it, you don't understand Bayesian probability and its implications.
However, calculation until the 7-men endgame table base or a prior 3-fold repetition or an otherwise known drawn endgame is reliable.
True, if ALL legal opposing play is dealt with. If not, you are just scamming.
The 109 ICCF WC Finals games link the initial position to a known reliable draw by average 39 moves. That is reliable.
Definitely not 100% reliable. It shows a complete lack of understanding to jump to this conclusion. You even said it was likely that it was very likely that some later games between the players would be decisive!
"For pruning moves all you have is the basic evaluation with no analysis"
++ Also logic and game knowledge.
No, YOU don't have logic. You have the word, but no idea of what it means.
And you have no understanding of the difference between deduced facts and UNCERTAIN inductive knowledge.
The latter DEFINITELY includes your beliefs about 1. e4 e5 2. Ba6
And, to be frank, you are an annoying, arrogant, ignorant person. [Ignorant means you don't know things that matter].
Bill Tutte is one of the unsung heroes. Also, the Tunny codebreakers in general.
Incidentally, I do think that we can know things due to inductive reasoning. After all, all the axioms and premises that deductive reasoning rests upon cannot come from deductive reasoning. That would imply an ideal system which may not relate to reality. There've been things written about that kind of idea or situation, which makes certain interpretations of "truth" far less certain by relying on coherence theories of truth.
Can't we perhaps agree that tygxc is to some extent both irritating and slow to respond to well-based challenges, which makes his input seem as though he isn't actually responding to anything? I still think his reliance on games played is at least a start, although I cannot understand why they do not focus on different types of game, such as 200 move games, rather than on the formula which seems to lead to draws all the time, when games typically last for 50 or 60 moves. It isn't a proper testing methodology and they ought to realise that.
Paul Thagard is the author of the first of two recent new arguments against the coherence theory. Thagard states his argument as follows:
if there is a world independent of representations of it, as historical evidence suggests, then the aim of representation should be to describe the world, not just to relate to other representations. My argument does not refute the coherence theory, but shows that it implausibly gives minds too large a place in constituting truth. (Thagard 2007: 29–30)>>>>
I came up with that in about 1993 as part of my philosophy degree. If I came up with it, the strong likelihood is that others presented it before I did, since it's obvious.
Optimissed wrote:
Paul Thagard is the author of the first of two recent new arguments against the coherence theory. Thagard states his argument as follows:
if there is a world independent of representations of it, as historical evidence suggests, then the aim of representation should be to describe the world, not just to relate to other representations. My argument does not refute the coherence theory, but shows that it implausibly gives minds too large a place in constituting truth. (Thagard 2007: 29–30)>>>>
I came up with that in about 1993 as part of my philosophy degree. If I came up with it, the strong likelihood is that others presented it before I did, since it's obvious.
Long quash is unmuted
u can setup a 4x8 board. no castling (no en passant ? / no promo ?). ea side gets 4 pawns, a king a knight a bishop & a rook (bishop & rook makes up a kmart queen). tweak a antminer or two & see what u get. once u get from 16 to 7 u can bridge to a 4x8 syzygy base (which would probably solve better than 7 on a half-board...and i bleeve its been done will check).
I feel this could be a good starting point 4da attemptor yes ?
hi opti !!...happy ur back luv 
L♥
(all need to be resolved).
nah worry about that later. u can build in a parsa a all sorts a dum moves. openly dropping a piece w/out compensation just makes a joke outta stuff. i say prune baby prune !
Jun 16, 2024
0
#12019
""that take a significant amount of computing time" ++ Average 5 days / move."
your calculation listed it at over a million positions per second, so you are contradicting yourself.
@12261
"we can know things due to inductive reasoning" ++ All we know is by inductive reasoning
"That would imply an ideal system which may not relate to reality" ++ Like Cantor
"tygxc is to some extent both irritating and slow" ++ Huh. Do I insult people?
I might be the fastest responder on this thread. I respond to trolls more often than I should.
"well-based challenges" ++ What well-based challenges?
"he isn't actually responding to anything?" ++ To that which makes sense.
I decide for myself what posts and what parts of them I respond to. I have no obligation.
"why they do not focus on different types of game, such as 200 move games"
++ The games end in draws in average 39 moves. What would you want them to do?
"the formula which seems to lead to draws all the time" ++ Chess is a draw, the formula cannot change that. There have been proposals to shorten the thinking time now 5 days/move.
That would mean a moderator taking something out of context. That's the problem.