Chess will never be solved, here's why

@13311

"quantum computing is not suitable at least right now for the problem of solving chess"
++ There are two reasons why it is or will be suitable

1. A conventional computer has a move generator that generates all legal moves from a given position, i.e. all positions that can result from it by a legal move. A quantum computer can resolve the reverse query: give all positions that can lead to a given position by a legal move. That kind of query is needed for retrograde analysis to generate an 8 men endgame table base from a 7 men endgame table base. Then 7 -> 8 -> 9 -> ... -> 32.
2. A conventional computer works sequentially. A quantum computer can process in parallel: give it an array of positions and get an array of positions that can lead to it by 1 legal move. That speeds up the process.

"who knows what researchers will discover the next couple of centuries"
++ Yes: the functionallity of the computer of 80 years ago in the Manhattan Project: a room full of vacuum tubes now fits inside a smart watch. That is why I expect a quantum computer to strongly solve Chess before 2100.

As per usual, your ideas of scale are way off. I will just state the simplest of reasons why quantum computers cannot solve chess: quantum computers can't even play a game of chess nor are they currently in any imminent danger of being able to play a game of chess. I could go into the problems of decoherence, the limited nature of matrix operations that quantum computers can do, etc. but they have already been covered before, and probably here in this thread.

people who are experts in the field dont dare make century length technology predictions.

tygxc: knows nothing yet confidently reports his predictions as facts to those around him.

tygxc you still havent addressed how ive had mathematicians personally verify my rebuttals to your delusional solution methodology claims, and you still havent explained how quantum computing is supposed to overcome the hurdle of that we dont have enough atoms on the earth to store 10^44 chess positions.

also tygxc why dont you just update the wikipedia articles since you say they are wrong.

oh i forgot about that.

H

Let me emphasise one of your most serious errors of reasoning - the well-known "hasty generalisation fallacy" which is to have a small sample and conclude that anything that has not happened in that small sample will never happen. I have marked where you are explicitly or implicitly using this.

Note that while sampling the whole of a population is the only way to draw a truly general conclusion without deductive reasoning, even reaching a high level of confidence about the general truth needs a very large sample, compared to the size of the population. For example if you know that there either is or is not one white ball in an urn with a large number of black balls and you take out half of the the balls and find they are all black, this is only very weak evidence that there is not a white ball in the urn - if there had been, there would be a 50% chance that you would miss it with your sample.

I understand that you have difficulty understanding the uncertainty in chess results and the appropriateness of viewing them as random samples, but that is your failing, not a real objection. The Elo rating system is of course based on dealing with the randomness in results, whether it is casual blitz or a world championship match with dozens of consecutive draws.

[Admittedly, sometimes you do contradict yourself and refute your previous arguments, such as the times you have said that it is not only possible but likely that there will be decisive games in a future ICCF world championship]

@13322

"conclude that anything that has not happened in that small sample will never happen"
++ We have data bases with millions of grandmaster games and of thousands of ICCF correspondence games. There is also logic and deductive reasoning.
As 10^38 = 3^80 3 non transposing branches per ply generate all chess positions in 40 moves.

"it is not only possible but likely that there will be decisive games in a future ICCF world championship" ++ Because of random external factors: clerical error, player illness, computer crash, power blackout, flooding, fire, war, revolution...

tygxc you still havent provided the " logic and deductive reasoning" you claim to have lmfao.

for something to be logical or deductive in terms of a game solution it must solely be derived of the axioms of the game rules.

all of your examples of "logic and deductive reasoning" are you claiming general chess knowledge as fact.

"We have data bases with millions of grandmaster games and of thousands of ICCF correspondence games"

still HGA LMFAO.

@13323

"AI algorithms are programmed by humans" ++ AlphaZero was only programmed with only the Laws of Chess. It acquired its own chess knowledge from autoplay.

"neither humans or AI will solve chess" ++ But the combination of 17 humans and twin computers 90 million positions/s at 5 days/move is doing it it now: 112 draws out of 112 games.

@Deslising im not here to attack or defend your opinions, im just here to point out that tygxc argues from a delusional fantasy.

tygxc's core arguments have been refuted by literal mathematicians and he refuses to admit to being wrong.

according to tygxc, the evidence that perfect chess is being played is that there are draws. yes, that is verbatim tygxc's logic. we have been asking for extra info for over 2 years and tygxc has simply repeated the claiom that the draws prove perfection. of course, this is ridiculous, but logic has never stopped tygxc before.

none of whats being done in the ICCF (what tygxc says is "doing it") counts towards any solution of chess. It's not mathematically verified in any sense. It does not draw the full game tree in order to verify the moves nor does it provide any rigorous algorithmic proof, it is just a single line of play. By definition it cannot represent any proof.

tygxc fundamentally doesnt understand what mathematical proof is, so he has been making the same fallacies for several years.

@13328

"Alpha Zero is run by algorithms."
++ But it acquired its chess knowledge from the Laws of Chess only.

"claim this variations is the only one right" ++ There are several good moves for white and several good responses from black, and they all draw: See here

1 e4 e6
1 e4 c5
1 e4 e5 2 Nf3 Nf6
1 e4 e5 2 Nf3 Nc6 3 Bb5 Nf6
1 e4 e5 2 Nf3 Nc6 3 Bb5 a6

1 d4 d5 2 c4 e6
1 d4 d5 2 c4 dxc4
1 d4 d5 2 c4 c6
1 d4 Nf6 2 c4 e6
1 d4 Nf6 2 c4 g6

1 Nf3 Nf6
1 Nf3 d5

It's almost baffling how consistently wrong tygxc is at all times.

You'll notice that tygxc has no counterargument to the core arguments of what I say, (because it has literally been verified by mathematicians), and instead resorts to down voting because he wants to try to visually discredit me.

@13318

"We only need to store about 10^2 positions"
++ We have 10^2 games, i.e. about 10^4 positions, but those are the summary of 10^17 positions considered and by coincidence that is the number of positions relevant to weakly solve chess.

tygxc here is literally too stupid to recognize that he is being satirized in the comment he is "addressing", but yet he still makes many errors.

the claim of 10^17 positions being relevant to solving chess has many fallacies and miscalculations.

first off, tygxc pulls a broader set of 10^34 positions out of LITERAL THIN AIR. We have been asking for any sort of source or calculation for this for years, but all tygxc does is link a sample that explicitly contradicts what he claims.

then, tygxc takes the square root of 10^34 to get 10^17- this step is not necessarily fallacious, taking the square root is a known heuristic for representing the branching of only one side, to get the estimated size for a game tree that guarantees a result for one side. But tygxc then claims that that the solution tree itself is the only positions that are relevant, ie taken into consideration in calculations. This ignores the positions that need to be searched in order to FIND the tree, which are just as relevant as the solution tree itself in calculations.

You are talking drivel, Mega. At first you pretended to be a mathematician and it turns out you seem to be a schoolboy. To you, a mathematician is someone two years closer than you to obtaining a degree in maths. The opinions of such people do not count.

You have resorted to constant personal attacks, backed up by Elroch. Neither of you are trustworthy in your opinions, since we do not even know that what you are giving is true opinion. It's clear that Elroch has used dishonesty to try to win points, whatever they may have been. This includes switching arguments to whatever seems to suit at the time, to denying making posts which he definitely made. Elroch has no credibility here or elsewhere and neither do you. tygxc's credibility is another matter. He doesn't seem to be dishonest and he is capable of making good arguments, which you would not understand, since your entire approach consists of an attempt to discredit him. Neither would the 16 year olds you claim are mathematicians understand it. Your arguments count for nothing and given his behaviour, neither do Elroch's.

Joined 19 hours ago and claims Mega is "smart". Is this someone to believe? Is he smart enough to know Mega is smart? Mega isn't smart so there's negative evidence on that score.

Optimissed I dont even need to worry about people believing you lmfao, your claims are so childish they literally have no substance to be believed, let alone addressed.

When I say mathematicians, I mean people who have PHD's in math and have written peer reviewed papers pushing the boundaries of mathematics. They actually told me that It was stupid of me to be wasting my time on people like you and tygxc.