LawAndOrderKing wrote:
btickle might be interested
Are you going to reply to your other sockpuppet all day? Even as a troll you are lacking.
Will computers ever solve chess?
Sort:
To solve chess practically "merely" requires chess software that never plays an inferior move. It may be hard to know when this has been achieved.
To solve chess completely requires software that not just plays perfectly but also knows for every legal position what the result would be with best play by both sides. It may be hard to know when this has been achieved.
btickler wrote:
...There are no papers showing a method for solving chess that can actually be achieved by mankind currently...
correct - a lack of a solution is not proof that a solution cannot be found.🤠
RCMorea wrote:
...do you have any reason to suspect it could be done? Personally, I suspect it can't, but obviously that isn't proof...
I'm with you 100% And I agree that a very legitimate argument can be made that chess will never be solved.
Since Shannon wrote his famous paper in 1949 we all know that an exhaustive analysis of the game tree can't be used to solved chess - it's just too complicated.
So solving chess will depend on new analytical methods. It can also be made easier if chess is a win for white, and it comes from less than 40 moves (the Shannon number is based on a game length of 40 moves I believe).
As for examples of other math problems that were considered "solved" but then found to be incomplete is the "tiling of a plane problem" (5 examples here - from another thread).
In 1968 Kershner said there are 8 types of pentagons that tile a plane, and said the list was "complete".
Then from 1975 - 1977 five more shapes were found for a total of 13. Progress stopped.
In 1985 another was found, and progress stopped again.
Then, two years ago, a 15th shape was found (here):
So chess could have a similar surprise. It just might not happen in 5 years, or 50 or even 200. But respected mathematicians aren't saying that "chess will never be solved". 😊
I am helping to solve chess by reminding the world that you two don't know what you are talking about. It saves people time when they read this debacle of a thread.
I have never said that "chess will never be solved". I said that it is currently impossible and that pining away for a solution in our lifetimes is a fantasy. Pin your strawman somewhere else.
vickalan wrote:
btickler wrote:
...There are no papers showing a method for solving chess that can actually be achieved by mankind currently...
correct - a lack of a solution is not proof that a solution cannot be found.🤠
Burden of proof is yours. Own it. Make a little graph if that helps you 
...
.. Earlier, in response, to my following, 'poker'- 'A.I.' software link ; The following question, was asked, by "ponz111" :
.. Read it, and Weep !
https://www.technologyreview.com/s/603544/an-ai-poker-bot-has-whipped-the-pros/
Will be interested to see what it means that it splits its bets into 3, 4, or 5 different sizes--something humans cannot do--what this means and how it works?
.. Anyway ; It has since occurred to 'moi'; What the above, posed question, {that appears, in the above link}, was referring to ; Namely ; If both, the 'A.I.' poker 'bot'; Plus, its' human opponents, were to play, 3, 4 or perhaps, as many as 5, simultaneous 'poker' hands ; With betting, at different levels, {known as 'split bets'} ; So as to offer, a Greater statistical sampling, of their respective, human- artificial intelligence, 'poker' playing sense.
This competition, conceivably, could go on for ages ; With, neither 'side', achieving, a likely insurmountable, statistical advantage ; Before, the spectators & 'players', decide to-call-it-quits ! <- Well, that's one, possible outcome ! o:
btickler wrote:
Burden of proof is yours. Own it. Make a little graph if that helps you ...
Brutum fulmen, Nihil est proven.👹
NelsonMoore wrote:
To me the surprise is not that Alpha Zero was able to draw with Stockfish8 but that it was able to defeat Stockfish8, and that Stockfish8 is even beatable. I would have expected every game to finish drawn. I believe they were playing 60 seconds per move which means that the engines had enough time to do a deep enough analysis.
Agree, AlphaZero is a pretty amazing machine. It did play on superior hardware, but it's still an amazing accomplishment. The fact that both sides had wins, and that some wins were even from Black means that the rumored "draw death" of chess isn't here yet.😊
Dec 11, 2017
0
#2532
Actually Stockfish won no games when AlphaZero had complete freedom: it achieved wins when Stockfish was forced to play a range of specific openings.
It will be a lot more impressive when AlphaZero actually beats Stockfish straight up at the next TCEC...
s23bog wrote:
btickler wrote:
I am helping to solve chess by reminding the world that you two don't know what you are talking about. It saves people time when they read this debacle of a thread.
I have never said that "chess will never be solved". I said that it is currently impossible and that pining away for a solution in our lifetimes is a fantasy. Pin your strawman somewhere else.
You are trying to help people solve chess by saying that it cannot be done?! Can you explain your reasoning, please?
I'm helping clear people's minds for the possibilities of future advances by helping to keep gibberish in its proper perspective.
It is not a task that can be accomplished in our lifetimes, nor in any foreseeable future. So, talking about it happening is like talking about Bigfoot.
You should solve P=NP next
Computers only do brute-search.. they can never "invent" strategies like center-control, mobility, pawn-structure, bishop/knight theories which drives a human mind towards the correct moves without having to do brute-search.
Dec 12, 2017
0
#2537
How? It can't happen with conventional computers (see the rest of this forum for why!)
.. Obviously, it's speculation ; But since, 'quantum' computers, are still in their development infancy ; Perhaps, there are already, mathematical projections, {or estimates} ; As to how long, it might take, a hypothetical, quantum 'supercomputer' ; Capable, of sorting through, mind-boggling, permutations ; Given that, the 'quantum' approach ; Has supposedly equal, daunting capabilities {!?! }
Stauntonmaster wrote:
In the next 10 to 20 years many chess engines will solve the problem of chess. Computers have alrady solved the chess problem against humans. No human can beat a good chess engine which means the computer cam always tell us exactly why we lost. Therefore, in a way chess problem is solved. The correct question is if computers can ever play 100% perfect chess? This question can become the topic of discussion.
By "problem of chess" perhaps OP meant 100% perfect chess. Also, if the computer does solve the problem of chess, maybe the humans will give it a tougher challenge? Maybe some sort of generalization of the game? I guess framing problems is something that the humans can do better than the machines.
.. Read it, and Weep !
https://www.technologyreview.com/s/603544/an-ai-poker-bot-has-whipped-the-pros/
Will be interested to see what it means that it splits its bets into 3, 4, or 5 different sizes--something humans cannot do--what this means and how it works?