Will computers ever solve chess?
Hans Berliner, a correspondence Master/Champion, wrote a book in 1999 called The System in which he puts forth analysis of various continuations after 1.d4 , depending on Black's adopted defense, in hopes of transmuting White's half move advantage, into a space advantage, and eventually into a Win for White. Note: Weaver Adams tried looking for a forced win after 1.e4, rather than Berliner's 1.d4.
Anyway, here I've assembled some contemporary computer analysis of Berliner's 1999 published analysis to see how it holds up and whether White's advantage is, in fact, increased:
http://oeco.hopto.org/mediawiki/index.php/Berliner,_Hans_-_The_System
*No ads, no sign-up... Just a hobby of mine that I make available to everyone.
Why not let Alpha Zero play against itself about 10,000+ times, and then make a rough assessment regarding whether a "perfectly played game" might be a "forced draw?"
Why hasn't that been done yet?
Sounds like a relatively simple and straight-forward proposition to me.
Why not let Alpha Zero play against itself about 10,000+ times, and then make a rough assessment regarding whether a "perfectly played game" might be a "forced draw?"
Why hasn't that been done yet?
Sounds like a relatively simple and straight-forward proposition to me.
Alpha Zero already played far more than that many games against itself during its machine learning...doesn't mean anything. Engines play might be better than humans, but are effectively no closer to solving chess.
Robots and computers are the raise of the machines. Chess is perfected in the future online chess computers dominate this game.
They can solve science and engineering problems thus : they live with humans and make war against us.
3000 AD only robots and computers survive. They play against another robots and computers from alien technologies.
What is silicon, is from alien comets. brain of robots destroyed by other aliens fall on earth and we make the computers.
I'm not sure that a computer, like Alpha, playing itself, makes any sense. To me, it seems like every time a computer plays itself it has to win or draw. If it lost it would be like it's playing a joke on itself. How could it lose when it's figured out it's move? It's next move would be what it already decided due to it's previous move. Unless it discovered some new move it didn't count on, which seems impossible.



