True or false? Chess will never be solved! why?

For Tic-tac-toe it's known (solved), for chess that's an educated guess.

An educated guess that I read in a dusty old book by Wilhelm Steinitz. An even older book, by Andre Philidor, made this claim not for chess in general, but for gambits.

http://chessskill.blogspot.com/2013/06/philidor-on-gambits.html 

I suspect it's an accurate guess.  I will never know for sure though.

Chess is already solved, with perfect play it is a draw.

You don't need proof of this for all possible chess moves, but you just use the general principles. A draw is too easily achieved by repetitions, pawn walls and fortresses. History shows draw is the most common result. And engines always show an advantage less than 0.5 with near perfect play which usually ends up as a draw.

That's not "solved" though, that's evidence of what a solution might be.

Yeah sorry, I was trying to say we already know what the solution is, we don't need the proof.

I don't think we truly know though, short of a proof....

See post #82.

I think one day chess will be solved, 

we already have enormous performant computers and have solved various Minichess-Variations (Wikipedia: Minichess), so it's only a matter of time before we will have solved chess, computers will be able to solve chess completely...

But chess will still be worth playing, tho. Although the best first moves are known, the best defense is also known and it would be impossible for a human to memorize the best move for every position. It would be like memorising openings, like today already. Not to mention one could still make a bad move to confuse the opponent, or one could still trap opponents etc. 

At one point the game will be at aposition where neither you neither your oponent will be able to know the best move. You willstill need chess-undrstanding ...

That is a good point, GMs of today know almost all the best moves so that is what both sides play in the game.

However they don't know the mistakes with bad moves as well, because they rarely play against them. That is when we truly see their chess skills in action.

But it must take one side to play "bad chess" s we get to see some "real chess" and not memory games.

yes, pruning can be done, but it does not reduce the things to analyse by much. Instead of having 10^123 position to analyse, you will have 10^120.

Actually a vast majority of moves can be pruned...so the number will not be 10^120 (which is an inflated number anyway).  

Chess will probably be solved by calculating the game as intersecting lines of influence...potential preponderance of force in proximity to opponent's king with minimum force affecting your own king = good.  Not using current piece valuation and programmed criteria that we use today as shortcuts (like how we tell computers that doubled pawns are worse than it initially thinks), but purely preponderance of potential force.  

You can set a threshold for the analysis, such that you only do brute force move-by-move calculations on positions that actually have a chance of being winning lines by their "potential force score".  You don't need analysis to "prove" that, for example, moving all your pieces as close to the a1 square as possible throughout the game, or trying to write the letter "W" with your pawn structure, is going to suddenly produce a revolutionary breakthrough and "solve" Chess.  You don't need direct analysis to prove that moving a knight back and forth between 2 squares for the entire game is not best play.

That is a massively oversimplified explanation, but hopefully you can extrapolate how the vast majority of "possible" moves are not really up for consideration/calculation at all.  In any given board position, the majority of continuations (not single moves, mind you) are provable, nonsensical, dreck...and in sum total, this majority over all possible board positions eliminates a vast majority of the 10^123 (or 10^52, the more reasonable number).

Now, people can continue to argue that not checking every possible variation directly is not proof, but that argument is similar to saying "Just because dropping a grand piano from 20 stories on someone's head has killed them 100% of the time does not actually prove it will kill me unless you try it...".

The premise is mathematically, the perfect game which could either be a win for black or white or a draw. Absolute proof of the most logical outcome. I should imagine calculating all possibilities is too vast at the moment. Similar struggle was for fermat's theorum. Bobby Fischer solved the 15 square game I think and Lasker puzzled over these issues too.

I know guys... I was joking

Kasparov strongly believes that despite whatever advancement in technology, chess can never be solved. He certainly knows much more about chess than any of us. 

Yes, but as he has proven, he knows little about computers ;).

What does it mean to "solve chess"?  And why would anyone want to?  I don't understand what it means to solve chess, wouldn't that be like solving the weather?

The weather is already solved. If u tell me where u live, I can tell u what the weather there will be like tomorrow.

You absolutely need the former in order to determine the latter.  Solving chess is a proof.  If you are discarding entire subtrees because they are "clearly" suboptimal you must first prove that they are suboptimal -- with very, very few exceptions this is, out of necessity, a brute force exercise that requires you to examine the entire tree just in case.... meaning you haven't discarded it at all.

You don't have to examine every part of the tree.

This is more like geometry.  Build and prove your theorems on top of each other, and you can eliminate most of the tree.  Not theorems based on human trends and understanding (like coding into the scoring that bringing out a queen early is bad, for example), but raw absolutes (restricting a king's movements is generally good, more material is generally good, etc.).

You can prove that moving knights towards an opposing king tends to increase your chances of winning versus always moving them away from an opposing king.  A ridiculously simple example :), but it makes my point.

We have not "solved" the number pi, and we ultimately cannot say that we know what that number is.  Yet, we have proven how it works and we use it every day.

Pi is actually a good analogy here, because the argument is that we have to have to search the tree of possible Chess moves exhaustively, methodically, and linearly, by process of direct elimination.  This leads to a problem similar to calculating pi out to infinity.  Chess is not that type of problem, though.  It has rules and goals that narrow the tree considerably, and allow for subproofs and extrapolations that can be reliably applied to the whole tree.

In that thread linked above, someone talks about having to calculate best play with 9 queens on the board.  This is a perfect example of what I am talking about.  I guarantee you that you can prove that beyond "X" queens, the game is won barring a forced stalemate precondition built into the scenario given.  Past "X" queens, you can discard the whole rest of the tree.  Do I know what X is?  No.  But I know it can be calculated without traversing 10^123rd possible moves...

O, I remember! I played against u yesterday!

Generally... tends to...

This kind of vocabularily undermines anything that could be considered a proof.  No, I'm sorry, guiding principles need not apply; this exercise is about proof, not approximation.  You can't discard an entire subtree because it looks like it's pretty much determined -- you must prove it, or the rest of the proof crumbles around it.