Will computers ever solve chess?

Prove it.

I think not but if they do the solution will be,  White plays and black can draw on all moves on several ways.

If white opens with e4 or d4, he already has a slight advantage (according to all chess engines).  A perfect computer would calculate this and would open with one of those two moves.  Therefore, the solution cannot be black wins.  White always has to make a mistake for black to win.  Knowing what white has done, before having to move, is not an advantage in top-level chess.   If you give any GM a choice of white or black, he will always take white.   I really believe that the solution to chess (if there ever is one) will be a draw, either by repetition, perpetual check or insufficient material to mate (K+N or K+B vs K+P).   Rather than making a sub-optimal move, a computer will take a draw by repetition or perpetual check or sac a piece for a passed pawn in the endgame to force a draw..  

I understand your reasoning now.

If you don't see it, it does not exist.

Please keep your eyes open, lest you make the entire universe disappear.

it may be a kid or teenager on this thread now that solves chess tomorrow👽

It is interesting to explore the imprecise practical notion of advantage in a way which may be able to make it concrete.   The key intuitive notion is that an advantageous position is one which requires less computation to achieve a result.

At stands, this is not properly defined, but one can first define a Universe of passable chess computers, defined by the axiom that for any of them, the probability that they will find a correct move in a position will tend to 1 as they are given more and more time.

But we should not treat all these computers equally. Some are very slow and are to be ignored. A notion of quality is that a computer achieves some accuracy fast given some sort of limitation on storage used. This limitation is crucial, as there is a "trivial" blindingly fast, perfect computer that consists of the 32 piece tablebase, but this requires stupendous amount of storage.

I have been a bit loose, but I suspect that with a bit of use of information theory (to deal with the issue of coping with all possible opponents in an efficient way, by appropriately weighting the code to what matters most), there are probably natural metrics for advantage in positions (or at least a range of equally valid metrics associated with alternative, optimal given constraints, imperfect chess computers.

when quantum computers learn chess the answer is yes

1. There are more "things" that are impossible to happen than the few things that actually do happen.

2. It is completely logical to expect Black has an equal chance to force a win.

3. s23bog is illogical, contradictory, irrational etc. etc.

It is however an interesting question what would be good odds to offer on chess being a win for black. I would feel very comfortable offering 10:1 odds against this (and reckon it would not be reckless to offer much higher ones, say 100:1 or more).

The reason is that the empirical results of games of chess are probably a good guide to the correct result. Zugzwang is a relatively unusual feature of endgame positions where one side has a clear advantage, and other than that, the empirical, practical value of a half move of time is not in doubt.

Actually, the main elements of chess are power, space and time.   Time is the initiative or tempo.  White has the tempo until he makes a mistake (which, in this perfect solution scenario, cannot happen).   The idea that black has more information and therefore has an information advantage is incorrect.  White can open with 1. Nf3 and transpose into a variety of openings, if he chooses.   1. Nf3 gives him both the tempo and space and also nullifies black's "information advantage" in the opening.   In fact, I used to open with Nf3 for that very reason. 

To be more precise about this notion of white giving up a move, all white has to do is to find an opportunity to move a pawn, a bishop, a rook or a queen twice to reach a square it could have got to in one move to reach a position that could be reached by black (and which would be acceptable for black).

There are such a lot of opportunities to waste a move, this does seem highly plausible.

And you'd take seriously someone who said white wins with perfect play ? 

This statement is invalid, based on assumptions. With several board games it is a disadvantage to move 1st. Counter-attack, sacrifice of material if not accepted may well lead to an advantage for black. 

Point is: There simply is no way of knowing. Part of the beauty and intrigue of the game imo. 

In a sense, a universal value can be proven to be impossible, except for the values 1, 0 and 1/2 for the result with perfect play.

My sketch proof is as follows. Suppose you have access to a tablebase, then you can pick any single best move out of this table base and include it as a hard-wired choice in any computer program at small cost in time and space. So no individual moves are hard to ALL programs (indeed ALL individual moves are easy to SOME primitive programs by this grafting technique).

Therefore you need to judge a program by how hard it finds a large number of positions. The easiest approach (conceptually) would be to judge it by a weighted sum of how good it does in ALL legally reachable positions. Somehow it necessary to achieve a higher weight for positions that are of any interest, by being plausible in real games against competent opponents.

I hypothesise that a good definition considers programs playing against each other, like a good poker computer might be one that does relatively well against a range of other poker computers.

By "give up his extra move" I think he meant playing a move like a3 or h3, early on (before it attacks a pinning bishop).  For example, white can probably still force a draw by opening with 1. h3, if he then follows it up with more solid opening moves (kinda like a reversed opening).  Not sure if he could force a win with it though.  Maybe white can force a win with 1. Nf3 though, since it's a very solid move and is very non-committal too (eliminating any chance of black having an "advanced knowledge" advantage).   GM's have even tested this theory by opening with moves such as g3, g4 or b3.   Those moves don't control the center and could be seen as "wasted" moves by some players.   

Moves like d3 offer many opportunities to entire precise colour-reversed positions. There are no moves by black that you cannot follow up with d4 to reach a playable position identical to one which black could sensibly reach after one move!

As normal,  no one is getting any where with the solution.

Will computers ever SOLVE chess ?   YES.

It won't be a child, kid, teen or even a chess player, it will be a programmer.  And the program isn't even hard to design. 

Anyone care to say "NO' ?   Hope you know programming.

First :  YOU have no estimate of the hardware needed.  Second : there is a lot of hardware on earth, and very big.  Third :  There are many designs that reduce the tables needed to cover EVERY possible game.  Would you like some ?   But you did agree that we  can solve chess by recording every possible game, right ?

Fourth :  what does solve mean ?  Does it mean that for a given possition to know this is a WIN.  or  that we need to know that whatever black move is , what white needs to do to Win.   That's a big difference in the data recorded.  Which do you want "solve" to be ?   The program can do either.

I believe that "space" has to do more with how many squares on the board that you control and how limited your opponent is.   For example, if white controls the center, plus squares in the opponent's territory, while black is cramped and his pieces can't be mobilized easily.   That's a space advantage for white.   It's similar to mobility, but not exactly the same.  Time (or the tempo) is a constantly changing element, which always favors white until white loses it.   In theory, if this supercomputer plays all perfect moves, white should never have to give up the tempo.  That's the key to this debate right there.  White will always have the tempo advantage, in a perfect game, therefore white should never lose.   Even after 1. Nf3 Nf6, white still has the tempo.   Tempo overrides your perceived "knowledge advantage".   True, if I play Nf3, then I can't play f4 any time soon.   Although, since f4 is considered by most GM's to be a weakening move (if played too early), that's not a big deal.  I'd rather play it safe, open with Nf3 and then try to get f4 in later on, if that turns out to be my best plan in a certain position.   Usually d4 is better than f4 though and Nf3 doesn't prevent or delay d4 in any way.  No matter what black plays, I can get in d4, with Nf3, e3, d4 or Nf3, c3, d4.   An e-pawn player probably wouldn't want to open with Nf3 though.   Black can usually prevent (or blunt the effectiveness of) e4 if it's not played on move 1.  

Yes, but if white opens with 1. e3, he has no intention of playing e4.   He basically gave up that right and is ok with that.   Btw, 1. e3 could be another example of "passing the tempo to black".   1. e3 is just a weak opening move, since 1. d4, 1. Nf3 and 1. e4 are much better.   I actually experimented with e3, back in the 1990's, as a way of playing a French, with colors reversed (since I played the French, as black, at the time).  However, my best friend explained to me why it was inferior to the main opening moves, so I stopped playing it.   I believe it's called the Van Kroy's Opening.  lol  In fact, the French Defense is a good example of why the solution to chess is probably a draw.  The French Exchange and Rubinstein Variations are dead draws, with best play by both sides.   So, in theory, if black plays the French, white can force a draw with best play, with 3. exd5.   This may possibly be true with all book openings (although we don't know that for sure yet).