Will chess ever be solved?

There are much less than one million positions in chess.
64 squares

• 13 possible figures on each square (K, Q, R, B, N, P with black and white or none piece at all)
• 32 (with white able to castle queen side or kings side same with black or no castle at all)
• 16 possible en passent squares
• 2 each position with white or black to move
  = 851968 possible positions most of which are illegal.

You can store this in a database.

After that you delete all positions with more then 16 black pieces.
After that you delete all positions with more then 16 white pieces.
After that you delete all positions with more then one white king
After that you delete all positions with more then one black king
After that you delete all positions with allowed white castle and no king on the first rank.
After that you delete all positions with allowed black castle and no king on the eighths rank.
After that you delete all positions with a pawn on the first or eighths rank.
After that you delete all positions with an en passent square but no pawn which can capture.
After that you delete all position with an en passent square but no pawn which can be captured.

That brings your position count down to five figures or a small six figure.

After that you let a multitude of computers access this database and

1. Find all positions which are checkmate and mark them as such
2. Find all positions which are stalemate and mark them as draw
3. Find all positions with to few material to win (K vs K, K vs K+B, K vs K+N) and mark them as draw
4. Find all possible connections from one position to another with a legal move and store the result in another table.

Then you just have to follow the path from the known start positions while avoiding recursion.
And voila you just soft solved chess and chess 960.

After that you just have to flip it around starting from all marked checkmate positions and go backward avoiding recursion and cut all branches at the checkmating party having the move which does not result in a checkmate.

And you have solved chess and chess 960.

If you reach a starting position position chess is won for the site resulting in checkmate.
If you do not reach a starting position chess is a draw.

With this database based design you can let millions of computer do the work. The database itself with a clever use of constraints (Primary Key, Unique Constraints and Foreign Keys) will make sure that no doubling takes place.

Problem being that the database will turn really big really fast

@fuxx_de said in #21:

There are much less than one million positions in chess.
64 squares

• 13 possible figures on each square (K, Q, R, B, N, P with black and white or none piece at all)
  I'm afraid you've already made a mistake here: if we assume that there are 13 possibilities for each square independently from each other, then the amount of possibilities is NOT 13 multiplied by 64, but 13 to the power of 64 - which is a "little" bit more.

@Ayushman_C said in #1:

With reference from en.wikipedia.org/wiki/Solved_game, "A solved game is a game whose outcome (win, lose or draw) can be correctly predicted from any position, assuming that both players play perfectly" and "In game theory, perfect play is the behavior or strategy of a player that leads to the best possible outcome for that player regardless of the response by the opponent."

Popular games that have been 'strongly' solved include tic-tac-toe (intuitively guaranteed a draw on perfect play, though certain simple strategies can be used to confuse the opponent), Connect-4 (where the first player can force a win) and loads more, which you can check out in the Wikipedia link above.

The article also states that 8x8 Draughts has also been solved, where, with perfect play, "both players can guarantee a draw".
Now what I'm really curious about is given the complex nature of chess (there are more possible positions than the number of atoms in the universe), will it ever be solved? And if yes, when do you think chess will be solved?

The number of positions exceeding some limit doesn't prevent a game from being solved. HeX for instance, is solved, in the sense that there's a simple argument which shows there is a winning strategy for the player going first, regardless of the size of the board. Even for a 1,000,0001 x 1,000,0001 sized board. (The strategy itself isn't known though). Many NiM games have been strongly solved as well: for any starting position it's known which player has a winning strategy, and the strategy is known. And NiM games can be as large as you want, resulting in more possible positions than chess.

A high number of possible positions makes it infeasible to use brute force to solve the game, but it doesn't prevent other ways to solve it.

@caughtinarut said in #22:

I'm afraid you've already made a mistake here: if we assume that there are 13 possibilities for each square independently from each other, then the amount of possibilities is NOT 13 multiplied by 64, but 13 to the power of 64 - which is a "little" bit more.

You sure are right and I stand corrected.

A apologize.

@fuxx_de said in #21:

There are much less than one million positions in chess.

Let's just count the number of ways to play 8 white pawns on the board, under the restriction no pawn as been captured, promoted, or made a capture itself.

There are 6 squares the a pawn can be.
There are 6 squares the b pawn can be.
There are 6 squared the c pawn can be.
There are 6 squares the d pawn can be.
There are 6 squares the e pawn can be.
There are 6 squares the f pawn can be.
There are 6 squares the g pawn can be.
There are 6 squares the h pawn can be.

And they are all independent. That already gives more 1.6 million ways just to place the pawns of white. If we factor in black pawns, under the same conditions of no captures nor promotions, there are 15 ways to place a white and black pawn on a file. With 8 files, that's already 2.5 billion ways to place pawns on the board, even given the restriction of no captures.

You are way, way, way off with your estimate of the number of positions.

@fuxx_de On my computer Stockfish reaches 1.9M nodes in one second. If chess had fewer than 1M positions, a modern engine would easily solve it in less than a second.

I've tried doing it before. One minute Stockfish spits out a Ruy Lopez, the other it tells me to go for a Nimzo-Indian. That's how I ended up with 10 "perfect" chess games - all at 100% acc and 2 average centipawn loss.

Bigger question: Will life ever be solved?

Answer: 42

Chess must be easier than that.

The wikipedia article linked by @Ayushman_C in the opening post does draw distinctions between different levels of a game being solved.

There will clearly never be a 32-piece tablebase because of practical limitations as already hashed out in this topic. But that isn't necessary for the outcome (win/draw/loss) with correct play to be known for any legal chess position, even though the outcome of each position will have to be calculated afresh each time rather than stored.