I don't know much about programming, but why wouldn't it be possible to make an absolutely PERFECT computer chess program that always played 100% perfect chess and always makes the bestest and exactest moves each and ever time for ever and ever?

Computer chess interests me because it is trying to use a combination of raw processor power & Brute Force, with elegant and super efficent algorithms to provide the best SOLUTION to a multivariable dynamic and ever changing 'problem'.

Life, like chess, is also a problem that has many solutions. Humans very seldom make the best choices, but so far sometimes our ESP and intuition and subconcious can still beat computers and brute force thinking.

The world at large can be seen as the grand chessboard that is a gigantic 'problem' just waiting to be 'solved' with the optimal solution. With peak oil, peak water, global warming and dimming, fall of us empire, decline of petrodollar hemegony, and ever increasing geopolitical tensions throughtout the world due to intrinsic lack of physical resources and extractable energy, our world and future timeline does look very very bleak and gloomy indeed. If only we had a computer program that would calculate all the variables, all the possible moves of each nation against one another in this dynamic arena, and output the best optimal path to take at each crucial junction in conciousness-space-time.

So let me get to the point. Chess has a finite number of pieces, with a finite number of squares, and a finite number of rules. Chess is also a game that has a finite number of moves, and only three endgame possibilities: Win, Lose, draw.

Thus,,,,, Chess is FINITE. In that sense, every game that COULD ever be played, every combination of moves, every permutation and every possible choice and action at every step along the way is simply another BRANCH to take. Some branches branch out more than others, some branches are long than others, but eventually when one finishes out all the branches you arrive at the leaf of win, lose or draw.

It is akin to the multiverse omninum existence that we live in, each and every possible different action at every singular distinct point in conciousness-space-time branches off into a different universe, a different dimension and a different existence. Our 'universe' is simply the one of the many that just so happened to go down *this* path in the space-time contiunumm timeline....

So with a perfect Chess program, with a perfect hardware (preferably quantum computer) we can 'map' out every chess game that has ever been played and that will EVER be played!!!!! There is no 'guessing' and no 'room for error'. The past, present, and future is all ONE!
So essentially even before the game ever started, before White's first move, the battle is already won or lost! It is a superposition of all things possible without having to do anything at all!

I don't see why not a chess program like this cannot be made.
There are only so many combinations and only so many possible ways to play the game. If we can map out every one of them, then there will be no winners and never any losers!

A 'perfect' game of chess can be played without either side doing anything at all! Infact, like a library index system where each character branches off a unique character of the book, eventually when you get to the book you realize the index IS the contents of the book itself, then a perfect game of chess is nothing more than both sides (players) choosing from and agreeing upon which particular 'index'/'path'/'game' of chess in the grand INDEX of the totality of ALL possible games of Chess GIGANTIC COSMIC CATALOG they want to 'play'.

They don't of course, actually 'play' at all, what they do is choose a particular timeline, a particular string of 'events' that is included in the comprehensive giant catalog of chess games, and simply let that 'game' BE. That abstract, platonic existence, simply IS, and exists... and without a need or compulsion to do *anything* from either side.

Imagine if CHESS was only 4*4 board with only 16 squares total and only 4 or less chess pieces per side with a miniturized board. Then it would not be difficult to write a 'perfect' chess program that knew ALL the position combinations and thus never made a mistake. (however such a smaller chess game is so simple it would be unplayable, but thats not the point) It is indeed possible to map out ALL possible games, thus making chess a 'catalog' so to speak... They might seem astronomically huge and impossible with current hardware, but at least on a hypothetical level it is possible, because it is NOT inumerable .

Zero and Infinity, everything and nothing, and one cannot exists without the other.

[Buffy]

The brute force algorithm is the most popular one in most chess programs, but more recently even those that have massive amounts of horsepower and memory use both arbitrary depth pruning of the branches and focus on standard openings and responses when there are lots of options seemingly available.

When you start to do the math on all those "finite" possibilities, you start to see that we're a long way from having that sort of power available on a desktop, or even with all the computers in the world tied together working on that one game.

I'll leave the math problem to you as an exercise: its quite fun and mind boggling!

Put one kernel on the first square, 2 on the second, 4 on the third,
Buffy

Originally Posted by Buffy

The brute force algorithm is the most popular one in most chess programs, but more recently even those that have massive amounts of horsepower and memory use both arbitrary depth pruning of the branches and focus on standard openings and responses when there are lots of options seemingly available.

When you start to do the math on all those "finite" possibilities, you start to see that we're a long way from having that sort of power available on a desktop, or even with all the computers in the world tied together working on that one game.

I'll leave the math problem to you as an exercise: its quite fun and mind boggling!

Put one kernel on the first square, 2 on the second, 4 on the third,
Buffy

I think the point was not to invent a computer program that could do brute force. (I don't think Star Trek LCARS computer could handle that kind of stress....) There are many programs out there that already play at the level of GM's or even better, such as Rybka or Hdyra for example.

My point was simply to point out that CHESS is a 'dead' game in the sense that there is a FINITE and definitely NUMBERABLE (COUNTABLE) number of possible unique games to the entire game of chess.

Thus purely HYPOTHETICALLY speaking, one could make a large enough hypothetical CATALOG of say EACH AND EVERY POSSIBLE chess game that COULD EVER BE PLAYED. And since this CATALOG will be a complete mapping to each and every game (since it is one on one and contains of course EACH and Every game) no matter what game you and your opponent chooses to play, it is already in the index of the CATALOG and you can simply look it up.

So instead of playing CHESS, we could just simply AGREE upon with our opponent how the game will begin, how it will end, and everything in between, we could simply come to a consensus and together CHOOSE a page from the CATALOG and say that we 'played' the game without either of us ever even moving a single piece.

[Buffy]

Originally Posted by chen2739

My point was simply to point out that CHESS is a 'dead' game in the sense that there is a FINITE and definitely NUMBERABLE (COUNTABLE) number of possible unique games to the entire game of chess....what game you and your opponent chooses to play, it is already in the index of the CATALOG and you can simply look it up.... So instead of playing CHESS, we could just simply AGREE upon with our opponent how the game will begin, how it will end, and everything in between, we could simply come to a consensus and together CHOOSE a page from the CATALOG and say that we 'played' the game without either of us ever even moving a single piece.

Existentially, sure. Practically, consider the time it would take to find the pages! Wouldn't it simply be easier to play a game? And maybe more fun too?

Again, do the math, if you haven't already because its illuminating even if you are asking the philosophical question....

And if you *are* asking the philosophical question, this belongs in Philosophy and Humanities, not Chess!

Your move,
Buffy

[Eclogite]

The rules of chess are not fixed. They have evolved over time. There is no reason to believe that they will never change again. Therefore the solutions are not necessarily finite. Even if they are, by the time you have figured out all the games for the current rules, the ruless will have changed again.

[pgrmdave]

Question - do the number of games converge near the end game?

[CraigD]

Originally Posted by pgrmdave

Question - do the number of games converge near the end game?

The number of possible choices for moves become smaller in the typical end game, because there are typically fewer pieces. The number of possible boards (states) is smaller with fewer pieces. However, the usual meaning of “number of games” is the count of leaf nodes in the game’s game tree, also called its “game tree complexity”. By this definition, it doesn’t make sense to say the number of games converges – although often difficult to calculate or even accurately estimate, for a particular well-defined game, this number is a constant.

The game tree complexity of chess has been studied for a long time. The famous “father of information theory Claude Shannon made a pretty good estimate of it in the 1950s: . More modern estimates are closer to .

This is a huge (though finite) number. By comparison, there are estimated to be atoms in the visible universe. So chen2738’s point that, hypothetically, one could write a book containing every possible chess game, is valid, as is Buffy’s suggestion to “do the math”, which will make clear that actually creating such a book, using any media, is impossible – there’s not enough material to do it. The best we can hope for is to build the part of the game tree we’re interested in “on the fly”, using a combination of brute force and clever pruning to make it possible on available computers.

In terms of actually playing competitive chess well, the best approach seems to be not brute force + pruning, but learning to recognize “book” positions by excluding irrelevant details – something human beings are amazingly good at, which has allowed human players to remain competitive against very recent computer opponents.

The near impossibility of calculating chess’s game tree is what prevents it from being, in any but the most abstract, hypothetical sense, a “dead game”. 3x3 tic-tack-toe, on the other hand, with a game tree complexity of only 26,830 (excluding rotations and reflections of the board) that has been implemented in media from paper to mechanical computers built of tinker toys to the brains of millions of six-year-old children, is for all practical purposes a “dead game”.

[ughaibu]

A short article on solved board games: Solved game - Wikipedia, the free encyclopedia

[nimzo]

Originally Posted by CraigD

The number of possible choices for moves become smaller in the typical end game, because there are typically fewer pieces. The number of possible boards (states) is smaller with fewer pieces. However, the usual meaning of “number of games” is the count of leaf nodes in the game’s game tree, also called its “game tree complexity”. By this definition, it doesn’t make sense to say the number of games converges – although often difficult to calculate or even accurately estimate, for a particular well-defined game, this number is a constant.

The game tree complexity of chess has been studied for a long time. The famous “father of information theory Claude Shannon made a pretty good estimate of it in the 1950s: . More modern estimates are closer to .

Originally Posted by Chen2739

Imagine if CHESS was only 4*4 board with only 16 squares total and only 4 or less chess pieces per side with a miniturized board. Then it would not be difficult to write a 'perfect' chess program that knew ALL the position combinations and thus never made a mistake. (however such a smaller chess game is so simple it would be unplayable, but thats not the point) It is indeed possible to map out ALL possible games, thus making chess a 'catalog' so to speak... They might seem astronomically huge and impossible with current hardware, but at least on a hypothetical level it is possible, because it is NOT inumerable .

The 3x3 chess board has been solved with 304,545,552 positions !

kd.lab.nig.ac.jp/3x3-chess/ (sorry can't post links)

nimzo

[DougF]

Playing Chess... is like watching a movie (for the first time) with a friend who has seen it before you are not going to enjoy it if he keeps telling you what is going to happen!