Computer Program/chess

Computer Program/chess

Post just about everything that isn't directly related to Spring here!

Moderator: Moderators

User avatar
Gota
Posts: 7151
Joined: 11 Jan 2008, 16:55

Computer Program/chess

Post by Gota »

Is there a perfect playing chess program yet?
User avatar
knorke
Posts: 7971
Joined: 22 Feb 2006, 01:02

Re: Computer Program/chess

Post by knorke »

perfect as in playing really good?
"Fritz" won against some chess worldchampion from russia a few years ago.
Kloot
Spring Developer
Posts: 1867
Joined: 08 Oct 2006, 16:58

Re: Computer Program/chess

Post by Kloot »

No.

Chess is not a fully solved game like tic-tac-toe, and will probably never be.

Unless by "perfect" you mean something else than a winning outcome from any position against an optimal opponent, in which case today's programs will keep you quite busy.
User avatar
Gota
Posts: 7151
Joined: 11 Jan 2008, 16:55

Re: Computer Program/chess

Post by Gota »

I mean being able to completely calculate all possibilities and moves from any position.

I dont get in in the 1990s there were already progs winning world champs...Were in 2010 with much more powerful computers and we still dont have a prog that can calculate the game out completely?
User avatar
Caydr
Omnidouche
Posts: 7179
Joined: 16 Oct 2004, 19:40

Re: Computer Program/chess

Post by Caydr »

Funny you should ask, I was reading about this just recently. They're working on it, but it's kind of like calculating Pi after a while.

They've got the "best" opening moves all worked out, the "best" ways to counter these, etc etc, and the same deal for every conceivable game ending. It's the in-between that's tricky.

Then you've got the scenario where a computer or the human plays not to win, but to stalemate, or to outlast, or with some other objective in mind, and these open another whole can of worms about what the best moves are.

Honestly I hope Chess is never "solved", there's something very deeply disturbing and unappealing to me about the idea that there is a right and wrong way to play a game besides following the rules and trying to outwit your opponent.

The best way to start your game is to move a certain pawn... I believe it's queen's pawn forward 2 squares. "English Opening"? With that knowledge in mind, I usually start with a knight to the middle or knight's pawn forward 1. I absolutely hate mechanically playing to win.
User avatar
Gota
Posts: 7151
Joined: 11 Jan 2008, 16:55

Re: Computer Program/chess

Post by Gota »

well It would be interesting what sort of results such a program would produce and how it would probably eliminate many wrong moves...
Who knows maybe moves that are taken for granted as good moves would actually turn out to be bad ones.
Obviously after such a program is made chess would immediately become less appealing.
Kloot
Spring Developer
Posts: 1867
Joined: 08 Oct 2006, 16:58

Re: Computer Program/chess

Post by Kloot »

You don't need to be able calculate all possible chess branches to beat a grand master.
gota wrote: dont get in in the 1990s there were already progs winning world champs...Were in 2010 with much more powerful computers and we still dont have a prog that can calculate the game out completely?
No.

The total number of valid game paths (you do a move, your opponent moves in response, etc.) in chess is somewhere in the order of 10^120. If you had a computer that could calculate ten billion (10^10) moves per second, your program would still take 10^110 seconds to evaluate them all.

The age of the universe (~14 billion years) is roughly 4.4*10^16 seconds. Still think it's possible?
Last edited by Kloot on 14 Apr 2010, 17:43, edited 2 times in total.
User avatar
Caydr
Omnidouche
Posts: 7179
Joined: 16 Oct 2004, 19:40

Re: Computer Program/chess

Post by Caydr »

Mmm... actually now I'm a lot more confident it's possible. While I know nothing about anything, every time someone says "such-and-such is impossible it would need this many calculations a second! Ha ha ha", I start planning for how to save up the money to buy one when it's released next year.

Thanks for ruining chess, ruiner.
User avatar
knorke
Posts: 7971
Joined: 22 Feb 2006, 01:02

Re: Computer Program/chess

Post by knorke »

I dont get in in the 1990s there were already progs winning world champs
I think I've picked up something how some of the sucess from those computers was because the human players were not used to their play style and could not prepare to their style. Like they could not study games previously played by the program while the programs have analyses of all games ever played and maybe were even adjusted to exploit weaknesses of the specific opponent.
User avatar
Caydr
Omnidouche
Posts: 7179
Joined: 16 Oct 2004, 19:40

Re: Computer Program/chess

Post by Caydr »

True, and besides they probably weren't taking them seriously at first, if ever. There's not a lot of career opportunities for chess players - when offered a sum of money to play against a computer, of course they'd take it regardless of whether they felt it was fair or not.

Real chess experts can study their opponent's prior games in advance if they choose, or study their composure as they play. They can bluff, frustrate, or trick their opponents in any number of ways. It's like playing poker against a computer... when half the game is studying your opponent, and your opponent is just playing the odds and counting cards, it's not poker anymore.

A big reason I enjoy chess is the social aspect. I don't play with a timer, I don't take it too seriously at all really. It's a game of wits - it doesn't matter if a computer solves chess or if a computer can consistently beat every human player on the planet 100% of the time, it's not playing the same game so it's irrelevant.
Google_Frog
Moderator
Posts: 2464
Joined: 12 Oct 2007, 09:24

Re: Computer Program/chess

Post by Google_Frog »

I don't see how you worked bluffing into chess. Sure you can lure them into a false sense of security but it is never a real bluff as all the information in the game is available. As it has complete information I think there is an optimal response to every move.

Look on the bright side. If chess is solved everyone can move on to Go.
User avatar
Gota
Posts: 7151
Joined: 11 Jan 2008, 16:55

Re: Computer Program/chess

Post by Gota »

You can be a romantic but ignoring reality is pointless..
You can go on saying what YOU want chess to be..
you dont take it seriously yet its a game of wits?Makes 0 sense.
If you dont try to win as hard as you can versus your opponent than your not playing chess your playing some made up game in your mind.
You talk of tricks and all sorts of way to win but without playing seriously none of them will be deployed.
Lets play chess,you will not play seriously and ill check mate you in 4 moves and we'll call it a game.

The Spring equivelent of what your saying would be to make "artillery on comet catcher".
the right thing would be to make flash but hey,your not taking it seriously,"its a game of wits".
If you dont make the flash on comet catcher than there is no game and there are no complex and interesting strategies since anyone else would steam role you.

All those psychological tricks are a result of us not being smart enough to calculate the game completely..It is not the computers fault the game can be calculated completely,maker a better game if you dont like it,don't blame the player for breaking the game for you by playing better.

If I play a game versus you will you complain when i send some scout over to kill your mexes?saying im a noob cause i use lame rush tactics?
User avatar
Gota
Posts: 7151
Joined: 11 Jan 2008, 16:55

Re: Computer Program/chess

Post by Gota »

Google_Frog wrote:I don't see how you worked bluffing into chess. Sure you can lure them into a false sense of security but it is never a real bluff as all the information in the game is available. As it has complete information I think there is an optimal response to every move.

Look on the bright side. If chess is solved everyone can move on to Go.
There is bluffing...it is again a result of our brains inability to calculate all possibilites..
There are 2 terms in chess one is true sacrifice and another is pseudo sacrifice.
You can sacrifice material,i.e. a piece to gain some other advantage like momentum or the ability to take important parts of the board or even to quickly regain your piece disadvantage plus some extra...

The other player does not know if you sacrifice cause you see some immediate pay off and he doesn't or its cause your hoping it will pay off on the long run.
Who knows maybe some pieces actually interfere the way to victory in some situations...Maybe at certain situations its better to sacrifice certain pieces cause they are only in your way...
That can only be known when the game will be completely solved.
User avatar
zwzsg
Kernel Panic Co-Developer
Posts: 7052
Joined: 16 Nov 2004, 13:08

Re: Computer Program/chess

Post by zwzsg »

Gota wrote:That can only be known when the game will be completely solved.
Did you get the message about how the game cannot be completly solved?

It's a physical limitation, not matter how far in the future, no matter how powerful computers will be, this universe is too limited to evaluate all chess states.
User avatar
Hobo Joe
Posts: 1001
Joined: 02 Jan 2008, 21:55

Re: Computer Program/chess

Post by Hobo Joe »

And once again we prove that the human mind > computers?


I remember seeing some high-budget presentation about the state of computers in our day and age, and it talked about how todays supercomputers are superior to the human mind, and in 10 years everyday personal computers will be equal to the human mind. My only response to that is that the person who made that presentation might not be smarter than a computer, but I sure as hell am.

Even disregarding hardware which may or may not be close to the level of the human brain, everything a computer does has to be programed. There are rudimentary 'learning' programs but they are just that, and can't match a human. Computers are really good at math but the human mind doesn't work by doing calculations, in fact we really know very little about how our minds work at all, even on a very basic level. I doubt we'll bring computers even close to that level anywhere in our lifetimes, if ever. Is it even possible for us to make something more intelligent than we are? (this is the part where we all cower in fear of a self-inflicted apocalypse at the hands of a superior race that we created)



Not really sure where I was going with that.
User avatar
JohannesH
Posts: 1793
Joined: 07 Apr 2009, 12:43

Re: Computer Program/chess

Post by JohannesH »

Better than a human grandmaster, or a great chess computer, is when they work on the same team.
User avatar
knorke
Posts: 7971
Joined: 22 Feb 2006, 01:02

Re: Computer Program/chess

Post by knorke »

zwzsg wrote:
Gota wrote:That can only be known when the game will be completely solved.
Did you get the message about how the game cannot be completly solved?

It's a physical limitation, not matter how far in the future, no matter how powerful computers will be, this universe is too limited to evaluate all chess states.
If possible chess moves=10^120 is correct it would be more than the number of atoms (10^80) so even just storing them would be difficult.

But I dont know if chess is perfectly solvable even in theory?
Completly solved/perfect game = one side can always win, no matter what.
Like in "3 in a line" (where you draw x and o on a # shaped board) the player wo makes the first moves can always win.
Dont think thats possible in chess?
It's a game of wits - it doesn't matter if a computer solves chess or if a computer can consistently beat every human player on the planet 100% of the time, it's not playing the same game so it's irrelevant.
People still compete in 100 meter sprints, bicycle races etc. even though every car is faster. If there was a perfect chess computer, chess might lose some of its "magic" though.
I doubt we'll bring computers even close to that level anywhere in our lifetimes, if ever. Is it even possible for us to make something more intelligent than we are?
Forgot the details but they made a maschine that did some experiments, calculations and other magic and then came up with some discovery on its on. It was not programed to find that special discovery (some physic law or something) yet somehow found it.
Better than a human grandmaster, or a great chess computer, is when they work on the same team.
boooo automicro widget :shock:
User avatar
Hobo Joe
Posts: 1001
Joined: 02 Jan 2008, 21:55

Re: Computer Program/chess

Post by Hobo Joe »

knorke wrote: boooo automicro widget :shock:

We're doomed! :shock:
User avatar
zwzsg
Kernel Panic Co-Developer
Posts: 7052
Joined: 16 Nov 2004, 13:08

Re: Computer Program/chess

Post by zwzsg »

Hobo Joe wrote:And once again we prove that the human mind > computers?
Nowadays even a medium chess program running on a cheap computer can beat any human. Where's your superiority now?

Also, I'd like to see a human running all the calculation of a typical BA DSD 8v8 game, and then drawing what happens 30 times per second, with only his organic brain, his hands, a paper, a set of pencils, and an eraser.

Hobo Joe wrote:and in 10 years everyday personal computers will be equal to the human mind
Even when computer will be superior to human mind in every aspect, there still will be people denying it, and claiming that the humand mind has such and such peculiarity that make it different and that computer can't have.

Anyway, already/still nowadays there are people claiming, believing and trying to convince others, that humans have a "soul" that transcend matters, is immortal, was granted by God, and similar nonsense.

So yeah, computer are already superior to human in many aspects, will probably one day be superior in every aspect, but that won't prevent some humans to keep believing they're "specials".

Is it even possible for us to make something more intelligent than we are?
Amoebas, fishs, lizards, mice, apes, all managed to "make" something more intelligent than themselves.
knorke wrote:
Indeed, the point could be made that maybe it's possible to solve chess without checking all possible states. Some problems can be solved with other ways than brute force...
User avatar
JohannesH
Posts: 1793
Joined: 07 Apr 2009, 12:43

Re: Computer Program/chess

Post by JohannesH »

knorke wrote:
Better than a human grandmaster, or a great chess computer, is when they work on the same team.
boooo automicro widget :shock:
Of course when playing a proper human vs human match, computer assistance like that ruins it in both at chess and rts.
Post Reply

Return to “Off Topic Discussion”