/sci/ - Science & Math

Providing you can think of every possible move by an opponent, Sure

No.

No, black always wins if both play perfect.

>>4554563
theoretically there can be infinite number of turns, so no

we're about at the stage where computers are just beating the best humans under match conditions.

>>4554581

>>4554581
We are talking about chess, not about finding a gf.

Such an algorithm can't exist. Imagine it playing itself. At most one player can win a game.

>>4554581
Doesn't white always win since it starts first and therefore has that advantage?

amazing game in your gif OP, pure skill

>>4554592
>>4554592

LOL

>>4554592

>>4554563

so far no there isn't

is it possible? not with our current computing power

the game theory complex tree of chess creates more possibilities than there are atoms in the universe

maybe quantum computers could do it..

http://en.wikipedia.org/wiki/Solving_chess

>>4554612
putin hates kasparov and vice versa

>>4554605
It is an unknown question whether one player can force a win with perfect play.

No.There is one for draughts, though.

Kind of meaningless though. I draw little fun from playing against computers. I love the look on people's face when you do something unexpected and ruin their plans. It's not all about doing the right or the wrong move, but appreciating the personality that each player bring to the game.

In the middle of a game, there is a certain number of moves you can do. Some are useless, others will make you lose a piece and so on. But the "best move" is only the best if it goes according to your plan and what you'll do in the future, it's about your instinct and also about ruining the other guy's plans, or misleading him to somewhere else. There is a wide variety of tricks, including stuff like "counting on the other guy's stupidity" or having a favourite and a least favourite piece to use, etc.

That's why it's hard for computers to calculate all the variables. Each move accounts for the whole game. They are getting there, though, and it's certainly possible. Personally, I think using computers for fapping and tweeting about doing your nails is more useful than having them playing chess. Just because we could build such computer, doesn't mean we should.

>>4554627
its pretty clear that chess is a draw

>>4554563
>win each time
Unknown.

The size of human played chess databases combined with clever and powerful computing will solve chess sooner than most think i believe. Perhaps even in our lifetimes.

>>4555447
We're not talking about 'should' or 'usefulness' you tryhard faggot.

>>4555468
But I am. If you can't handle it, close your eyes. I'm going in dry.

>>4555447

Holy fuck you are dumb.

Counting on the opponent's stupidity is for scrubs. Learn to play.

>>4554563
Sure. Build an exhaustive state transition matrix and then backtrack from wins, marking the transitions that lead to wins. Infinite moves would not matter since there are finite states.

During play, just pick whatever transition marked for your current state.

The only problem is that the storage space required for this is horrifically large.

chess has been solved for positions with 7 or less peices

http://en.wikipedia.org/wiki/Endgame_tablebase

chess has been practically solved with modern engines, meaning a skilled human player with a chess engine will be able to figure out the truth (win, loss, or draw) behind 99% of positions

http://lichess.org/us7fdjth

>>4555612
pascal lost

>>4555573
Anyone have that chart which includes progress on other games like Go and backgammon?

>>4554563
d4, your move, faggot

>>4556358
d5 nigger.

>>4556361
e4
WAT DO NAO

no

>>4556378

Nf6

ITT: http://www.youtube.com/watch?v=P9mwELXPGbA

>>4556390
3. e5
attacking the Knight
NAO WAT

>>4556451
Ne4

>>4556456
4. f3
bring it on!

>>4556470
e6

>>4556473
5. g3
closing in!

>>4556493
Qg5.

>>4556496
6. Bxg5
You're dead now.

>>4556504
Shit, I am. I concede. I screwed up. Good game sir, good game.

Computers are fast enough to play out every move combination and choose the optimal in a matter of seconds. So maybe not mathematician level "solved" but I don't think a human has beaten the top computer in over a decade or something like that.

They're working on one, I believe.
I remember a while back they mapped out every possible way a checkers game can go and then using a search algorithm from Google they figure out "optimum" strategy.

>>4556545
>Computers are fast enough to play out every move combination
Yes.
>and choose the optimal
No. Choosing the best move (or even a merely good move) is not the result of brute-force computation.

>>4556568

http://en.wikipedia.org/wiki/Human-computer_chess_matches

wikipedia disagrees. Matches *have* to give humans advantages or they can't even eek out draws. I would say that is checkmate in all practical terms, even if that is only 99.9% instead of an actual algorithm. Hell a cell phone program can reach grandmaster level. Game over humans. At least we can crush them at go... for now.

>>4556572
>weakiepedia disagrees
Oh, those weakies* who write Weakiepedia must be right about that! L0L
*weakies = Fischer's name for "weak players"

end games are getting solved because they are much, much simpler. i think they are solved to the 7th move out currently, but i havent checked in a while. computers will have to get much, much, much more powerful before a complete solution is ever attained. its at the point now that a computer could be designed to be nearly unbeatable though, if the money and effort was given to it in the same way that Deep Blue was funded. That isnt likely to happen though unless a company wants to show off or we get another bonjwa like Kasparov. The tech is there though.

>>4556564

I highly doubt anyone is working on an optimal chess strategy. Checkers is orders of magnitude smaller.

Most chess work is simply to make better, suboptimal algorithms.