Gizmodo

This will take the fun out of the game if you can't beat the computer. But eventually a human will beat this new AI.

Yes, but the computer can't account for "feminine wiles"

@Kohono: Or if she's on that emotional rollercoaster when is "aunt Flow" visiting town.

It doesn't actually calculate all the moves. It calculates them from when it gets down to 10 pieces, and then never loses. Checkers ain't exactly rocket surgery.

No, a person will never beat this AI.

Checkers isn't as complicated as chess.

To put it another way, that's like saying that someday, a human will figure out how to beat a computer at Tic-Tac-Toe. It simply won't happen because the game is too simple. There is no clever strategy the human to use to fool the computer.

With this computation, checkers has become as trivial for a computer to beat as Tic-Tac-Toe.

in the WarGames remake (coming out in 2009 starring Shia LaBeouf and "The Rock" as the W.O.P.R.) i heard that they trick it at the end by playing endless games of checkers instead of tic-tac-toe!

looks like it was freshly (and factually!) updated after all! impressive!

I wonder what would happen if they pitted computer vs. computer?

@chemicalfish:
It would prob be endless. Like 2 peices left constantly moving around.

@chemicalfish:
the world would explode!

@CURSEMASTER0: Actually, a human will never beat this AI because it's unbeatable... the best they could do is a draw. That's because it's already been calculated exactly what the computer should do in response to any board state you can throw at it.

And, actually, on that note, there's an error in this article. They didn't analyze all 500 billion billion checkers moves. Instead, they only had the computer calculate until it found a winning move for each possible game state. Because of this, only 1 out of every 5 million possible moves were analyzed.

This is all according to the article in Nature, wherein the researchers announced that they had solved the game of checkers. For those keeping score, the researchers involved speculate that it will be another 50 or 60 years before chess is "solved" in a similar fashion. Who knows how long until there's an unbeatable Go AI.

So, I can still beat the computer at Sorry! right?

@Gannoc: My dream is ruined!!!!!! :*( now i cant beat it at checkers or tic-tac-toe! Suicide my life meaning is gone.

I for one welcome our new computer overlords... in the game of checkers.

Don't hate me, you knew somebody was going to say something inane like this. >:3

I'd cheat.

Call me when it can solve world peace.

@Lorticon: That's easy, Kill All Humans!

I'm beating it hasn't calculated the move where I unplug it and go outside, AND that's why artifical intelligence still lags behind REAL intelligence.

@The-Spectre:

Disagree. That's where artificial INTELLIGENCE still lags behind real STUPIDITY.

I think it'd be really fun to have two computers playing endless, unwinnable games of chess against each other as some kind of modern art exhibition.

@Cursemaster0: Actually, no. Checkers (as we normally play it) has been "solved". If you basically start from the beginning on a standard board, no matter what move you make, the best you can hope for is a draw (which is what happens if you let it play against itself).

Tic-tac-toe also has a solution. If you go first, you can always win or draw. If you go second, you can always draw. It's trivial to calculate in maybe a few minutes these solutions.

That's what the meaning of term "the game is solved" really means - that you can play a legal game (i.e., no cheating or alternate rules), and come up with a matching move that will either guarantee a draw or a victory (the former if both play a perfect game, the latter if the other person makes a mistake).

Checkers, while having easier rules than Chess, actually is harder to solve. It turns out that a victory in checkers relies on getting a tiny advantage (because your opponent makes a mistake). In chess, if you make a mistake, there is still a possibly of a recovery, while in checkers, once you gain the slight advantage, the odds are heavily tilted in that you'll win - i.e., unless you do something completely stupid, there's no way for your opponent to win.

The search space for chess is larger due to the rules, but people are interested in solving that as well (to answer the question - does white always win in a perfect game, or does it always end in a draw?).

Go's search space is much larger than that of chess. The only way to make progress is via AI. But Tic-tac-toe and now checkers doesn't need an AI to be "good".

All your pieces are belong to us.

I'll just go 12 Monkeys on it and start eating the pieces, virtual or not!

If computers and robots can beat the best Checker players in the world, then why not make android Checker players?