Slashdot Mirror
16th World Computer Chess Championship In Progress
vmartell writes "The 16th World Computer Chess Championship is now in progress in Beijing, as part of the Computer Games Championship. Currently in the lead are Rybka 3.0, recognized as the world's strongest chess engine and Hiarcs, another commercial engine. Another curiosity is a Java ME based engine running on a Nokia phone, which is currently being trounced by the other engines. A very interesting sideline: before the computer tournament, a Women's Grandmaster played two games against Rybka. The result? Rybka won both games!"
Dear Slashdot Admin,
What did you just do to tags? I previously ignored them, and now I am ignoring them at two or three times the size they were a few hours ago.
To add insult to injury, I just disabled "display tags" in my user preferences and they haven't gone away! What the hell?!
-Dave
How much of the "skill" in computerized chess comes from the programmers and how much comes from the raw cpu horsepower available? TFA was quoting 40-core boxes competing with Nokia cell phones.
Transliterated back into Russian, Rybka means Fish. Maybe I don't know the joke, but I've never known fish to be particularly strong chess players.
Sam ty sig.
Does chess really need to separate the rankings between male and female champions? Isn't this a sport that gender really doesn't factor in?
Chess has become boring, like checkers or backgammon.
To even competitively play at the local club level you really need a ridiculously deep memorization of openings and endings. At the grandmaster level, they've basically memorized the tables used by computers.
Average games of chess only last around 60 moves. The depth of opening and closing books increasingly has reduced the middle game of actually interesting play. If it's not down to only 1-5 moves, it will be soon.
The game will be dead--or at least not interesting enough to be seriously played--long before it is solved.
P.S. You arrogant fans of Go can frak yourselves. Where do you think the scientists will go once they're done with chess. Enjoy it while it lasts.
obviously no deficiencies vs. no obvious deficiencies
http://en.wikipedia.org/wiki/Rybka
The page also goes into detail on the name...
No sig today...
It's the same in Russian. I know it from Russian so I posted about the Russian.
Sam ty sig.
"P.S. You arrogant fans of Go can frak yourselves. Where do you think the scientists will go once they're done with chess. Enjoy it while it lasts."
Considering Go's harder. I'd say they're welcome to try.
Shai Schticks:"You don't make peace with friends, you make peace with enemies"
"So in other words, it's not the the number of cycles you have, but how you use them."
That's what she said.
Shai Schticks:"You don't make peace with friends, you make peace with enemies"
When I think of powerful chess programs, I think of Deep Fritz. It did beat the human World Champion after all. Does anyone know why Deep Fritz isn't competing?
Queen to Bishop 6. Check.
Today in Beijing is also starting the first World Mind Sport Games, a series of team and individual tournaments of chess, go, bridge, draughts and Xiangqi (chinese chess).
I'm actually curious. The engine used by the free chess program supplied with OS X has no trouble whipping me and I'm playing it at one of the easier levels. Fair enough I'm not a great player but I'm not a moron either. My guess is that a significant portion of those that enjoy a game of chess would find many of the free offerings sufficiently challenging.
I understand the desire to create better and more efficient algorithms but I'll be damned if I can see their commercial use.
I thought it is accepted that Hydra Scylla is by far the best. It is estimated to have a strength of over 3000+ ELO; it haven't been rated since it has never lost a competitive game...
http://en.wikipedia.org/wiki/Hydra_(chess)
Have you looked into Chess960 (Fischer Random Chess)? You basically randomize the back rank of pieces into one of 960 different configurations before starting. It makes opening books useless and the game becomes more about skill than memorization. Of course, you'll probably still get slaughtered by a GM, but that's more to do with skill than rote memorization.
Tabloid headline: Chinky chess chick checked?
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
.. ther may well be a market for the specialised hardware and what batter way than for example IBM to advertise their hardware & software abilities as a company to the triumph at chess so the marketing paybacks are enormous , far bigger than slapping your logo on the front of a football player or F1 car.
...play a nice game of Global Thermonuclear War?
After having dealt the last seven years with the anime-obsessed crowd invading Go, and (at least in my limited experience) being nothing but whiny about how much there is to learn, I'd say that engineers, chess players, and heads of state (though to be fair, in this case, he was merely Prime Minister and was assassinated trying to avoid the militarization of Japan pre-WWII) would be welcome back into the fold.
On that note, I wouldn't say I'm an arrogant Go player. I've played chess at the local club level for awhile, but then I realized that I'd rather enjoy a game of chess with a friend over a beer out at the park than several hours of study to keep my opening and endgame up. I picked up Go about a year after that, and while I'm not very good, I enjoy it and enjoy teaching it to others.
A couple of years ago the best chess computer was IBM's Deep Blue, a massive custom machine with terabytes of innards.
Now it's a standard desktop PC worth a couple of thousand dollars.
Brute force can compensate for lack of insight but insight advances steadily.
No sig today...
...and preloaded winning moves can compensate for either
Most of the learning that chess players do is learning known good move sequences .. chess computers are just getting preloaded with more of these - this is why a handheld computer can even think about competing
Puteulanus fenestra mortis
> Sweeping generalizations are always wrong. :)
Sweeping generalizations are always wrong.
There's something ironic about how you worded that. I can't quite put my finger on it.
Because if Fritz stays out than people like you will assume it still might be world champion when it isn't even close any more. Well, and me since I bought a copy.
Rybka would tear it to shreds since all the Rybka programmers are all grandmasters who fill Rybka to the brim with chess ideas.
You'll notice this isn't the first competition Fritz stayed out of but all the other engines have stayed in. Fritz has stayed out since Rybka showed up IIRC.
/.is against patents.
Shut up already. Why don't you go look up the definition of the no-limit fallacy, and then get back to us?
I had a similar experience in 1978: I had written a neat but slow chess program in Basic for the Apple II. The guy who organized the first West Coast Computer Fair chess tournament encouraged me to enter, and I did not fair so well against the programs nicely crafted in assembler language. Still, since I was handing out free copies of my program, and people liked that, it was a fun experience.
I actually did a little chess programming last week. I am finishing up the 3rd edition of my Java AI book (self plug: a PDF version will always be available on my web site) and since I have generally "caught the Java generics" disease, I re-coded the chess alpha-beta search example using the new collection classes and generics - it ran so much slower than the old native data type + array version that I archived and tossed the new version :-)
I'm on a chess binge lately and have to say that the youtube repository of videos about Bobby Fischer and Gary Kasparov is just fascinating. Check out Fischer on the Dick Cavett show to see the how earnest and human Fischer was, for example. The game of chess itself is a formal system with a finite vector space that machines can eventually brute force, but what is interesting is how humans and their neural processing can match the machine approach. The people who play chess well, like Fischer and Kasparov and others, are much more interesting than any algorithm will ever be.
If anyone is interested, I implemented a java applet chess game a long time ago, and the source has been open for some time now. Check it out at http://www.wanfear.com/~mbrito/games/twoplayer/chess.html If anyone is interested in enhancing / commercializing it, I'd be happy to accomidate.
Deep Blue was capable of 11.38 gigaflops in 1997.
You can go to the store and buy an off-the-shelf PC today which will do 2 teraflops.
The fact that a standard desktop PC can best Deep Blue has little to do with "insight" and everything to do with Moore's Law.
If you mod me Overrated, you are admitting that you have no penis.
P.S. You arrogant fans of Go can frak yourselves. Where do you think the scientists will go once they're done with chess. Enjoy it while it lasts.
This is obviously trolling, but what the heck. Chess has been dominated by computers for quite some time now - many have moved on to Go, and they still fail to beat amateur-level Go players. There are practical reasons why a Go AI is more difficult to program than a Chess one, which I'm sure by the time this has been posted will be explained in great detail by other replies.
What I want to explain is why Go is better than Chess. It is not because it is more difficult for computers. While Chess doesn't really scale well, Go can be scaled down quite nicely to where the number of moves that have to be read out is equivalent to that of a Chess game. On 9x9 boards, Go AIs are rapidly catching up to humans. MoGo, for instance, consistently beats many amateur-level go players on 9x9's. There is more lost in the transition to a smaller board than simply brute-force reading, but the example still stands.
The reason Go is better than Chess is that unlike Chess much of it's depth is a natural, mathematical result of the very simple rule set. Essentially the entire game:
Players take turns placing pieces on the board. When a piece or group of pieces by one team is surrounded (ie, no empty spaces touching any of them in the group), the pieces are removed from the board. Whoever controls the majority of the board at the end wins.
There are details missing there about things such as defining who controls the majority of the board, but with one exception (ko) that's really all the game is. Everything, all of it's depth, is a mathematical result. Interesting patterns emerge, comparable to things like prime numbers or pi. For example: From the rules I put above it should quickly become apparent that I could just surround the opponents pieces that surrounded mine, and my opponent could do the same. No piece would ever really be safe, or ever really be captured that could not be replaced. No rules had to be made to remedy this: go naturally has a system in place where pieces cannot be captured. It's just plain cool. (See: "eyes").
For the most part, Go is not about brute-force reading so much as recognizing and understanding the patterns. I don't have to mathematically derive the formula for the area of a circle each time I need to calculate it - once I understand the concept I can retain it. Go is very similar. That's not to say there is no brute-force reading. For example: one of the interesting phenomenon in Go is the "ladder," which results in one player chasing the other across the board diagonally. The winner will be determined based on whose pieces (if any) are met along the way. As a result, people just read out the very simple pattern across the board rather than playing the whole thing out.
Go isn't flawless. There is a situation (ko) which essentially breaks the game. It's comparable to where in Chess if both players repeatedly make the same moves the game becomes a draw. There had to be rules added to deal with this situation, and while they do add even more depth and change the way the game is played, it is an area where Go's awesome depth-through-simplicity is marred.
I recognize that Chess does have some depth to it other than just brute-force reading. There are concepts like pinning which result from the rules, but they're far fewer and generally not as interesting. There's also many concepts which carry between the two, such as Go's sente/gote vs Chess's tempo.
Go is less about reading than it is about understanding. This, combine with the cool patters and concepts which emerge, is why Go is better than Chess. This is largely opinion; someone who plays Go is not necessarily a better, smarter person than someone who plays Chess. It is possible for an intelligent, healthy, mentally-stable human to prefer Chess's convoluted rules, simpler concepts and overwhelming amounts of brute-force reading. They'd just be silly for it.
"A witty saying proves nothing." - Voltaire
Are there any challenges out there that limit what sort of computing power you may have? Sort of like the Loebner prize, but for chess. I think algorithms would get much more interesting if every contestant had to limit themselves to some standard configuration like a dual core 5000+ with 2 gigs of ram and 64 gigs of SSD. Granted, that hardware is going to be long in the tooth in 2015, but the algorithm that wins in 2015 will probably give mr 40 core supecomputer a run for its money.
I'm rooting for the Turk. Keep him happy winning chess tournaments and maybe he won't get all discourage and turn into Skynet instead.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
The difference between men and women is even more drastic in Shogi (Japanese chess).
In order to qualify to become a professional, first you have to enter the Shoreikai and climb all the way to 3-dan (equivalent is amateur 6-dan) before entering a round robin tournament exam against all of the other 3-dan players. The top two get crowned as 4-dan and get to become professionals.
No woman has ever advanced past Shoreikai 1-kyu (amateur 4 dan). Hence, the shogi association created a separate womens league. The titleholders of the womens league are even seeded in the regular championship league, but they almost always lose horribly, and they can almost never beat even the newly crowned 4-dan pros (they have something like a 20% win percentage against male professionals as a whole).
Didn't mean to post this as AC:
The difference between men and women is even more drastic in Shogi (Japanese chess).
In order to qualify to become a professional, first you have to enter the Shoreikai and climb all the way to 3-dan (equivalent is amateur 6-dan) before entering a round robin tournament exam against all of the other 3-dan players. The top two get crowned as 4-dan and get to become professionals.
No woman has ever advanced past Shoreikai 1-kyu (amateur 4 dan). Hence, the shogi association created a separate womens league. The titleholders of the womens league are even seeded in the regular championship league, but they almost always lose horribly, and they can almost never beat even the newly crowned 4-dan pros (they have something like a 20% win percentage against male professionals as a whole).
...a Women's Grandmaster played two games against Rybka...
There's a women's league for chess ??
the number 500 supercomputer has an Rmax of 9 teraflops; it wasn't long ago (~4 years) when half a teraflop would get you on this list.
well, unless you're talking about GPUs, which don't really have anything to do with generalized computing tasks, and those numbers are dubious, anyway. it should also be noted that, apparently, the fastest PC processors are up to the 50 gigaflops range now.
I was talking about GPUs. Of course the numbers don't completely match up, but the point is that last decade's ultra-specialized supercomputer is this decade's beige-box PC. Less, even.
Additionally, with CUDA and the like now available, I don't see any reason why a GPU couldn't be used for a chess engine. The kind of deep game-tree searching they use is pretty much inherently parallel. No doubt it would take a lot of work (and would probably not be worthwhile compared to improving the existing algorithms on normal CPUs) but I don't see why it couldn't be done.
If you mod me Overrated, you are admitting that you have no penis.