On the rapid proliferation of powerful chess software

January 26, 2010 5:35 PM

"It was my luck (perhaps my bad luck) to be the world chess champion during the critical years in which computers challenged, then surpassed, human chess players. [...] What if instead of human versus machine we played as partners? My brainchild saw the light of day in a match in 1998 in León, Spain, and we called it "Advanced Chess." Each player had a PC at hand running the chess software of his choice during the game. The idea was to create the highest level of chess ever played, a synthesis of the best of man and machine." The Chess Master and the Computer: A article/book review on computer chess and the state of the top-level chess world by Garry Kasparov.

Here are some Wikipedia links to Jonathan Shaeffer and his programs Chinook and Polaris. His home page is here.

posted by painquale (43 comments total)

This post gave me an instant geekgasm. Thanks.
posted by Joe Beese at 5:44 PM on January 26, 2010

And what if all of the best chess players played together instead of against one another? What would happen then?!?

Oh, I guess we'd get Chess Life Magazine. Nevermind, then.
posted by kaibutsu at 5:50 PM on January 26, 2010

Call me when computers can beat a serious Go player.
posted by mullingitover at 5:52 PM on January 26, 2010 [3 favorites]

Though you'd think geeks would be more historically aware of the dangers of creating intelligent cyborgs.
posted by Joe Beese at 5:52 PM on January 26, 2010

"Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process."

Fascinating.

And, wow, Kasparov can write.
posted by bpm140 at 6:02 PM on January 26, 2010 [2 favorites]

Kasparov is an interesting guy. I remember reading an interview with him years ago hwere he discussed at length that the whole point of chess isn't calculation, but understanding the mind/ego of your opponent and beating that.

That said ... interesting stuff in his article.

Now if I could just find a woman like Kasparov.
posted by Relay at 6:08 PM on January 26, 2010 [2 favorites]

Call me when computers can beat a serious Go player.

Ring-ring.
posted by mr_roboto at 6:12 PM on January 26, 2010 [1 favorite]

bpm140: "And, wow, Kasparov can write."

Also, brave in the face of danger.

My love of that clip intends no disrespect towards the man's actual physical courage in challenging Russian political authority - both during the Soviet Union and afterwards.
posted by Joe Beese at 6:12 PM on January 26, 2010 [1 favorite]

And, wow, Kasparov can write.

I came here to say this.
posted by empath at 6:16 PM on January 26, 2010 [1 favorite]

I'd really like some more details on how computer + machine chess is better than pure machine and computer chess. Some specific examples of moves or strategies.
posted by empath at 6:33 PM on January 26, 2010 [1 favorite]

So mullingitover, when that does happen (and it will), what happens next? Do we make more excuses for ourselves?
posted by wilful at 6:39 PM on January 26, 2010

wilful: "So mullingitover, when that does happen (and it will), what happens next? Do we make more excuses for ourselves?"

Nope, after that I'll point out that computers are terrible football players. And when that weakness is overcome, I'll point out that computers are terrible at bringing me apple pie ala mode and massaging my feet.
posted by mullingitover at 6:53 PM on January 26, 2010 [10 favorites]

I had a programming 101-type class with Shaeffer way back in the nineties. He's an excellent teacher (I could mentally revisit his lectures after picking programming back up after almost fifteen years off) and a super guy. If he were to get filthy rich helping the pokerwebs, I would feel no begrudgement at all.
posted by Casimir at 6:53 PM on January 26, 2010

Call me when computers can beat a serious Go player.

I should have linked to the University of Alberta GAMES group, of which Jonathan Schaeffer is a member, and which hosts Chinook. You can find Fuego there, which is the best automated Go player right now. Here's an account of a recent series of games between Fuego the top Go players in the world:

At this event, held August 21-22 2009 on Jeju Island, Korea, Fuego played a total of three official games. Two were 9x9 games against the top-rank (9 Dan) professional player Zhou Junxun (his name is also transliterated as Chou Chun- Hsun). Fuego won the first game playing white against Mr. Zhou by 2.5 points. It is the first time that a computer program has won a 9x9 game on even against a top-ranked player, so this represents a milestone for Computer Go. Fuego lost the second game against Mr. Zhou, playing Black. In the final official game, Fuego got an upset win on 19x19 against Mr. Chang Shen-Su, a strong 6-Dan amateur player, with 4 handicap stones.

My favorite thing on that GAMES page is the RoShamBo competition, which I described in a past comment here.
posted by painquale at 7:03 PM on January 26, 2010 [1 favorite]

And yes, Kasparov is a surprisingly masterful writer. I was impressed.
posted by painquale at 7:08 PM on January 26, 2010 [1 favorite]

Wow that is a really fantastic read, thanks for posting it.
posted by vito90 at 7:15 PM on January 26, 2010 [1 favorite]

Like so much else in our technology-rich and innovation-poor modern world, chess computing has fallen prey to incrementalism and the demands of the market.
Hot damn, that sentence.
posted by Paragon at 7:18 PM on January 26, 2010 [1 favorite]

I thought it was interesting for computer science when Kasparov lost to Deep Blue, but it didn't make me think any less of humans. Kasparov was using almost wholly different procedures to play. He was using thinking, reasoning, learning, memory and other capacities in an astoundingly complex way. The match shows that there are other procedures for winning the game. But they're different, and not nearly as complex. (Maybe if they weren't so complex they'd be better understood, hence programmable.) Computers can do lots of things better than people. Chess is different because it required some particularly innovative computer engineering and programming? Good for computer science! Why should this be threatening to people's sense of their own intelligence? Doesn't it matter that Kasparov actually writes better than any computer anywhere ever did?
posted by cogneuro at 7:23 PM on January 26, 2010

(Maybe if they weren't so complex they'd be better understood, hence programmable.)

I meant, maybe if the HUMAN procedures weren't so complex....
posted by cogneuro at 7:24 PM on January 26, 2010

masterful writer

Ho ho ho?

But I don't know why we should be particularly surprised: being a great chess player doesn't mean being a monomaniacal savant.
posted by kenko at 7:25 PM on January 26, 2010

kenko: "being a great chess player doesn't mean being a monomaniacal savant."

It significantly improves the odds.
posted by Joe Beese at 7:27 PM on January 26, 2010

I'm sort of surprised Kasparov hasn't yet been killed after speaking out against Putin's government.
posted by exogenous at 7:29 PM on January 26, 2010 [1 favorite]

I never liked Kasparov. I have changed my mind.
posted by hexatron at 7:33 PM on January 26, 2010 [1 favorite]

That was a truly terrific article. Thanks a lot for posting this; I would not have read it otherwise.
posted by Sticherbeast at 7:42 PM on January 26, 2010 [2 favorites]

I like Garry even though Kasparov lost his match with Putin.
posted by twoleftfeet at 8:40 PM on January 26, 2010

Just to be clear, computer Go programs have a long, long way to go before beating the top human Go players without a handicap, on a full board (9x9 is basically a toy board, mostly used for teaching beginners).

They will most probably win eventually, but it will require a lot more sophistication than the brute-force approach used to win at Chess.

Brute-force programs play the best chess, so why bother with anything else? Why waste time and money experimenting with new and innovative ideas when we already know what works? Such thinking should horrify anyone worthy of the name of scientist, but it seems, tragically, to be the norm.

That this article does not mention Go is a strange omission. Fortunately Go is immune to the approach Kasparov outlines. Though I wonder if one day the statistical shenanigans of computer Go programs will face similar criticisms when they begin to approach, or even beat the best humans.

Ultimately, our desire for machine intelligence will not be quenched until we have an Artificial General Intelligence which just happens to play Go, and can beat the best humans. But by the time that day comes, we'll have far better things to worry about, like Skynet.
posted by MetaMonkey at 9:34 PM on January 26, 2010

I enjoyed this article a lot more than I thought I would. Thanks!
posted by djgh at 10:10 PM on January 26, 2010

I am beginning to have mad respect for Kasparov.
posted by blue shadows at 10:34 PM on January 26, 2010

Kasparov is currently coaching world number one Magnus Carlsen who is one of the new generation of chess players who's learned from computers. What's most interesting about Carlsen is his amazing intuition. Time profiles him here, a Q & A and Jonah Lehrer at the Frontal Cortex writes about Carlsen and intuition in the age of computers.
posted by grounded at 11:29 PM on January 26, 2010 [4 favorites]

Cool, I've always wondered if or when Polaris would get mentioned on MetaFilter. I'm a PhD student on the project.
posted by Fully Completely at 11:44 PM on January 26, 2010

That is a great article
posted by criticalbill at 12:13 AM on January 27, 2010

Interesting.
posted by flippant at 4:06 AM on January 27, 2010

He should move on to go.

Computers suck at go.
posted by Sukiari at 4:44 AM on January 27, 2010

I'm not worried about the future of chess as a pastime. We still paddle boats and run long distances for fun 200 years after the development of steam transportation.
posted by KirkJobSluder at 6:35 AM on January 27, 2010

This would be the perfect disaster movie. Just have an endless succession
of these scenes where Deep Blue is just sitting there. "It's the locusts!" or whatever...and Kasparov is just running his little piggy legs out of the room. "I'm free again, you fucker!"

--Moxy Fruvous
"Kasparov vs. Deep Blue"
Live Noise
posted by Mr. Bad Example at 6:57 AM on January 27, 2010

That really was a great article. Thanks for posting it here.
posted by Wolfdog at 6:58 AM on January 27, 2010

Call me when computers can beat a serious Go player.
Ring-ring.
posted by mr_roboto at 9:12 PM on January 26

Hello? Oh hi! What's that? 2 dan skills? Better luck next decade.
*click*
posted by Theta States at 7:45 AM on January 27, 2010 [1 favorite]

"Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process."

This is why shouldn't worry about a computer taking over the world.

It will be a computer-enhanced human who will take over the world.
posted by eye of newt at 8:04 AM on January 27, 2010

> being a great chess player doesn't mean being a monomaniacal savant.

What Joe Beese said. Nabokov knew all about this stuff; I urge anyone with an interest to read The Defense.
posted by languagehat at 10:22 AM on January 27, 2010

This is the bit that struck me the most about the article:

Even more notable was how the advanced chess experiment continued. In 2005, the online chess-playing site Playchess.com hosted what it called a "freestyle" chess tournament in which anyone could compete in teams with other players or computers.....

Lured by the substantial prize money, several groups of strong grandmasters working with several computers at the same time entered the competition. At first, the results seemed predictable. The teams of human plus machine dominated even the strongest computers....

The surprise came at the conclusion of the event. The winner was revealed to be not a grandmaster with a state-of-the-art PC but a pair of amateur American chess players using three computers at the same time. Their skill at manipulating and "coaching" their computers to look very deeply into positions effectively counteracted the superior chess understanding of their grandmaster opponents and the greater computational power of other participants. Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process.

Just like a top athlete's coach does not have to be as good as the athlete himself at his sport to be a good coach, the amateur players didn't have to be top-level at chess in order to coach their machines to outplay even the grandmaster+computer combinations. Perhaps I'm reading too much into the description, but it sounds like the grandmasters were trying to partner with their computers, attempting to supplement the computer's tactical abilities with their own strategic intuition, and while that was certainly better than computers alone, it was still inferior to amateurs (who know full well they're not as good as the computers or the human grandmasters and don't try to substitute their intuition for the computers' calculations) who knew how to coach their machines.

I wonder what other fields could be improved by understanding that we need only to coach computers, rather than, in our hubris, trying to supplement computer algorithms with our own fallible intuition?
posted by DevilsAdvocate at 11:51 AM on January 27, 2010

He should move on to go.

Which "he" do you mean here? Kasparov? Russkin-Gutman? Shaeffer?
posted by DevilsAdvocate at 11:55 AM on January 27, 2010

brave in the face of danger

As Kasparov demonstrates, the helicockter is not all that intimidating without Ride of the Valkries playing in the background.
posted by minimii at 12:39 PM on January 27, 2010 [1 favorite]

wow kasparov is actually quite a good writer. or has a great editor.
posted by jcruelty at 8:11 PM on January 28, 2010