Chess - 2070 CPUs vs 1 GM

jvarsoke writes "ChessBrain.net broke the world's record for 'largest number of distributed computers used to play a single game' by holding a chess match between Danish GM Peter Heine Nielsen and the equivalent of SETI@home (which similarly, has some people looking for a Mate). 2070 CPU's from 56 countries aided Black by running the chess program Beowulf, including a couple of University clusters. Their supernode ran Linux, and MySQL. The game was relayed by FICS. Results can be viewed here(1) and here(2)."

Comments

I'd love to see a Beowolf cluster of those

[odeee]

I'd love to see a Beowolf cluster of those... Oh damn... it is =:-)

Re:I'd love to see a Beowolf cluster of those

[brad-d]

Oh yeah, that'd be right.

Finally I thought I could get a 5+ funny and here you go and steal my joke. I mean, what are the chances of somebody else thinking of this exact same joke on Slashdot? 1 in 3?

For those too lazy to read the article...

[Gogl]

Or in case it gets Slashdotted or something, I may as well note who actually won the game (although I do think that is something that should have been noted in the submission itself but oh well).

Our World Record attempt is now complete. We had serious technical difficulties early in the game, but managed to resolve them! The result of the game was a draw.

Well that's great....

[filtur]

Sure Chess it great, but can it find me a date?

Here is mirror of the game :)

[doomy]

Nielsen,P - ChessBrain [E94]
Guinness record attempt, 30.01.2004
1.d4 g6 2.c4 Bg7 3.e4 d6 4.Nc3 Nf6 5.Nf3 0-0 6.Be2 e5 7.0-0 a5 8.Re1 exd4 9.Nxd4 Bd7 10.Bg5 Nc6 11.Nxc6 Bxc6 12.f3 Qd7 13.Qd2 Rfe8 14.Rac1 h5 15.Kh1 Nh7 16.Bh6 Bxh6 17.Qxh6 Re5 18.Nd5 Rae8 19.Qd2 b6 20.Bd3 Qd8 21.Rf1 Nf6 22.b3 Bb7 23.Qc2 Nd7 24.f4 R5e6 25.e5 c6

Re:Here is mirror of the game :)

[wan-fu]

You're missing the remainder of the game:

26.f5 gxf5 27.Bxf5 cxd5 28.Bxe6 Rxe6 29.Rxf7 Kxf7 30.Qh7+ Ke8 31.Qxh5+ Ke7 32.Qg5+ Ke8 33.Qh5+ Ke7 34.Qh7+ 1/2-1/2.

What's the point?

[syrion]

The problem with this is that it seems to assume that chess is a difficult problem. It isn't. Modern chess algorithms are really simple search-and- prune systems, relying on the computer's immense number-crunching ability to overcome the more heuristic human mind. Unfortunately, this isn't very interesting. What's the point? We know that computers can search faster than a human. See: Google. All these projects (DeepBlue, Fritz, this) accomplish is trivializing the game of chess, which is rather sad. Now, I'll be really annoyed when Go programs start improving to a 'decent amateur' level...

Re:What's the point?

[vidarh]

Any turn based board game are "really simple search-and-prune" systems. The problem is how to minimize the time taken by the search, and how to decide what a to prune, and how to decide which move to take.

Brute forcing a chess game tree based on basic alpha beta minimax for instance is no way to play well against an experienced human player - first of all you won't get many moves ahead, and a good player that know how the computer work can easily set up a trap that will make the board look good X moves ahead, to make the computer to do stupid moves they can't easily reverse later.

Second you face the problem of definining and weighting what a "good position" is. What is a good position depends on the strategy of the opponent.

Most modern chess programs will augment the basic search and prune with a lot of heuristics to guide the search and weighting of choices, exactly for that reason. They also often contain massive databases of games, sequences of moves etc., to hunt for known strategies that humans might try to recycle against it.

Chess isn't "simple". Chess is a game where it's easy to beat beginners, possible to beat intermediate players on modest hardware, and possible to face grand masters if you have lots of time and access to millions of dollars worth of hardware, and you can still expect to be surprised every now and again.

It makes it interesting, because you have a good foundation to research algorithm improvements on, and because a good algorithm will be more and more useful as hardware costs come down, but it certainly doesn't invalidate the need for better algorithms.

It's also interesting because better algorithms might help us appreciate how humans approach the problem, and as such benefit AI research.

Bullshit...

[Transient0]

It is very rare that a common opener played at the GM level results in a discrepancy greater than about a quarter of a pawn. And it takes a great strategic thinker to understand the advantages and disadvantages of all the available branches in the opening against different types of players.

Of course, it should be obvious that your line of reasoning is totally bogus. The totality of possible moves in chess is simply incomputable and somehow magically trimming this tree to "good" moves still leaves a fundamentally unmemorizable realm of possibilities even at only ten moves depth.

Re:Bullshit...

[DAldredge]

10^120 is the number of possible chess moves. From a google link.

" If you were to fully develop the entire tree for all possible chess moves, the total number of board positions is about 1,000,000,000,000,000,000,000,000,
000,000,000,00 0,000,000,000,000,000,000,000,000,00 0,000,
000,000,000,000,000,000,000,000,000,000,00 0,000,00 0,000,
000,000,000,000, or 10120, give or take a few. That's a very big number. For example, there have only been 1026 nanoseconds since the Big Bang. There are thought to be only 1075 atoms in the entire universe. When you consider that the Milky Way galaxy contains billions of suns, and there are billions of galaxies, you can see that that's a whole lot of atoms. That number is dwarfed by the number of possible chess moves. Chess is a pretty intricate game!"

Re:Bullshit...

[ObviousGuy]

For example, there have only been 1026 nanoseconds since the Big Bang. There are thought to be only 1075 atoms in the entire universe.

The universe must be much smaller than I am prepared to comprehend.

Re:Bullshit...

[troon]

For example, there have only been 1026 nanoseconds since the Big Bang. There are thought to be only 1075 atoms in the entire universe.

Mental note: <sup> doesn't work on /.

Re:Bullshit...

[product byproduct]

Watch your terminology:

• The number of chess moves is at most 218.
  
• The number of chess positions is estimated to be between 10^43 and 10^50.
  
• The number of chess games is infinite, as the 50-move rule and the draw by repetition of position don't apply if no player makes the claim.
  
• The game tree complexity is about 10^123. That's the number of chess games you may have to consider to play perfect chess.
  

Source: http://en2.wikipedia.org/wiki/Chess

Re:Understanding vs. Processing

[njan]

The theorists would disagree with you; computers are extremely good at assessing a *large* number of potential outcomes. Humans, however, are much better at pattern recognition and whilst they can only consciously assess a dozen or two moves, they have most of the work done for them by the functionality in the human brain which causes them to recognise patterns and possibilities far more efficiently than any computer we have now (or will in the forseeable future) will.

Computers can certainly give GM chess players a run for their money - no-one's disputing this; but ultimately, barring a total change of direction in programming/processor/logic/chess theory, they're still just applying what basically boils down to a probability-based brute force method to chess-playing - the human method is far more elegant.

May I suggest...

[kamapuaa]

which similarly, has some people looking for a Mate

May I suggest, that neither the SETI@Home, nor Chessbrain.net, is the best place where one can find a Mate.

Draw game against 2070 CPUs?

[vchoy]

To give credit to Danish GM Peter Heine Nielsen, I would have to say if there were only 2069 CPUs then he might of just won... :P (J/KING)

More interestingly, would the ChessBrain.net team would of won with more CPUs?

Re:Draw game against 2070 CPUs?

[Migrant Programmer]

"might of"

"would of"

Make the hurting stop!

The sad part is you correctly said "would have" earlier in the post.

Yeah yeah, evolving language. Some adaptations should be thrown in the chlorinated pool!

I'm not usually a grammar nazi. But hey, chess is neat. Those fancy chess playing computers are going to take over the world some day, yessirree!

Re:Understanding vs. Processing

[vontrotsky]

We're getting closer and closer to the days when humans won't be able to compete with computer's at chess. Even so I don't think this is such a big deal. We haven't be able to compete with computers at arithmetic for half a century and this doesn't bother anyone.

Losing to computers in chess will be like losing to calculators in a addition match. People and computers aren't really in competition. They do very different things.

Results

[Stalyx]

"The game lasted several hours before resulting in a draw. Chess Grandmaster Peter Nielsen commented that he had set several traps for ChessBrain which computers normally fall for... but was surprised that ChessBrain refused them! "

So what does this tell us? Nothing really, however it would be interesting how the computer will perform in a 5 match series.

Although I still think the GM would win handily.

Re:Results

[azaris]

Why do you believe that over a 5 match series the GrandMaster will win handily?

If you look at the position at move 26, it's obvious ChessBrain is being pressured. In fact the article gives a possible move that could have resulted in ChessBrain losing. Instead Nielsen went for a forced draw because he only cared about not losing to a computer.

If ChessBrain refused some normal traps that computers normally fall for, then could it be the case that the computer is better than you realise. What if the drawn match was a bad one for the computer?

I suspect Nielsen sacrificed the win to see if ChessBrain would fall in his standard tricks, and when it didn't he settled for a draw. With that knowledge he'd probably play the second game much differently, and based on ChessBrain's poor position in the first game, would likely win.

But the fact that ChessBrain didn't fall in those standard traps tells us it's better than most computer opponents.

Obligatory Slashdot Comment

[Crypto Gnome]

It's only a large aggregation, not really a cluster in that sense.

Anyway apparently it worked! (ie not a cluster in that sense either)

If it WAS implemented on the clustering technology we-all-know-and-love as Beowulf, would that make it a Beowulf-Squared?

And, of course, we have to ask the (obvious) question(s)

• Who was 13th?
• And did anyone Eat their Victims?

PS

[Gogl]

It was a draw by repetition. The human grandmaster had a position advantage and was able to force a draw that way despite being down a significant amount of material.

Re:PS

[arvindn]

Not exactly. Nielson had a positional advantage but decided to force a draw anyway by sacrificing material to obtain a draw by repetition. Your version sounds more romantic, but is not accurate :-)

Intangibles...

[John Seminal]

Computers playing chess is not the same thing as two people playing the game.

With two people, there are some elements that can not be programmed into a chess game. I remember in high school playing chess, there was a differance between playing a math academy team and a school best known for its basketball program. Expectations were different, the pressure was different. I remember the pressure of the state finals. There is the look the other person has, almost like poker. Can I bluff this person? Can I trick this person? What about the clock, can I manipulate that to cause an emotion in the other person.

Maybe Spock can play a PC and have no differance in quality of play. But I prefer humans.

Comp. vs. Comp.

[John Hurliman]

I want to see this cluster take on IBM's system!

GM vs. thousands of humans?

[schm00]

Has anyone ever written a system by which a large number of average chess players could collaborate to play a single game? The individuals could vote for the best move, and the majority would rule. Would a group like this be able to beat a high ranking player?

Re:GM vs. thousands of humans?

[sciencewhiz]

There are many systems like this. Chessworld.net is one, and they just challenged chessbrain to a match. You can see a full list of chessworld.net's ongoing games here: http://chessworld.net/chessclubs/event_show_chessw orld_summary_rowgames.asp

Re:GM vs. thousands of humans?

[JuggleGeek]

Would a group like this be able to beat a high ranking player?

I seriously doubt that the group would win. Some of the moves suggested by individuals in the group would likely be the best choice. But more votes would probably come in for another move - one which doesn't hold up as well.

Some time back, I saw an average or slightly above average player play "everyone at the event" by allowing anyone who wanted to make one move in the game. Many people felt this put him at a disadvantage. But it actually gave him a huge advantage. 10 people make make reasonable moves - but all it takes is one guy to make a really stupid move, and now the individual has a big advantage over the "group".

Chess is a game of mistakes. If neither side makes a mistake, draws are very common. That's why when you see games between two GM's, you see a lot of draws. In games where both are "average" players (not serious chess players) then mistakes are common, and generally the guy that made the last mistake is going to lose. (Not always - especially if he already had a demanding lead at the time.)

Things that look very minor - or which are not noticable at all to the average player - are very important to top players. Letting 5,000 average players vote on each move pretty much guarantees that any slightly-above average player would win.

Re:Understanding vs. Processing

[arvindn]

Hey! This is pure FUD.

GMs don't even play to mate anymore

Only rank beginners (say less than a couple months into chess) ever play to mate. Its obvious who's going to win long before mate happens. To continue playing is a waste of both players' time, not to mention an insult to the opponent's intelligence.

they just play out an opening move .

I don't even know what this is supposed to mean. Grandmasters do an enormous amount of research into finding new moves in openings. They don't "memorize" them. There are five volumes of the ECO chess encyclopedia, and that just covers the basics!

and whoever has the upper hand at the end takes the game

No of course they don't. This is simply false, period. Why do you think there are things called "middlegame" and "endgame"??

Its sad that because most moderators aren't chess players, anyone can write ridiculous BS and get modded up "+5, interesting".

Losing to Computers

[rynthetyn]

It's gotten to the point that even Kasparov is only playing the best chess computers to draws. Of course, he did lose to Deep Blue, but despite all his insistance that IBM cheated, he got beat mentally, not necessarily because the computer was better.

Incidentally, there is a new documentary, Game Over: Kasparov and the Machine about the Deep Blue rematch, which I had the opportunity to see at the US premier a few weekends back. I'd link to the review I wrote on my blog, but I don't think the sysadmin would be very happy with me if I did.

Re:Understanding vs. Processing

[prockcore]

Losing to computers in chess will be like losing to calculators in a addition match. People and computers aren't really in competition. They do very different things.

Damn straight. A computer may be able to beat me at chess, but at least I can visually identify a chess set in a crowded room.

2070 CPUs

[Duncan3]

1 CPU to beat the GM.
+2069 CPU's so it could get on Slashdot.

There are very few humans on the planet that can beat even one computer. That's been true for how many years now? Neither beating a GM or 2070 CPU's is impressive anymore.

Someone go built a robot that can shovel snow, now THAT would be useful.

Re:Understanding vs. Processing

[spun]

You do know FUD means "Fear, Uncertainty, and Doubt", right? I think the acronym you are looking for is "BS."

Am I doing my maths correctly?

[Knx]

There are approximately 35 moves per position in Chess (average value). Thus, the branching factor of the search tree is ~35 with a simple min-max search. Assuming that the program is always picking the best move to search first -- which is obviously not systematically the case -- alpha-beta pruning allows us to get a branching factor equal to approximately the square root of 35, that is: close to 6.

Assuming that 2070 CPU are able to do the calculations 2070 times faster than 1 CPU -- which, again, is not the case -- it appears that the resulting supernode is able to 'see' up to 4 or 5 half-moves deeper than a single CPU in the same amount of time:

6^4 < 2070 < 6^5

It doesn't seem to be *that* useful. For most strategical positions, thinking 5 half-moves deeper just doesn't make any difference. Game 3 of 'Kasparov vs X3D Fritz' is a good example: I'd be willing to bet that 2070 X3D Fritz playing together would have lost the game the same way, since the serious troubles caused by the pawns diagonal are still far beyond the resulting analysis depth. (Well... At least, I think so. I'm not a Chess expert!)

Anyway, this is quite an interesting project. I hope to see it grow up in the future.

-- Arnauld