Hacker News new | past | comments | ask | show | jobs | submit login
In 1999, Kasparov played chess against 50000 people. (wikipedia.org)
33 points by shadytrees on Dec 10, 2008 | hide | past | favorite | 25 comments



To clarify the title, all 50000 people were working as a team in a single game of chess.

I originally thought this would be a pointless exercise, sort of like 50000 monkeys banging away at typewriters. It appears that was not the case. I found this passage about how one of the players effectively lead the team quite fascinating:

"15...Ra8

The results of the vote were a reflection of the increasing coordination of the World Team. Krush was maintaining an analysis tree, and continually updating with all the suggestions and refutations from the bulletin board. Not only did the analysis tree allow the World Team to work with less duplication of effort, it served as a standing, detailed argument for the correctness of the recommended move."

It makes sense that some kind of unified document to keep track of everything would help. It is fascinating to me how much like the data structure you would probably see inside a computer chess program this document resembled.

Next, in the commentary for move 16:

"This sparked loud grumbling on the bulletin board that Krush had "taken over the game". Those who complained were not overstating Krush's influence; her recommendations were selected every single move from the 10th to the 50th."

This was really surprising to me, how much influence her recommendations had after that point. I wonder how much of it was: first mover advantage vs form of coordination vs having the best ideas.


One thing which it fails to state is that this was a major event for casual / hobbyist chess players. I don't have any statistics but every friend of mine who played chess was taking part in this.

So I suspect that the majority of those who voted moves regularly was a majority of people who understood the game and could follow the logic.

I don't think it can be overstated the sort of impact recommended moves had. This was largely a game between each GM and Kasparov with people voting on who to follow. As I said in another comment, had there been no recommendations, the game would have probably been much shorter and much more juvenile in complexity.


Sadly, I think that article needs to be rewritten to conform to Wiki standards of studied, patient mediocrity in writing. :(

>> The black queen forked Kasparov's central pawns; he could not save them both. Counterattacking with 11.Qb3 would have been met by 11...0-0 12.Qxb7 Rfc8, and the World Team would have won back a central pawn with a favorable game. Kasparov was forced to enter the maelstrom with the next several moves. >>

Great story but it has POV and citation issues. Here:

>> The black queen forked Kasparov's central pawns. Chess experts have noted[123] that, hypothetically assuming Kasparov had played 11.Qb3, that 11...0-0 12.Qxb7 Rfc8 was a possible outcome. Some experts[124] consider this gain of a pawn to give strategic advantage to the World Team, others[125] disagree. >>


Yeah, sometimes I wish wikipedians weren't as strict on those rules....


Why not make the change on Wikipedia?


I used to be an English teacher so forgive me for two things:

1) I think that my suggested edit, and the other few dozen similar ones I could make, do not improve the article even if they are more in line with the Wikipedian NPOV and NOR policies.

2) The words "sadly" and "mediocrity" are cues that would have suggested #1 already.


I wasn't suggesting you should have. I just wondered what your reasoning was. Thanks for explaining.


I remember this game fondly. Sadly, the game has a serious flaw. I don't remember the exact move, but Irina Krush was late on posting her move, the best one for the position, and as a result, it did not get nearly as many votes as it should have. Had the move been played, the world would have had a much better position.


reposting a comment I saw elsewhere, google it for more

--

Let me, as a chess player, say that I was much less surprised by the resistance the World put up against Kasparov:

A) Kasparov is indeed considered the greatest player by most and the reason for his strength when playing ‘over the board’ (OTB, as it is termed) was threefold: 1) he prepared more and better than anyone else 2) his calculating powers OTB were quite magnificent and 3) his understanding of dynamic chess was ahead of his time. Now when using the format of the Kasparov vs the World game these advantages largely (not completely) disappear: 1) everyone can consult books (in particular about openings), 2) everyone can move the pieces around on the board instead of inside their heads 3) the game was not a messy game Kasparov used to excel in but quickly went into a middle game with endgame features.

B) Kasparov may be “much better” than other GMs and than Krush in particular, but this difference in strength should be qualified. The difference between me and most tournament players in the US player is much larger than that between Kasparov and Krush! The median rating of US chess players [those who play tournaments and actually do have an official rating] is somewhere around 1000-1100, mine is more than 1100 points higher. Kasparov is “only” about 300-400 points above Krush (and like 150 points above Khalifman).

C) There is a good comparison with correspondence chess here: this is where people play by email, for instance, (it used to be by snail mail). I just started playing this myself and the tempo in most games is about 1 move per day. Now I play several people whose OTB ratings are way below mine, indeed a 1000 points below mine. But I don’t have an easy time beating those players in correspondence chess. A difference of 1000 points implies I can beat them, when playing OTB, literally blindfolded (just seeing the board in my head). But in correspondence chess it’s much much harder. They can use opening books and look up how grandmasters have played in the position we happen to have on the board.


This doesn't change the argument, but the median rating of 1100 is misleading since most of under 1100 are children who participate in various Chess in Schools programs and who never go onto play rated chess as adults. Among active playing adults the median must be closer to 1400 if not more.

However this Kasparov vs. The World game does show the case where the "tragedy of commons" comes into play. 50,000 players playing without "guidance" from the top would quickly lose to Kasparov.

One last point, 400 point rating difference is actually considered quite significant, top player then is supposed to get a 99:1 score in 100 games. When there is 200 point difference, advantage slips to 75:25.


400 points is actually more like a 90:10 advantage, but that's a minor point.


You are correct! It is the 800 point difference which comes out to roughly 99:1. Finally, took the time to calculate correctly from: http://en.wikipedia.org/wiki/Elo_rating_system


I think also that it greatly influenced the play by having the four GMs make recommendations. Had they (MSN) instead let players choose any move on the board, then the game would have been much shorter and much less interesting.



I remember participating in that game -- I was in New York when Kasparov opened the game by playing 1. e4, I voted on a lot of the moves, posted my analysis (1350 USCF analysis, but still) to the forum there, and was greatly disappointed when the World realized it had blown a drawn endgame. This brings back the memories, and it was indeed a great game of chess.


A Spanish tv channel did something similar (viewers voting on a few options against Kasparov) and on that occasion 'Spain' won. Edit: I remembered it wrong; I looked it up and it was 1991 with a win for Kasparov and a tie (I guess the tie felt like a victory for the viewers).


One summer, I interned at a think-tank where there were a lot of Go players. One of them invited a Go master (7+ dan) to play 15 separate games against 15 people. He never delayed: he would take his turns in each of the games during the amount of time that his opponents would use to take one turn. He won all of the games, but none by anything that would seem like a blowout margin.

After he left, people began discussing and comparing scores. He had beaten every single one of them by exactly 13 points.


Its sort of like playing horse with Michael Jordan. (My Dad has -- they happened to go to the same health club in Chicago. He tells the story much better.)

You may think you are playing horse. Michael Jordan is playing another game entirely. You probably don't understand the rules but don't worry, you're not winning in that one either.


Funny you mention it, a reasonably normal person beat Jordan one-on-one:

http://www.youtube.com/watch?v=5B7U74Dg04k

And a reasonably normal person also beat LeBron in horse too:

http://www.neatorama.com/2008/09/18/video-of-lebron-james-lo...


What game is he playing?

I've heard that, among elite Go players, it's most prestigious to beat an esteemed opponent by a small margin, such that a 1-point win is considered ideal.


i think that could be a separate issue from the tactics of the actual game. if an opponent blunders such that you have 2 moves that seal the game for you, one which gives a large point victory and one which gives a small one, it is considered polite to take the small one as a show of respect for your opponents skills. that's the way I've heard it at least. Go etiquette reflects japanese culture and not having grown up in it I'm sure none of us truly appreciate all the subtleties.


That's incredible. Go has always fascinated me on principle, though I have yet to find the time to learn the game. Thanks for this.


If you have 5 minutes, you can learn the rules from http://homepages.cwi.nl/~tromp/go.html -- a much neater explanation than most. It takes a little practice to get good, of course. Just a few years of full-time playing.

The best general resource online is http://senseis.xmp.net/ .


Cool video of game of Go on Microsoft's "Surface", played against an HPC cluster: http://www.youtube.com/watch?v=Qe0o-IvHOa0


As celoyd said, it only takes a few minutes to learn the handful of rules. It's just a matter of finding time to play.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: