The Consensus of Crowds

September 18, 2005 5:52 PM

Consensus View
New Yorker columnist James Surowiecki's book The Wisdom of Crowds "explores a deceptively simple idea that has profound implications: large groups of people are smarter than an elite few, no matter how brilliant—better at solving problems, fostering innovation, coming to wise decisions, even predicting the future." Now this idea has been put into practice with Consensus View, a site where you can enter your predictions on stocks, commodities, and currencies, and view the group consensus. (from wsj.com)
posted by reverendX (46 comments total)

It's like Wikipedia for guessers!
posted by Mach5 at 6:05 PM on September 18, 2005

For the sake of democracy this idea must be true, but I must admit I have developed certain doubts in the "wisdom of crowds". Wouldn't that suggest that mobs riots and pogroms are Good Ideas? Not sure I'm ready to sign on with that.

In other words: how's that president owrking out for you?
posted by freebird at 6:14 PM on September 18, 2005

But aren't speculation based markets famous for being driven by the whim of the crowd? Isn't that all Consensus View is? How does this prove greater wisdom?
posted by PinkStainlessTail at 6:16 PM on September 18, 2005

Sir Bedevere: What makes you think she's a witch?
Peasant 3: Well, she turned me into a newt.
Sir Bedevere: A newt?
Peasant 3: ...I got better.
Crowd: [shouts] Burn her anyway!
posted by mullingitover at 6:24 PM on September 18, 2005

I haven't read the book yet, but I seem to recall from some of his columns that crowds aren't better at everything, but rather certain sorts of problems. The one that stuck in my head was the location of a sunken submarine; it had been lost for decades, so (the proverbial) they brought together a large group of people, showed them maps of where the submarine had last been seen, where it was going, etc. Each person then guessed where it was; the answers were aggregated, and it turned out that the submarine was almost precisely at that point. Apparently this sort of thing happens enough to make it significant.
posted by Coherence Panda at 6:27 PM on September 18, 2005

Brian: You are all individuals.
Crowd: We are all individuals!
Man in Crowd: I'm not! (gasps from crowd)

The tension between common cause and the tragedy of the commons is unresolvable except through education. Even then, there will be outliers who challenge us all to define the center, the commonweal. Lately, it's been tilting toward tragedy on a grand scale.
posted by realcountrymusic at 6:28 PM on September 18, 2005

Einstein vs the crowd at Yankee stadium .... you decide.
posted by R. Mutt at 6:29 PM on September 18, 2005

Coherence Panda:

I think that was the USS Scorpion.
posted by JB71 at 6:37 PM on September 18, 2005

Oh this is nonsense. What's the mechanism supposed to be?
Of course the small group of people who really know what's going on are going to be better at solving a problem than leaving it to some aggregate opinion.
posted by snoktruix at 6:39 PM on September 18, 2005

crowds aren't better at everything, but rather certain sorts of problems.

...for instance, crowds are sometimes pretty good at predicting what the crowd is likely to do next.
posted by sfenders at 6:43 PM on September 18, 2005

There's already a contradiction on the front page: If what matters is the "consensus view" why do they have a list of "Top Forecasters" - individuals? Thats hugely misleading.

I mean, if I had 100 monkeys making predictions with dice, one of them would inevitably be a "brilliant forecaster" just because someone has to be at the end of the curve. Put another way, its like calling the winner of a lottery "a top lottery forecaster"
posted by vacapinta at 6:44 PM on September 18, 2005

Anyone want to take a guess where the next terror attack is going to be?
posted by Hobbacocka at 6:53 PM on September 18, 2005

Anyone want to take a guess where the next terror attack is going to be?

I dunno, in the ghost train ride when the zombie jumps out?

Terror...I don't think that work means what you think it does. Terrorism, however, might be more appropriate...
posted by Jimbob at 6:57 PM on September 18, 2005

vacapinta: it does seem like a contradiction, however I suspect the 'top forecaster' feature is an incentive for people to submit their predictions and thus give the system more data to work with.
posted by reverendX at 7:01 PM on September 18, 2005

Terror...I don't think that work means what you think it does. Terrorism, however, might be more appropriate...

Right you are, Jimbob!
posted by Hobbacocka at 7:02 PM on September 18, 2005

I'm as mad as hell, and I'm not going to take this anymore!
posted by kingfisher, his musclebound cat at 7:23 PM on September 18, 2005

R. Mutt : In a fair fight, Einstein's pretty comprehensively outnumbered. No way he could take 'em, not even with a flamethrower.
posted by kaemaril at 7:35 PM on September 18, 2005

Does this not mean that moderate takes on factual issues tend to be more correct, as aggregates water down the extreme views?
posted by dreamsign at 7:41 PM on September 18, 2005

Does this not mean that moderate takes on factual issues tend to be more correct, as aggregates water down the extreme views?

I think so, yeah. Kind of like in signal processing, where you use noise to cancel out noise so you can see the signal.
posted by vacapinta at 7:52 PM on September 18, 2005

As a counter-example, 50,000 were unable to outsmart 1. (They did make a pretty good showing, though.)
posted by justkevin at 8:02 PM on September 18, 2005

MetaFilter: using noise to cancel out noise

(but freebird already won, for sure.)
posted by blacklite at 8:05 PM on September 18, 2005

Without even looking at this WSJ Consensus View I think it contains a fundamental flaw. How people actually act when they invest is different from how they say they act. What we'll get here is a consensus view of views, not a consensus view of behavior.

Also, deep down most people share the same fundamental superstition, call it Evil Eye, Knock on Wood, or what you will: If you predict something good, you prevent it from happening. So we can expect this WSJ experiment to be biased toward the downside.
posted by mono blanco at 8:23 PM on September 18, 2005

Mayor Quimby: "You people are nothing but a pack of fickle mush-heads."

Springfieldian #1: "Hey, he's right!

Springfieldian #2: "Give us hell, Quimby!"

Crowd: "Yaayyyy!"
posted by uncanny hengeman at 8:55 PM on September 18, 2005

The main flaw is that the aggregate effects of the crowd, are enjoyed by the crowd as the crowed, not as individuals. It's the elite few who are the drivers. The consensus just represents the ability to know a good thing when you see it, but not the ability to come up with it in the first place.
posted by Gyan at 9:06 PM on September 18, 2005

I assume I'm not the only one who thinks about John Brunner's Shockwave Rider every time something like this comes up....
posted by dersins at 9:09 PM on September 18, 2005

The Iowa Electronic Markets have been pretty interestingly close over the years -- for instance, they pegged Bush in '04 when the media was generally choosing either "too close to call" or Kerry by a nose.
posted by dhartung at 9:14 PM on September 18, 2005

This "wisdom" tends to converge on correct prediction as the time horizon of prediction grows short. In other words, the "wisdom" grows in proportion to the surety of the bet.

I agree with Freebird : The "Wisdom of Crowds" theory is directly contradicted by boom/bust cycles in animal populations, the stock market, and countless other mass cycles.
posted by troutfishing at 9:34 PM on September 18, 2005

Well, "wisdom" is being used rhetorically, it doesnt mean crowds are actually wise in the literal sense (if such a thing exists).
posted by stbalbach at 10:58 PM on September 18, 2005

Aren't prices for stocks, futures and currencies *already* concensus opinions on their future values?
posted by ManInSuit at 11:04 PM on September 18, 2005

Aren't prices for stocks, futures and currencies *already* concensus opinions on their future values?

alright now, in caps, NEITHER THIS NOR 'MARKET' DECISION-MAKING IS A PROCESS OF CONSENSUS.

(with a flowchart, for the picture lovers)

These are merely the aggregate statistics of a herd of isolated, atomised individuals, and not a process of creative call-and-response that yields to no compromise. This is moronic majoritarianism. This squelches new ideas, that by definition start small.

there is no feedback among the individuals, leading to a deep, analysis-based, or creative solution, no cultivation of any diversity of ideas, no taking advantage of the collective power of argument and logic--nothing but mindless herd behavior that, if followed, will drain your very dignity and the dignity of those around you. stop it already, you know better.

questions?
posted by eustatic at 11:37 PM on September 18, 2005

More specifically, the financial markets reflect the wishes of those with means. if someone makes 600 times your wage, that person has 600 times the influence on an aggregate statistic, and is 600 times the human being you are, if you really believe that markets are 'democratic.' it used to be, in the United States, that we were at least 2/3 the human being our masters were. 1/600th?

The "Corporate Crime Busters" van is touring locally to draw attention to inequitable pay scales, said Nader campaigner Jim Senyszyn. Utilizing a "Big Mac Index," Senyszyn and fellow campaigner Patrick O'Brian told workers that while they could purchase two or three Big Macs on an hour's salary, company CEO Charlie Bell can purchase 690 of the sandwiches on an hour's pay, about $1,725.

The injustice of markets is aggrevated as the gap between the poor and the wealthy increases, and there is nothing in the market system to correct this fatal allocation flaw.

ok, done with the pet peeve for tonight, i promise.
posted by eustatic at 11:57 PM on September 18, 2005

"Give us hell, eustatic!"
posted by uncanny hengeman at 12:04 AM on September 19, 2005

look, i was the first one to use the word aggregrate.
posted by snoktruix at 12:56 AM on September 19, 2005

Charles Mackay argued the opposite of Surowieki's thesis, in 1841, in a book called Extraordinary Popular Delusions and the Madness of Crowds. He analyses more than a dozen examples (the South Sea Bubble, witch-hunts, the tulip frenzy) when mass insanity provoked an economic or social disaster. It's Mackay, rather than Surowieki, who gives a clear indication of how the dotcom boom happened.
posted by MinPin at 1:55 AM on September 19, 2005

large groups of people are smarter than an elite few, no matter how brilliant

The precise opposite is true, I believe. But then, much depends on how you define slippery words like 'smarter' or 'wisdom'.
posted by stavrosthewonderchicken at 4:31 AM on September 19, 2005

Isn't this just the Delphi Effect?
posted by sandking at 5:39 AM on September 19, 2005

So life echos William Gibson? Neuromancer featured an "oracle" that was an aggregate of human decisions on subjects...

Personally I've always felt that the intelligence of a crowd is directly inversely proportional to the size of the crowd. The lowest common denominator theory - in the simplest (and least politically correct) terms, that the average intelligence of a crowd is brought down to the level of the least intelligent people in the group, which tends to rule the activity of the group as a whole. Larger crowds are less intelligent as they are more likely to include a greater number of dumb people. I'm not sure how scientific the idea is, but I'm still pretty convinced that the theory accounts for mob behavior and other related idiocy that doesn't happen in smaller groups.
posted by caution live frogs at 6:36 AM on September 19, 2005

I think this pretty much sums it up: ...for instance, crowds are sometimes pretty good at predicting what the crowd is likely to do next.

And: What stavros said: What [the HELL] does "wisdom" mean, anyway? (For that matter, what's a "crowd"?)

And how do you get to get your decisions into the aggregate? And at what point does knowledge of the aggregated decision(s) begin to affect the outcome? These aren't untestable questions; but people tend to be so psychically (emotionally, intellectually) invested that they can't formulate objective tests, and don't understand the results when they get them.

In the final analysis, it's probably true that both Surowiecki and Mackay are right: "Crowds" are both "wise" and "stupid"; they can lead us to success (as defined by someone) or disaster. Which we focus on in the moment probably has a hell of a lot more to do with what we're afraid of (as a "crowd") than in any real evidence for one view or another.
posted by lodurr at 7:28 AM on September 19, 2005

BTW: There was an oracle in Neuromancer? I don't remember that at ALL in Neuromancer. Was it just some kind of throwaway? Does Wintermute imply at some point that that's how it makes decisions?
posted by lodurr at 7:30 AM on September 19, 2005

...large groups of non-panicked, un-rushed, unfearful people are smarter than an elite few, no matter how brilliant—better at solving problems, fostering innovation, coming to wise decisions, even predicting the future as long as sufficient facts are presented, the terms of the question to be decided are made clear, and the decision does not endanger the participants...
posted by Mo Nickels at 8:45 AM on September 19, 2005

Kind of like on Who Wants To Be a Millionaire?, when they poll the audience, the audience always shows a strong preference toward the correct answer!
posted by knave at 11:39 AM on September 19, 2005

No... no.... I'm absolutely certain this is not true. As Bierce said:

MULTITUDE, n. A crowd; the source of political wisdom and virtue. In a republic, the object of the statesman's adoration. "In a multitude of counsellors there is wisdom," saith the proverb. If many men of equal individual wisdom are wiser than any one of them, it must be that they acquire the excess of wisdom by the mere act of getting together. Whence comes it? Obviously from nowhere – as well say that a range of mountains is higher than the single mountains composing it. A multitude is as wise as its wisest member if it obey him; if not, it is no wiser than its most foolish.
posted by BlackLeotardFront at 12:26 PM on September 19, 2005

Cosma Shalizi has had a couple interesting things to say on the topic:

Popular Delusions of Distributed Multi-Agent Systems
The Logic of Diversity
posted by sonofsamiam at 2:03 PM on September 19, 2005

"Yaayyyy!"
posted by uncanny hengeman at 4:48 PM on September 19, 2005

"my god, we've given mobs a bad name"
posted by mowglisambo at 9:54 PM on September 19, 2005

lodurr writes "BTW: There was an oracle in Neuromancer?"

I'm sure there was an oracle - however I read the book in 1993, so I might be totally and completely wrong here. (Hell, it might have been in a Heinlein book now that I think of it.) I am pretty sure I remember mention of this oracle thing, basically like a public opinion lottery used for predictive value, not a singular person. I was sure I remembered it as a phenomenon not explained until later in the novel, at first you're just given the idea of the predictive power and later the details on how predictions were made was filled in.
posted by caution live frogs at 12:12 PM on September 20, 2005