Sunday, October 02, 2005

(***) The Wisdom of Crowds by James Surowiecki

(***) The Wisdom of Crowds by James Surowiecki

[In 1906, British scientist Francis Galton analyzed the guesses to a contest where patrons wagered on the weight of butchered meat from a fat ox.] Galton arranged the guesses (787 in total) in order from highest to lowest and calculated the mean of the group’s guesses… Galton undoubtedly thought that the avg guess of the group would be way off the mark. After all, mix a few very smart people [butchers, farmers, cattlemen, well smart when it comes to slaughtering cattle at least] with some mediocre and dumb people [county fair patrons], and it seems likely that you’ll end up with a dumb answer. But Galton was wrong. The crowd had guessed that the ox, after it had been slaughtered and dressed, would weight 1,197lbs… The ox [actually] weighed 1,198lbs. In other word, the crowd’s judgement was essentially perfect.” Pxiii

In 1968, the US sub Scorpion disappeared in the North Atlantic… Although the Navy knew the sub’s last reported location, it had no idea what had happened to the Scorpion, and only the vaguest sense of how far it might have traveled after it had last made radio contact. As a result, the area where the navy began searching was a circle 20 miles wide and many 1000s of feet deep. You could not imagine a more hopeless task… A naval officer, John Craven concocted a series of scenarios… Then he assembled a team of men with a wide range of knowledge including mathematicians, submarine specialists, salvage men, etc. Instead of asking them to consult w/each other to come up with an answer, he asked each of them to offer his best guess about how likely each scenario was… He took all the guesses, and used a formula called Baye’s theorem to estimate the Scorpion’s final location. The location that Craven came up with was not a spot that any individual member of the group had picked… The final estimate was a genuinely collective judgement that the group as a whole had made, as opposed to representing the individual judgement of the smartest people in it… 5 months after the Scorpion disappeared, a navy ship found it. It was 220 yards from where Craven’s group had said it would be. Pxx

[From Who Wants to Be a Millionaire, you are allowed to ask the audience for help or an expert]. The experts did okay, offering the right answer – under pressure – almost 65% of the time… The random crowds of people with nothing better to do on a weekday afternoon than sit in a TV studio picked the right answer 91% of the time. P4

A classic demo of group intelligence is the jelly beans in the jar experiment, in which invariably the group’s estimate is superior to the vast majority of the individual guesses. When finance professor Jack Treynor ran the experiment in his class with a jar that held 850 beans, the group estimate was 871. Only 1 of the 56 people in the class made a better guess. P5

There are 2 lessons to draw from these experiments. 1st in most of them the members of the group were not talking to each other or working on the problem together. They were making individual guesses, which were aggregated and then averaged…2nd the group’s guess will not be better than that of every single person in the group each time… But there is no evidence in these studies that certain people consistently outperform the group… Over 10 experiments, the group’s performance will almost certainly be the best possible. P5

Groups that are too much alike find it harder to keep learning, because each member is bringing less and less new information to the table…Bringing new members into the organization, even if they’re less experienced and less capable, actually makes the group smarter simply because what little the new members do know is not redundant with what everyone else knows. P31

Between 1984 and 1999, almost 90% of mutual fund managers underperformed the Wilshire 5000 index [a relatively low bar]. P33

A number of studies have concluded that non-psychologists, for instance, are actually better at predicting people’s behavior than psychologists are… The between expert agreement in a host of fields, including stock picking, livestock judging, and clinical psychology is below 50% [meaning that at best the experts could be no better than below 50% right!]… More disconcertingly, one study found that the internal consistency o medical pathologists’ judgments was just .5, meaning that a pathologist presented with the same evidence would, ½ the time offer a different opinion. [Experts are prone to bias, and quick judgment, leading to overestimation of the likelihood that they’re right.] p33

A survey found that physicians, nurses, lawyers, engineers, entrepreneurs, and invst bankers all believed that they knew more than they did… It wasn’t just that they were wrong; they also didn’t have any idea how wrong they were. P34

Homogeneous groups become cohesive more easily than diverse groups, and as they become more cohesive they also become more dependent on the group, more insultated from outside opinions, and therefore more convinced that the group’s judgment on important issues must be right. [This is called groupthink, and it is not good.] p36

The classic illustration of the power of conformity is an experiment in which groups of people to judge which of 3 lines was the same size as a line on a white card. The groups of 7 to 9 had only 1 subject and the rest, unbeknownst to the subject, were confederates of the experimenters. The subject was placed at the end of the row of people, and each was asked to give his choice out loud. There were 12 cards in the experiment, and with the 1st 2 card, everyone in the group identified the same lines. Beginning with the 3rd card, the confederates began to pick lines that were clearly not the same size… The unwitting subjects changed the position of their heads to look at the lines from a different angle. They stood up to scrutinize the lines more closely… Most importantly, a significant number of the subjects simply went along with the group… 70% of the subjects changed their real [accurate] opinion at least once, and 33% went along with the group at least ½ the time. In one variant of the experiment, a confederate instead of going along with the group, picked the lines that matched the line on the card, effectively giving the subject an ally. And that was enough to make a huge difference… The rate of conformity plummeted. P38

The smartest groups are made up of people with diverse perspectives who are able to stay independent of each other. Independence doesn’t imply rationality or impartiality. You can be biased and irrational, but as long as you’re independent, you won’t make the group any dumber. P41

Social proof is the tendency to assume that if lots of people are doing something or believe something, there must be good reason why. This is different from conformity because you are not afraid of peer pressure or being reprimanded. [To test this] experimenter’s put a single person on a street corner and had him look up at an empty sky for 60 seconds. A tiny fraction of the passing pedestrians stopped to see what the guy was looking at, but most just walked past. Next time around, they put 5 skyward looking people on the corner. This time 4 times as many people stopped to gaze at the empty sky. When they put 15 people on the corner, 45% of all passersby stopped, and increasing the cohort yet again made more than 80% of the pedestrians tilt their heads. P43 [Could private religious faith be explained by this?]

People are in general overconfident. They overestimate their ability, their level of knowledge and their decision making prowess. And people are more overconfident when facing difficult problems than when facing easy one. This is not good for the overconfident decision makers themselves, since it means that they’re more likely to choose badly. But it good for society as whole, because overconfident people are less likely to get sucked into negative information cascade, and in the right circumstances are even able to break cascades. Remember that a cascade is kept going by people valuing public information more highly than private information. Overconfident people don’t do that… They make the public information seem less certain. And that encourages others to rely on themselves rather than just follow everyone else. P61

If you want to improve an organization’s decision making, one of the best things you can do is make sure that decisions are made simultaneously rather than one after another… One key to successful group decisions is getting people to pay much less attention to what everyone else is saying. P65

“There is no democracy in physics. We can’t say that some second rate guy has as much right to an opinion as Fermi” [Luis Alvarez, physicist]… You can’t listen to or read everyone, so you only listen to the best - it has a number of dubious assumptions built into it, including the idea that we automatically know who the second-rate are, even before hearing them, as well as the idea that everything [the expert] had to say was inherently valuable. The obvious peril is that important work will be ignored. P171

“If you can name for me one great discovery or decision that was made by a committee, I will find you the one man in that committee who had the lonely insight that solved the problem and was the basis for the decision.” Former GE Chairman Ralph Cordiner. [Yes, but can you tell me before the meeting who that one person will be?]

Social scientists who study juries often differentiate between 2 approaches. Evidence based juries usually don’t take a vote until after they’ve spent some time talking over the case, sifting the evidence. Verdict based juries see their mission as reaching a decision as quickly as possible. They take a vote before any discussion, and debate after that tends to concentrate on getting those who don’t agree to agree. [You can probably guess which type of jury you would like to have evaluate your next court case.] p178

One of the consistent findings from decades of small group research is that group deliberations are more successful when they have a clear agenda and when leaders take an active role in making sure that everyone gets a chance to speak... That matters because, in small groups, diversity of opinion is the single best guarantee that the group will reap benefits from face to face discussion. Berkely political scientists have shown in mock juries that a presence of a minority viewpoint, all by itself, makes a group’s decisions more nuanced and its decision making process more rigorous. [Always have everyone speak, and have a devil’s advocate] P182

As a general rule, discussions tend to move both the group as a whole and the individuals within it toward more extreme positions than ones they entered the discussion with. Why does polarization occur? One reason is because of people’s reliance on ‘social comparison’… It means that people are constantly comparing themselves to everyone else with an eye toward maintaining their relative position within the group. In other words, if the you start out in the middle of the group, and you believe the group has moved, as it were, to the right, you’re inclinded to shift your position to the right as well, so relative to everyone else you’re standing still. [And that is how groupthink can lead to shocking positions that no individual would ever have come up with eg. think OJ jury] P185

The virtue of decentralization are twofold... The more responsibility they have for their own environments, the more engaged they will be… 2 groups were put in rooms to solve puzzles while loud, random noises recurred in the background. 1 group was left alone. The other was given a button they could press to turn off the sound. The 2nd group solved 5 times as many puzzles… No member of the group ever pressed the button. Knowing it was there was all that mattered. The 2nd thing decentralization makes easier is coordination… Companies can rely on workers to find new, more efficient ways of getting things done. P212

The idea of the wisdom of crowds is not that group will always give you the right answer but that on average it will consistently come up with a better answer than any individual could provide. P235

The problem with the stock market is that there never is a point at which you can say that it’s over… This is one reason why a company’s stock price can easily soar far past any reasonable valuation, because people can always convince themselves that something in the future will happen to make the company worth it… Even if the market does eventually get the price right, it can be wrong for a long time, because there is no objective means to demonstrate it’s wrong… This is what Keynes meant when he said markets can stay wrong longer than you can stay solvent. P237

You don’t see bubbles in the real economy, which is to say the economy where you buy and sell TVs, apples, haircuts, etc. In other words the price of TVs doesn’t suddenly double overnight, only to crash to a few months later… And you never end up with a situation where the fact that prices are rising makes people more interested in buying… [When you buy a stock] you’re buying the right to resell that share of stock to someone else – ideally someone who has a more optimistic view of the company’s future than you do, and will therefore pay you more for the stock. P246

[A Caltech economic experiment to show how easily bubbles form] Everyone was given 2 shares to start and some money to buy more shares. Each share paid a dividend of $.24 at the end of each period [15 periods in the experiment]. If you owned 1 share for the entire experiment you’d get 15x.24=$3.60. So before the game started, if someone asked you how much you’d pay for a share, the correct answer would be no more than $3.60. After the 1st period ended, you’d be willing to pay no more than $3.36. And so on… Yet, when the experiment was run, the price of the shares jumped immediately to $3.50 and stayed there almost until the very end… What were the students thinking? “They’d say ‘Sure I knew that prices were way too high, but I saw other people buying and selling at high prices. I figured I could buy, collect a dividend or 2 and then sell at the same price to some other idiot.” P250

It’s not clear that barrage of news [and information] is necessarily conducive to good decision making. [An MIT experiment let students build a portfolio of stocks of their own choosing]. One group was allowed to see only the changes in the prices of their stocks. They could buy and sell if they wanted, but all they knew was the price… The 2nd group was allowed to see the changes in price, but was also given a constant stream of financial news… Suprisingly the less well informed group did far better… The reason was that news reports by their nature, overplay the importance of any particular piece of information. P254

Studies have shown that ideology does a much better job of predicting attitudes on issues than self interest does. For example, conservatives without health insurance still opposed national health insurance, while liberals who had health insurance favored it. P265

http://encyclopediaindex.com/b/tcrwd10.htm The crowd: A study of the popular mind 1982

With only a few exceptions, the market’s ranking of the horses predicted exactly the order in which horses finished. This was true no matter how many horses were in the face: the favorite finished most often, the 2nd favored horse finished second most often, and so on… Even more impressively is how well the odds predicted the frequency of victory. Take for instance, 312 races in 1 year which 7 horses ran. The favorite was predicted to win 33%. It won 34% of the time. 2nd place horse: 22% predicted, 21% real; 3rd place: 16%/16%; 4th place, 12%/12%, 5th place, 9%/8%; 6th place, 6%, 8%; 7th place, 3%, 2%.

3 comments:

Anonymous said...

This blog cites Surowiecki's criticism of medical pathologists. Since this is my profession, I know something about the issue of the accuracy of our work and I know that what Surowiecki writes is highly misleading. The evidence that he cites says nothing about the diagnostic accuracy of pathologists. See my critique at http://e-immunohistochemistry.info/Surowiecki.htm. If he is so wrong on something that I know about, I wonder how reliable is he on anything else?

Anonymous said...

Who knows where to download XRumer 5.0 Palladium?
Help, please. All recommend this program to effectively advertise on the Internet, this is the best program!

hcg said...

Hey, Really great work, I would like to join your blog anyway so please continue sharing with us,