On Saturday I had an opportunity to hear James Surowiecki discus the potential of intelligent group forecasting and decision-making from his book,The Wisdom of Crowds

3 things necessaryto make groups intelligent:
(1) aggregating people’s judgments in a genuine bottom-up decision making process
(2) diversity: experience, background, pov
(3) independence of thought: Independence based solely on individual  knowledge, information or intuitions (No piggy-backing on others' ideas)

Diverse crowds are far likely to be more intelligent.  Important in two ways: expands range of information we have access to

What intelligence crowds look like?

Some support for crowd argument:
* Gameshow: Who wants to be a millionaire.  Crowd gets it right 90% of time
* Racetrack: odds at track predict finish almost perfectly
* Prediction Market: Business School @ University of Iowa
Iowa Electronic Market
Check out: http://www.Newsfutures.com

* JS builds case of crowd’s intelligence with stories….
* Points to companies: HP set up internal stock market.  Buy and sell shares during lunch hours.  Over 3 years, that market outperformed the organization's internal forecast over the last 3 years
* Google, Siemens, Lilly are using this methodology

When you aggregate judgments, the errors people make kinda fall away.  The key is that this works only during certain circum:

Some thoughts:
* We live in a society that worships expertise.
* Argument is not about experts are not irrelevant
* Mistake to try and seek out the 2 or 3 experts and to rely solely on their judgment
* Problem with experts that they don’t have a keen sense of their own blind spots
* Diversity allows us to stay away from “groupthink.”
* Devil’s advocate emerged in 13th & 14th century Catholic Church
* Important thing is that the same person can’t be the same devil’s  advocate every time or people will simply ignore them

See book: The madness of crowds

Groups are more often than not watered-down in their thinking-- mediocrity at best

* Organizations and teams often put too much weight on consensus:  This is when you end up with watered-down decisions (lowest common denominator)

Groups are smartest when everyone in them is acting as independently as possible.  

Why this stuff is so hard:
The reality is that human beings have a natural tendency to copy others.

The net is a classic double-edged sword:
1.  Opportunity to tap into collective intelligence of diverse group
2.  Makes  it very easy to break down independent thought
(instead of drawing into wide range of information, they got locked-in)
From perspective of collective intelligence, you end up with “circular mills.” EX: Happens with army ants. When they get lost, they just do what the other ant does.  At the end of the circular mill is hunger and death.
The great strength of the web and this model is in the cacophony of voices
* Randomize our connection

Collective intelligence parable: See example of USS Scorpion.  May, 1968 where Kraven assembled diverse crowd to find lost sub.  See work for details…

Two other things are necessary for the wisdom of the crowd to prevail:
* Need to know a right answer
* Need people to agree on what the problem is they’re trying to solve

Groups over 50 people are going to be pretty good with random questions

Even in small groups (8 people), the crowd is better than the smartest member