4 ms·

I think you are missing the point of the wisdom of crowds. Crowds are great at aggregating the influence the various factors that go into a complex problem. T

I think you are missing the point of the wisdom of crowds. Crowds are great at aggregating the influence the various factors that go into a complex problem. The example from the book was (I think) that crowds were great at say, estimating the weight of a cow. Or, in modern terms, the price of a stock. Crowds are terrible at being creative whether they are lead by a person or not.

ghaff 6y ago
"Wisdom of crowds" is pretty fascinating in that it does really well under some circumstances. A friend of mine has been running an Oscar contest for years and I grabbed some data from him a few years ago. We knew that "consensus" did well but I actually did some graphs and analysis; it's pretty striking: https://bitmason.blogspot.com/2012/04/crowdsourcing-prediction-and-other-data.html https://bitmason.blogspot.com/2012/04/crowdsourcing-predicti... Undergrad, we were were both on the movie committee where we had an attendance-guessing contest that had similar results. But... outside of some very specific and often artificial circumstances where the effect is very robust. It doesn't really work.
yboris 6y ago
An example I remember from the book was about locating a sunken submarine. The 'experts' related to submarine/naval stuff were all briefed together on the problem and were asked to think about it independently and then their individual ideas were brought together in some fashion. Their composite proposal ended up being very successful. That example from the book was meant to illustrate (as I understood it) that you can aggregate individual thoughts on complex subjects and arrive at a useful conclusion that may be closer to the truth than any individual alone. I shared my comment in response to a claim that brainstorming is a waste of time. It is definitely the case that certain ways to aggregate people's thoughts (e.g. "groupthink") can be worse than having a well-informed individual solve the problem. What I believe _is_ possible is to combine ideas of a group of people to get better results that choosing just a few "smart" individuals to make the decision. A far analogy is the voting system in the US. Because we don't have Approval Voting or RCV / IRV (Ranked Choice Voting / Instant Runoff Voting) we now have a 2 party system where 3rd parties can have deleterious effects. The way you aggregate REALLY matters.

ghaff 6y ago

"Wisdom of crowds" is pretty fascinating in that it does really well under some circumstances. A friend of mine has been running an Oscar contest for years and I grabbed some data from him a few years ago. We knew that "consensus" did well but I actually did some graphs and analysis; it's pretty striking: https://bitmason.blogspot.com/2012/04/crowdsourcing-prediction-and-other-data.html https://bitmason.blogspot.com/2012/04/crowdsourcing-predicti... Undergrad, we were were both on the movie committee where we had an attendance-guessing contest that had similar results. But... outside of some very specific and often artificial circumstances where the effect is very robust. It doesn't really work.

yboris 6y ago

An example I remember from the book was about locating a sunken submarine. The 'experts' related to submarine/naval stuff were all briefed together on the problem and were asked to think about it independently and then their individual ideas were brought together in some fashion. Their composite proposal ended up being very successful. That example from the book was meant to illustrate (as I understood it) that you can aggregate individual thoughts on complex subjects and arrive at a useful conclusion that may be closer to the truth than any individual alone. I shared my comment in response to a claim that brainstorming is a waste of time. It is definitely the case that certain ways to aggregate people's thoughts (e.g. "groupthink") can be worse than having a well-informed individual solve the problem. What I believe _is_ possible is to combine ideas of a group of people to get better results that choosing just a few "smart" individuals to make the decision. A far analogy is the voting system in the US. Because we don't have Approval Voting or RCV / IRV (Ranked Choice Voting / Instant Runoff Voting) we now have a 2 party system where 3rd parties can have deleterious effects. The way you aggregate REALLY matters.