Crowd-sourcing intelligence was the idea behind the IARPA initiative that funded FWE.org, “Forecasting World Events.”
A crowdsourced prediction is that of a large group of unqualified people (henceforth referred to as “idiots”), which may be no more than feelings and whims are tallied day by day. The prediction is the outcome to a question worded something like,
“What are the chances that Moody’s will upgrade Tunisia’s bond rating by October, 2014?”
If you are one of the idiots trying to guess the answer, you enter a percentage into a box, and a little pie-chart on your screen is updated. You can do this as often as you like. Daily, a computer grinding away in a back room somewhere adds up how often you feel one way or the other. The secret formula to be discovered by the researchers is how to weigh the opinions of all the idiots to come up with an accuracy rating that beats the experts.
The C.I.A. was interested in this, because they have a large staff of professional analysts who should provide a better product.
A few individuals, also idiots, had individual abilities that were about as good as all the idiots summed together by secret formulas. For example, I was able to answer a similar question about Tunisian bond ratings. I did this by getting into the head of a hypothetical bond rater, imagining his disgust at all the numbers in the face of a volatile situation, and made a decision based on sentiment, not financial facts.
The reason this works so well so often is that people have processes that they use to bolster their decisions, but in many cases yield to sentiment. The intuition of that particular question has to do with anticipating when that occurs.
Because a crowd-sourcing participant does not make a face-to-face presentation, or write a report, one can change one’s mind on a whim. One can wind-mill without embarrassment, following the breeze of one’s mood and thoughts.
If you were in a situation where people were relying on you for guidance, this lack of consistency would not inspire confidence. So inspiring confidence is in opposition to being an accurate predictor. Ergo, never be loyal to yourself.
How do we solve this problem?