Korva is a scroll of the internet’s best ideas.
Don't tell advisers how it ended
Describing the hand
A poker hand shows how this works. Holding king-queen, a player opens with a raise, and an aggressive opponent who raises a lot responds. The question to bring to a group is whether to fold or raise again, with the hand described exactly that far.
Hiding the ending
Leaving out what happened next matters because these problems are infectious. If an adviser hears that the player raised again and the opponent moved all-in, the adviser will build a story that makes that result make sense, because that is how minds like to work.
A second trick is to tell the same story two ways, telling one adviser that it worked out well and another that it worked out poorly. The analyses that come back look strikingly different, and merging them gets closer to the true quality of the decision.
Opinions first
Beliefs spread the same way. When several people have interviewed a job candidate, each should write down an opinion separately before anyone talks, because discussion infects everyone like a virus. The person making the final call should keep from saying who they think should be hired, and the same goes for asking a friend about an article: ask for their opinion before giving yours.
Why do groups lose independent opinions?
In Solomon Asch's conformity experiments of the early 1950s, participants judged the lengths of lines after a group of actors had given an obviously wrong answer aloud. Roughly a third of the participants' answers followed the group's mistake, and about three quarters conformed at least once. When even one actor gave the correct answer, conformity dropped sharply. The results are among the best known in social psychology.
Does averaging separate guesses beat one expert?
In 1907, Francis Galton collected about 800 guesses at the weight of an ox at a livestock fair in Plymouth, England. The median guess was 1,207 pounds, and the ox's actual weight was 1,198 pounds. He reported the finding in the journal Nature under the title Vox Populi. The method works best when each guess is made independently of the others.
How does the Delphi method keep opinions separate?
The Delphi method was developed at the RAND Corporation in the 1950s by Olaf Helmer and Norman Dalkey. Experts answer questionnaires separately and anonymously, then see a summary of the group's answers and revise in further rounds. Because nobody knows who said what, a senior or loud voice cannot sway the others. The approach was first used for forecasting military and technological questions.
More from Talks at Google
