PhilGoetz comments on Extreme Rationality: It's Not That Great - Less Wrong

140 Post author: Yvain 09 April 2009 02:44AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (269)

You are viewing a single comment's thread.

Comment author: PhilGoetz 10 April 2009 03:43:21AM 10 points [-]

Sometimes, people do worse when they try to be rational because they have a poor model of rationality.

One error I commonly see is the belief that rationality means using logic, and that logic means not believing things unless they are proven. So someone tries to be "rational" by demanding proof of X before changing their behavior, even in a case where neither priors nor utilities favor not X. The untrained person may be doing something as naive as argument-counting (how many arguments in favor of X vs. not X), and is still likely to come out ahead of the person who requires proof.

A related error is using Boolean models where they are inappropriate. The most common error of this type is believing that a phenomenon, or a class of phenomena, can have only one explanation.