Today's post, The Error of Crowds was originally published on April 1, 2007. A summary (from the LW wiki):

Mean squared error drops when we average our predictions, but only because it uses a convex loss function. If you faced a concave loss function, you wouldn't isolate yourself from others, which casts doubt on the relevance of Jensen's inequality for rational communication. The process of sharing thoughts and arguing differences is not like taking averages.

Discuss the post here (rather than in the comments of the original post).

This post is part of a series rerunning Eliezer Yudkowsky's old posts so those interested can (re-)read and discuss them. The previous post was Useful Statistical Biases, and you can use the sequence_reruns tag or rss feed to follow the rest of the series.

Sequence reruns are a community-driven effort. You can participate by re-reading the sequence post, discussing it, posting the next day's sequence reruns post, summarizing forthcoming articles on the wiki, or creating exercises. Go here for more details, or to discuss the Sequence Reruns.

New Comment
5 comments, sorted by Click to highlight new comments since: Today at 6:30 AM

The post seems to imply that you should put your fingers in your ears; but wouldn't it be better to randomly choose between the single highest and single lowest first-round answers?

Nicely spotted.

The article contains a link to an image at, which is now bad. Does anyone have a similar image to replace it?

In potentially relevant research news: "How social influence can undermine the wisdom of crowd effect". (I haven't read the paper yet, only skimmed it.)

New to LessWrong?