Discussion about this post

User's avatar
Overcoming Bias Commenter's avatar

The first criterion there reminds me of Norbert Weiner's humble suggestion(1) that the amount of information carried by a message can be evaluated by finding the probability of that message's occurring in the space of all possible messages (pretty loosely defined), and then taking the negative logarithm of that quantity, which reminds me in the vaguest sense to the log-barrier function from ...econometrics? (I might really be stretching that reference there).

(1)http://open-site.org/Scienc...

Expand full comment
Robin Hanson's avatar

Stuart, those do seem like promising tests.

Expand full comment
1 more comment...

No posts