Bakkot comments on Open Thread, December 1-15, 2012 - Less Wrong Discussion
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (177)
Is anyone doing self-scoring on everyday predictions? I've been considering doing this for a while - writing down probability estimates for things like 'will have finished work by 7PM' and 'will have arrived on time' and even 'will rain' or 'friend will arrive on time', so I can detect systemic errors and try to correct for them. (In particular, if you are consistently over- or under-confident, you can improve your score by correcting for this without actually needing to be any more accurate.) This seems like a pretty straightforward way of getting `better' at predicting things, potentially very quickly, so I'm curious about other's experiences.
I do (sometimes) do this with private predictions on prediction book.
I haven't. Before I imitate you, I'd like to know, what specific implementations have you performed yet, and what are your results so far?