← Back to context

Comment by Aerbil313

10 hours ago

These "competitive forecasting" platforms seem very interesting. What do you need to be good at this sort of skill?

Strong beliefs that

- things rarely change,

- noise dominates most measurements,

- a good story is not evidence,

- it is very rarely possible to know things more certainly than 10--90 %.

Phrases to look up are "confirmation bias", "availability heuristic", "uncertainty calibration", "outside view", "Fermi method".

A good understanding of the world, your own capabilities, and how they relate to the capabilities of other players.

There is also some technical knowledge but that's easy to pick up.