I've been looking at Brier scores and log scoring from Tetlock-style forecasting recently and am trying to determine if they could be applied to my personal decision making.

However, I'm not really sold on how useful they are for everyday decisions. Most real situations have tiny samples and very messy outcomes, so I'm curious if anyone's actually used this stuff outside of a research environment. Does it tell you anything useful outside of an academic scenario, or does it only work when you have a ton of predictions?