Thursday, December 22, 2016

Superforecasting

Forecasts are everywhere. But surprisingly, few ask and receive the track record of the forecasters. As a result, we end up consuming a lot of bad forecasts, as we end up hearing from those who are better at entertaining than they are at forecasting. But there is a better way.

In Superforecasting: The Art and Science of Prediction, the authors tell us what happened when they measured forecasters. They found that some repeatedly outperformed as forecasters, and tried to reverse-engineer their thinking.

There is some great stuff in the book that can make you a better forecaster. A lot of it is avoiding behavioural biases, the same ones that come up if one is familiar with Munger, Kahneman, Tversky etc. But a refresher on these biases is never a bad thing.

If you are going to read a book on forecasting though, I'd first recommend Nate Silver's terrific book The Signal and the Noise: Why So Many Predictions Fail--but Some Don't

No comments: