How to be an accurate political pundit

It’s very simple really:

Top reshuffle prediction tip: for any job, tip a man for the role. In the long run, you’ll be massively right.

— Mark Pack (@markpack) October 7, 2013

Which also shows how predicting ‘more of the same’ is a pretty sure way of both (1) being right most of the time, but also (2) of missing all the really interesting stories.

It’s also a good example of why judging people simply by the number of predictions they get right can be very misleading – especially in American politics where most political races are in safe seats.

An apparently very high percentage accuracy rate simply tells you someone is good at predicting what will happen when nothing very much of interest happens.

It’s the ability to spot turning points and the unexpected that is the real mark of a brilliant forecaster, and that gets lost in the accumulation of statistics about ‘what everyone thought was bound to happen did happen’ easy calls on safe seats.