(Quotes are capsules of information, reinforcement or enlightenment.)


“All models are wrong, but some are useful.”

George Box


Coaching Point: He further said, “Since all models are wrong the scientist must be alert to what is importantly wrong. It is inappropriate to be concerned about mice when there are tigers abroad.” That’s the Signal vs. Noise issue [S:1134].

In his work as a mathematician and statistician Box knew and used many models. Effectively. But he also knew that to blindly trust a model is dangerous. Let a model guide you, let it be useful. It’s also healthy to assume that, at some level, it’s wrong.

This is a skill which is not easily developed, the skill of finding and using useful models without being blind to their ‘wrongness.’

What is a model you find useful? And how is it also ‘wrong’?



Copyright 2019 Steve Straus. All rights reserved.