All models are wrong (forged), but some are useful.

George Box

George E. P. Box (1919–2013) was a famous British statistician who made major contributions to statistics, quality control, and experimental design. He is known for above quote which caught my attention as it harbours some truth which I cannot deny.

All models are wrong (forged), but some are useful.

Consider the wording, in the quote. It claims that "All" models are wrong. Forget the usefulness of some models, which are wrong. I would rarther ask the question what is a model. A model by definition is just an interpretation of some process or some natural phenomena. Consider a model of the water cycle. Simply we can teach a five year old kid about how the water gets vaporised from the surface of the sea, and form clouds, to rain down on earth, and flow back to the sea. This is accurate model to explain some phenomena. But this doesn't account for glaciers. The water evaporate from trees and exhaled air. Not even water boiled in the kettle this morning, when you made a cup of coffee. Truth is we cannot ever make a complete model of the water cycle. That would need to track every water molecule that ever exist in this instance of time. Including the ones freshly arriving to the earth's atmosphere in commets. Only the real water cycle itself would mount to such a model. Then we can just call it a reality, not a model. But we don't have do that. We don't need to make perfect model which describe every variable. We can create a wrong model of the water cycle and still use this to explain some phenomena. Draw it on a board and be amazed ourselves.

Water as a model

Let's get back to the quote. Mind you, this is said by a statistician. If you do not know they tend to be good and obsessed with numbers, which are considered accurate. The models they make help us to determine amount of active compound which an 80 kg man with a stomach cancer needs, in his chemotherapy regimen. Which can mean the life and death if not done correctly. Similarly models are the first to be built upon, before any skyscraper reach its future glory in any modern city. If someone told me that these complex models are wrong, and by all means if that is an statistician, I would consider him insane. How can that be possible. After all I have never heard failure of any skyscraper. (Chemotherapy, a different story)

Why all models are wrong?

Truth is only the skyscraper itself knows how it's going to respond to the enviornment over time. It is true non of them have ended catastrophically. Thanks to modern mathematical and architectural models are being accurate enough, to persist during short span of human life.

Skyscraper Model

Here we can take a rest and ask, what else is a model in your understanding about life. How about your finance? A tree is a model. A family is a model. Society is a model. A country is a model. Philosophy mathematics, language and science itself are models. How about weired things that we cannot see or touch. How about dreams, soul, religion? If every model is wrong then all of the above models harbour some error in them. It is not obvious to someone who is blinded by their own believes. But if we look at ourselves throughout the life, we can see our models change over time when we gather more knowledge. Yet we can never reach a fundamental truth in any of the models. It is like riding an asymptote towards the eternity. Never reaching there.

Understanding above quote does two things to oneself. First is it let's identify the idiot within ourselves who wants to believe in a perfect model, and even such exists. Second is it let's you identify other idiots who are fooled by their own models of the world as they haven't even considered that there can be a flaw in their model. And its better to keep the mouth shut, in such arenas of egos.