Originally circulated on 24 April 2020
Hi fellow Ruminants & Groupies in Lock Down
Greetings from 28 day of lock down. This week I have received a number of worthy submissions for Ruminant Pink Friday’s ™ which will be fleshed out in the coming weeks. Thank you for this. This week I need to complete the modelling topic because it’s one of the topics I have been pontificating on for many years. I can truly bore you to tears on the topic of modelling.
Last week I focussed on what we see being a simulation of reality and not reality itself. This does not mean that the simulation is not a good simulation. It is a very good simulation and we rely on what we see to avoid injury and death.
Today I am going to focus on the topic of modelling, what constitutes a good model and to what extent we properly understand the models we create.
I am going to start with a quote attributed to Albert Einstein regarding what constitutes a good model.
Everything should be made as simple as possible, but not simpler
Understanding must be conveyed as simply as possible for greatest communication. If you try to give too many rules without order or structure enabling you to see the connection, then understanding breaks down. You’ve got to see the forest before you can see the trees: the simple before the complexity of the details. https://fs.blog/2014/12/albert-einstein-simplicity/
I will now turn to the British statistician George Box who did a great deal of work on the study of modelling and provide his most relevant quote on modelling:
“Since all models are wrong the scientist cannot obtain a “correct” one by excessive elaboration. On the contrary following William of Occam he should seek an economical description of natural phenomena. Just as the ability to devise simple but evocative models is the signature of the great scientist so over-elaboration and over-parameterization is often the mark of mediocrity“.
For the more diligent readers I provide this short reference which is worth reading. https://en.wikipedia.org/wiki/All_models_are_wrong
Finally I will turn to the topic of understanding and explanation from one of the worlds true mavericks, David Deutsch. https://www.ted.com/talks/david_deutsch_a_new_way_to_explain_explanation?language=en
Deutsch essentially argues that if you can’t clearly explain your model based on guiding physical principles that are hard to vary then your model lacks meaning. So for example over the centuries there have been many theories and models to try to explain the seasons on earth. The current elegant and correct explanation is that the earth is tilted at an angle of 23.5 degrees relative to the orbital plane of earth’s orbit around the sun. That is all you need to know to build a model. The mathematics might get complicated but are built on this simple explanation. If you create a model without a simple explanatory framework then what you doing is not scientific and does not contribute to true understanding.
Please keep the submission ideas flowing.
Regards
Bruce
Ruminant Pink Friday 