So said statistician George Box.
What is it with us humans that we need to know what will happen, when the only certainty is uncertainty?
Much is currently being discussed (with some benefit of hindsight) about the accuracy or not of the various mathematical models which have been predicting the impact of Covid-19.
One of the problems with increased human sophistication and advances in artificial intelligence is that it seems that many expect computers to give an instant answer about what the future will bring. This simply is not realistic. Models cannot foresee the future and technology cannot manage the unpredictable.
Essentially there are two parts to any model:
- The assumptions
These will never be entirely correct and are not under our control. However, they should be based on sensible ranges and rigorous research.
- The mechanisms
These are under our control and should always be correct, but human error means that this is not always the case.
No model can exactly represent any system in the real world and therefore can never be entirely true.
But some can be useful.