How to calculate the unknown

There are a few methods for making valuable “guestimations” about many systems.

Something that appears almost everywhere is the Gaussian curve. There is a theorem, called the central limit theorem, which states that the mean of numerous independent random variables is approximated by the normal distribution, regardless of the nature of the distribution they actually belong to.

Another law that is found to be reliable in many complex systems is the famous Pareto law, which states that 80% of the effects are determined by 20% of the causes and vice versa.

Obviously, percentages are quite approximate (although they very often get it right), but this ‘law’ can be verified in many systems, due to the fact that in a non-linear system, inputs affect the output unpredictably. This means that some of the initial smaller variations lead to large changes in the output, while others are cancelled out. The random character is precisely due to the impossibility of predicting which of the causes will be relevant and which will not.

A final useful and ubiquitous tool are Fermi’s estimates. They involve making reasoned assumptions about quantities that seem impossible to calculate from the limited information available. It is, in short, a technique for approximating with little or no information at hand by multiplying reasonable…