FUTUREMATH Ltd
There are many kinds of unpredictability that we encounter in everyday life. That is why it is so hard to pick which fixed interest rate to choose, when to buy or sell a stock, or whether you can expect to be rained on tomorrow.
Surprisingly, all these things are governed by the same kind of statistical pattern. Rather than just being random events, these unpredictable phenomena display distinct patterns and structure.
In recent decades advances by mathematicians of the likes of Benoît Mandelbrot have shown that these phenomena display self-similarity, that is, small pieces of behavior resemble scaled-down versions of the whole.
Another key property is that periods of volatility are clumped together in time.
So, when it rains it really does pour!
The method used to characterise this is called multifractal analysis. It allows much more reasonable estimates of a system’s future behaviour than would be possible with more traditional statistical techniques.
Use of such techniques helps to understand the impact and statistics of unforeseen events - a subject that has been spoken about in depth by a recent book called "The Black Swan" by Nassim Nicholas Taleb.
These models aren’t better predictors, but they certainly describe risk in a more robust manner.