Black Swans are what applied statistician Nassim Nicholas Taleb calls high impact economic or technological upheavals, similar to what Clay Christensen has termed disruptive innovations. In an interview with New Scientist, Taleb talks about his belief that we can’t forecast these events and that ‘experts’ who believe they can forecast them can be dangerous.
Taleb points to the problem of post-hoc rationalisation as part of the problem. We look at major events in the past and we think we can see the patterns that caused them, but we are leaving out all the events that could have happened, but didn’t. In the field of science Taleb talks about the problems caused by the choice of which studies are reported (generally the ones with successful results), causing us to confuse a finding of absences with an absence of findings.
As an example of the problems and dangers of forecasting Taleb quotes US oil price forecasts shown to him by somebody responsible for producing 25 year forecasts. In January 2004 the 25 year forecast was for $27 a barrel, six months later the 25 year forecast was $74 a barrel.
Taleb draws a distinction between what he refers to as Type I randomness and Type II. Type I randomness was much more common in pre-technological society and relates to things like the throw of a die, or whether you catch flu or not, this type of randomness averages itself out over time, so that in aggregate the patterns make sense, and can be predicted. Type II events are one-off events, they don’t average out because the distributions they belong to are so large and the events so infrequent that most types of events do not happen, or they happen only once, making predicting them impossible.
Taleb is publishing his ideas more fully later this year in his new book The Black Swan.
Comments